diff --git a/content/ar/404.md b/content/ar/404.md new file mode 100644 index 0000000000..da192c53c0 --- /dev/null +++ b/content/ar/404.md @@ -0,0 +1,8 @@ +--- +title: 404 +sidebar: false +--- + +Oops! You've reached a dead end. + +If you think something should be here, you can [open an issue](https://github.com/numpy/numpy.org/issues) on GitHub. diff --git a/content/ar/_index.md b/content/ar/_index.md new file mode 100644 index 0000000000..be88f9e642 --- /dev/null +++ b/content/ar/_index.md @@ -0,0 +1,49 @@ +--- +title: null +--- + +{{< grid columns="1 2 2 3" >}} + +[[item]] +type = 'card' +title = 'Powerful N-dimensional arrays' +body = ''' +Fast and versatile, the NumPy vectorization, indexing, and broadcasting concepts are the de-facto standards of array computing today. +''' + +[[item]] +type = 'card' +title = 'Numerical computing tools' +body = ''' +NumPy offers comprehensive mathematical functions, random number generators, linear algebra routines, Fourier transforms, and more. +''' + +[[item]] +type = 'card' +title = 'Open source' +body = ''' +Distributed under a liberal [BSD license](https://github.com/numpy/numpy/blob/main/LICENSE.txt), NumPy is developed and maintained [publicly on GitHub](https://github.com/numpy/numpy) by a vibrant, responsive, and diverse [community](/community). +''' + +[[item]] +type = 'card' +title = 'Interoperable' +body = ''' +NumPy supports a wide range of hardware and computing platforms, and plays well with distributed, GPU, and sparse array libraries. +''' + +[[item]] +type = 'card' +title = 'Performant' +body = ''' +The core of NumPy is well-optimized C code. Enjoy the flexibility of Python with the speed of compiled code. +''' + +[[item]] +type = 'card' +title = 'Easy to use' +body = ''' +NumPy's high level syntax makes it accessible and productive for programmers from any background or experience level. +''' + +{{< /grid>}} diff --git a/content/ar/about.md b/content/ar/about.md new file mode 100644 index 0000000000..243f7083ae --- /dev/null +++ b/content/ar/about.md @@ -0,0 +1,90 @@ +--- +title: About Us +sidebar: false +--- + +NumPy is an open source project that enables numerical computing with Python. It was created in 2005 building on the early work of the Numeric and Numarray libraries. NumPy will always be 100% open source software and free for all to use. It is released under the liberal terms of the [modified BSD license](https://github.com/numpy/numpy/blob/main/LICENSE.txt). + +NumPy is developed in the open on GitHub, through the consensus of the NumPy and wider scientific Python community. For more information on our governance approach, please see our [Governance Document](https://www.numpy.org/devdocs/dev/governance/index.html). + + +## Steering Council + +The NumPy Steering Council is the project's governing body. Its role is to ensure, through working with and serving the broader NumPy community, the long-term sustainability of the project, both as a software package and community. The NumPy Steering Council currently consists of the following members (in alphabetical order, by last name): + +- Sebastian Berg +- Ralf Gommers +- Charles Harris +- Stephan Hoyer +- Inessa Pawson +- Matti Picus +- Stéfan van der Walt +- Melissa Weber Mendonça +- Eric Wieser + +Emeritus: + +- Alex Griffing (2015-2017) +- Allan Haldane (2015-2021) +- Marten van Kerkwijk (2017-2019) +- Travis Oliphant (project founder, 2005-2012) +- Nathaniel Smith (2012-2021) +- Julian Taylor (2013-2021) +- Jaime Fernández del Río (2014-2021) +- Pauli Virtanen (2008-2021) + +To contact the NumPy Steering Council, please email numpy-team@googlegroups.com. + +## Teams + +The NumPy project leadership is actively working on diversifying contribution pathways to the project.
NumPy currently has the following teams: + +- development +- documentation +- triage +- website +- survey +- translations +- sprint mentors +- optimization +- funding and grants + +See the [Team](/teams) page for more info. + +## NumFOCUS Subcommittee + +- Charles Harris +- Ralf Gommers +- Inessa Pawson +- Sebastian Berg +- External member: Thomas Caswell + +## Sponsors + +NumPy receives direct funding from the following sources: +{{< sponsors >}} + + +## Institutional Partners + +Institutional Partners are organizations that support the project by employing people that contribute to NumPy as part of their job. Current Institutional Partners include: + +- UC Berkeley (Stéfan van der Walt) +- Quansight (Nathan Goldbaum, Ralf Gommers, Matti Picus, Melissa Weber Mendonça) +- NVIDIA (Sebastian Berg) + +{{< partners >}} + + +## Donate + +If you have found NumPy useful in your work, research, or company, please consider a donation to the project commensurate with your resources. Any amount helps! All donations will be used strictly to fund the development of NumPy’s open source software, documentation, and community. + +NumPy is a Sponsored Project of NumFOCUS, a 501(c)(3) nonprofit charity in the United States. NumFOCUS provides NumPy with fiscal, legal, and administrative support to help ensure the health and sustainability of the project. Visit [numfocus.org](https://numfocus.org) for more information. + +Donations to NumPy are managed by [NumFOCUS](https://numfocus.org). For donors in the United States, your gift is tax-deductible to the extent provided by law. As with any donation, you should consult with your tax advisor about your particular tax situation. + +NumPy's Steering Council will make the decisions on how to best use any funds received. Technical and infrastructure priorities are documented on the [NumPy Roadmap](https://www.numpy.org/neps/index.html#roadmap). + +{{}} + diff --git a/content/ar/arraycomputing.md b/content/ar/arraycomputing.md new file mode 100644 index 0000000000..abd29d11c1 --- /dev/null +++ b/content/ar/arraycomputing.md @@ -0,0 +1,21 @@ +--- +title: Array Computing +sidebar: false +--- + +*Array computing is the foundation of statistical, mathematical, scientific computing in various contemporary data science and analytics applications such as data visualization, digital signal processing, image processing, bioinformatics, machine learning, AI, and several others.* + +Large scale data manipulation and transformation depends on efficient, high-performance array computing. The language of choice for data analytics, machine learning, and productive numerical computing is **Python.** + +**Num**erical **Py**thon or NumPy is its de-facto standard Python programming language library that supports large, multi-dimensional arrays and matrices, and comes with a vast collection of high-level mathematical functions to operate on these arrays. + +Since the launch of NumPy in 2006, Pandas appeared on the landscape in 2008, and it was not until a couple of years ago that several array computing libraries showed up in succession, crowding the array computing landscape. Many of these newer libraries mimic NumPy-like features and capabilities, and pack newer algorithms and features geared towards machine learning and artificial intelligence applications. + +arraycl + +**Array computing** is based on **arrays** data structures. *Arrays* are used to organize vast amounts of data such that a related set of values can be easily sorted, searched, mathematically manipulated, and transformed easily and quickly. + +Array computing is *unique* as it involves operating on the data array *at once*. What this means is that any array operation applies to an entire set of values in one shot. This vectorized approach provides speed and simplicity by enabling programmers to code and operate on aggregates of data, without having to use loops of individual scalar operations. diff --git a/content/ar/case-studies/blackhole-image.md b/content/ar/case-studies/blackhole-image.md new file mode 100644 index 0000000000..22334ebed0 --- /dev/null +++ b/content/ar/case-studies/blackhole-image.md @@ -0,0 +1,80 @@ +--- +title: "Case Study: First Image of a Black Hole" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/blackhole.jpg' title = 'Black Hole M87' alt = 'black hole image' attribution = '(Image Credits: Event Horizon Telescope Collaboration)' attributionlink = 'https://www.jpl.nasa.gov/images/universe/20190410/blackhole20190410.jpg' +{{< /figure >}} + +{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="Katie Bouman, *Assistant Professor, Computing & Mathematical Sciences, Caltech*" +> }} Imaging the M87 Black Hole is like trying to see something that is by definition impossible to see. +> +> {{< /blockquote >}} + +## A telescope the size of the earth + +The [Event Horizon telescope (EHT)](https://eventhorizontelescope.org) is an array of eight ground-based radio telescopes forming a computational telescope the size of the earth, studing the universe with unprecedented sensitivity and resolution. The huge virtual telescope, which uses a technique called very-long-baseline interferometry (VLBI), has an angular resolution of [20 micro-arcseconds][resolution] — enough to read a newspaper in New York from a sidewalk café in Paris! + +### Key Goals and Results + +* **A New View of the Universe:** The groundwork for the EHT's groundbreaking image had been laid 100 years earlier when [Sir Arthur Eddington][eddington] yielded the first observational support of Einstein's theory of general relativity. + +* **The Black Hole:** EHT was trained on a supermassive black hole approximately 55 million light-years from Earth, lying at the center of the galaxy Messier 87 (M87) in the Virgo galaxy cluster. Its mass is 6.5 billion times the Sun's. It had been studied for [over 100 years](https://www.jpl.nasa.gov/news/news.php?feature=7385), but never before had a black hole been visually observed. + +* **Comparing Observations to Theory:** From Einstein’s general theory of relativity, scientists expected to find a shadow-like region caused by gravitational bending and capture of light. Scientists could use it to measure the black hole's enormous mass. + +### The Challenges + +* **Computational scale** + + EHT poses massive data-processing challenges, including rapid atmospheric phase fluctuations, large recording bandwidth, and telescopes that are widely dissimilar and geographically dispersed. + +* **Too much information** + + Each day EHT generates over 350 terabytes of observations, stored on helium-filled hard drives. Reducing the volume and complexity of this much data is enormously difficult. + +* **Into the unknown** + + When the goal is to see something never before seen, how can scientists be confident the image is correct? + +{{< figure >}} +src = '/images/content_images/cs/dataprocessbh.png' title = 'EHT Data Processing Pipeline' alt = 'data pipeline' align = 'center' attribution = '(Diagram Credits: The Astrophysical Journal, Event Horizon Telescope Collaboration)' attributionlink = 'https://iopscience.iop.org/article/10.3847/2041-8213/ab0c57' +{{< /figure >}} + +## NumPy’s Role + +What if there's a problem with the data? Or perhaps an algorithm relies too heavily on a particular assumption. Will the image change drastically if a single parameter is changed? + +The EHT collaboration met these challenges by having independent teams evaluate the data, using both established and cutting-edge image reconstruction techniques. When results proved consistent, they were combined to yield the first-of-a-kind image of the black hole. + +Their work illustrates the role the scientific Python ecosystem plays in advancing science through collaborative data analysis. + +{{< figure >}} +src = '/images/content_images/cs/bh_numpy_role.png' alt = 'role of numpy' title = 'The role of NumPy in Black Hole imaging' +{{< /figure >}} + +For example, the [`eht-imaging`][ehtim] Python package provides tools for simulating and performing image reconstruction on VLBI data. NumPy is at the core of array data processing used in this package, as illustrated by the partial software dependency chart below. + +{{< figure >}} +src = '/images/content_images/cs/ehtim_numpy.png' alt = 'ehtim dependency map highlighting numpy' title = 'Software dependency chart of ehtim package highlighting NumPy' +{{< /figure >}} + +Besides NumPy, many other packages, such as [SciPy](https://www.scipy.org) and [Pandas](https://pandas.io), are part of the data processing pipeline for imaging the black hole. The standard astronomical file formats and time/coordinate transformations were handled by [Astropy][astropy], while [Matplotlib][mpl] was used in visualizing data throughout the analysis pipeline, including the generation of the final image of the black hole. + +## Summary + +The efficient and adaptable n-dimensional array that is NumPy's central feature enabled researchers to manipulate large numerical datasets, providing a foundation for the first-ever image of a black hole. A landmark moment in science, it gives stunning visual evidence of Einstein’s theory. The achievement encompasses not only technological breakthroughs but also international collaboration among over 200 scientists and some of the world's best radio observatories. Innovative algorithms and data processing techniques, improving upon existing astronomical models, helped unfold a mystery of the universe. + +{{< figure >}} +src = '/images/content_images/cs/numpy_bh_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} + +[resolution]: https://eventhorizontelescope.org/press-release-april-10-2019-astronomers-capture-first-image-black-hole + +[eddington]: https://en.wikipedia.org/wiki/Eddington_experiment + +[ehtim]: https://github.com/achael/eht-imaging + +[astropy]: https://www.astropy.org/ +[mpl]: https://matplotlib.org/ diff --git a/content/ar/case-studies/cricket-analytics.md b/content/ar/case-studies/cricket-analytics.md new file mode 100644 index 0000000000..d0be271f50 --- /dev/null +++ b/content/ar/case-studies/cricket-analytics.md @@ -0,0 +1,72 @@ +--- +title: "Case Study: Cricket Analytics, the game changer!" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/ipl-stadium.png' title = 'IPLT20, the biggest Cricket Festival in India' alt = 'Indian Premier League Cricket cup and stadium' attribution = '(Image credits: IPLT20 (cup and logo) & Akash Yadav (stadium))' attributionlink = 'https://unsplash.com/@aksh1802' +{{< /figure >}} + +{{< blockquote cite="https://www.scoopwhoop.com/sports/ms-dhoni/" by="M S Dhoni, *International Cricket Player, ex-captain, Indian Team, plays for Chennai Super Kings in IPL*" +> }} You don't play for the crowd, you play for the country. +> +> {{< /blockquote >}} + +## About Cricket + +It would be an understatement to state that Indians love cricket. The game is played in just about every nook and cranny of India, rural or urban, popular with the young and the old alike, connecting billions in India unlike any other sport. Cricket enjoys lots of media attention. There is a significant amount of [money](https://www.statista.com/topics/4543/indian-premier-league-ipl/) and fame at stake. Over the last several years, technology has literally been a game changer. Audiences are spoilt for choice with streaming media, tournaments, affordable access to mobile based live cricket watching, and more. + +The Indian Premier League (IPL) is a professional Twenty20 cricket league, founded in 2008. It is one of the most attended cricketing events in the world, valued at [$6.7 billion](https://en.wikipedia.org/wiki/Indian_Premier_League) in 2019. + +Cricket is a game of numbers - the runs scored by a batsman, the wickets taken by a bowler, the matches won by a cricket team, the number of times a batsman responds in a certain way to a kind of bowling attack, etc. The capability to dig into cricketing numbers for both improving performance and studying the business opportunities, overall market, and economics of cricket via powerful analytics tools, powered by numerical computing software such as NumPy, is a big deal. Cricket analytics provides interesting insights into the game and predictive intelligence regarding game outcomes. + +Today, there are rich and almost infinite troves of cricket game records and statistics available, e.g., [ESPN cricinfo](https://stats.espncricinfo.com/ci/engine/stats/index.html) and [cricsheet](https://cricsheet.org). These and several such cricket databases have been used for [cricket analysis](https://www.researchgate.net/publication/336886516_Data_visualization_and_toss_related_analysis_of_IPL_teams_and_batsmen_performances) using the latest machine learning and predictive modelling algorithms. Media and entertainment platforms along with professional sports bodies associated with the game use technology and analytics for determining key metrics for improving match winning chances: + +* batting performance moving average, +* score forecasting, +* gaining insights into fitness and performance of a player against different opposition, +* player contribution to wins and losses for making strategic decisions on team composition + +{{< figure >}} +src = '/images/content_images/cs/cricket-pitch.png' title = 'Cricket Pitch, the focal point in the field' alt = 'A cricket pitch with bowler and batsmen' align = 'center' attribution = '(Image credit: Debarghya Das)' attributionlink = 'http://debarghyadas.com/files/IPLpaper.pdf' +{{< /figure >}} + +### Key Data Analytics Objectives + +* Sports data analytics are used not only in cricket but many [other sports](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx) for improving the overall team performance and maximizing winning chances. +* Real-time data analytics can help in gaining insights even during the game for changing tactics by the team and by associated businesses for economic benefits and growth. +* Besides historical analysis, predictive models are harnessed to determine the possible match outcomes that require significant number crunching and data science know-how, visualization tools and capability to include newer observations in the analysis. + +{{< figure >}} +src = '/images/content_images/cs/player-pose-estimator.png' alt = 'pose estimator' title = 'Cricket Pose Estimator' attribution = '(Image credit: connect.vin)' attributionlink = 'https://connect.vin/2019/05/ai-for-cricket-batsman-pose-analysis/' +{{< /figure >}} + +### The Challenges + +* **Data Cleaning and preprocessing** + + IPL has expanded cricket beyond the classic test match format to a much larger scale. The number of matches played every season across various formats has increased and so has the data, the algorithms, newer sports data analysis technologies and simulation models. Cricket data analysis requires field mapping, player tracking, ball tracking, player shot analysis, and several other aspects involved in how the ball is delivered, its angle, spin, velocity, and trajectory. All these factors together have increased the complexity of data cleaning and preprocessing. + +* **Dynamic Modeling** + + In cricket, just like any other sport, there can be a large number of variables related to tracking various numbers of players on the field, their attributes, the ball, and several possibilities of potential actions. The complexity of data analytics and modeling is directly proportional to the kind of predictive questions that are put forth during analysis and are highly dependent on data representation and the model. Things get even more challenging in terms of computation, data comparisons when dynamic cricket play predictions are sought such as what would have happened if the batsman had hit the ball at a different angle or velocity. + +* **Predictive Analytics Complexity** + + Much of the decision making in cricket is based on questions such as "how often does a batsman play a certain kind of shot if the ball delivery is of a particular type", or "how does a bowler change his line and length if the batsman responds to his delivery in a certain way". This kind of predictive analytics query requires highly granular dataset availability and the capability to synthesize data and create generative models that are highly accurate. + +## NumPy’s Role in Cricket Analytics + +Sports Analytics is a thriving field. Many researchers and companies [use NumPy](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx) and other PyData packages like Scikit-learn, SciPy, Matplotlib, and Jupyter, besides using the latest machine learning and AI techniques. NumPy has been used for various kinds of cricket related sporting analytics such as: + +* **Statistical Analysis:** NumPy's numerical capabilities help estimate the statistical significance of observational data or match events in the context of various player and game tactics, estimating the game outcome by comparison with a generative or static model. [Causal analysis](https://amplitude.com/blog/2017/01/19/causation-correlation) and [big data approaches](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4996805/) are used for tactical analysis. + +* **Data Visualization:** Data graphing and [visualization](https://towardsdatascience.com/advanced-sports-visualization-with-pandas-matplotlib-and-seaborn-9c16df80a81b) provide useful insights into relationship between various datasets. + +## Summary + +Sports Analytics is a game changer when it comes to how professional games are played, especially how strategic decision making happens, which until recently was primarily done based on “gut feeling" or adherence to past traditions. NumPy forms a solid foundation for a large set of Python packages which provide higher level functions related to data analytics, machine learning, and AI algorithms. These packages are widely deployed to gain real-time insights that help in decision making for game-changing outcomes, both on field as well as to draw inferences and drive business around the game of cricket. Finding out the hidden parameters, patterns, and attributes that lead to the outcome of a cricket match helps the stakeholders to take notice of game insights that are otherwise hidden in numbers and statistics. + +{{< figure >}} +src = '/images/content_images/cs/numpy_ca_benefits.png' alt = 'Diagram showing benefits of using NumPy for cricket analytics' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} diff --git a/content/ar/case-studies/deeplabcut-dnn.md b/content/ar/case-studies/deeplabcut-dnn.md new file mode 100644 index 0000000000..257905de91 --- /dev/null +++ b/content/ar/case-studies/deeplabcut-dnn.md @@ -0,0 +1,102 @@ +--- +title: "Case Study: DeepLabCut 3D Pose Estimation" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/mice-hand.gif' title = 'Analyzing mice hand-movement using DeepLapCut' alt = 'micehandanim' attribution = '(Source: www.deeplabcut.org )' attributionlink = 'http://www.mousemotorlab.org/deeplabcut' +{{< /figure >}} + +{{< blockquote cite="https://news.harvard.edu/gazette/story/newsplus/harvard-researchers-awarded-czi-open-source-award/" by="Alexander Mathis, *Assistant Professor, École polytechnique fédérale de Lausanne* ([EPFL](https://www.epfl.ch/en/))" +> }} Open Source Software is accelerating Biomedicine. DeepLabCut enables automated video analysis of animal behavior using Deep Learning. +> +> {{< /blockquote >}} + +## About DeepLabCut + +[DeepLabCut](https://github.com/DeepLabCut/DeepLabCut) is an open source toolbox that empowers researchers at hundreds of institutions worldwide to track behaviour of laboratory animals, with very little training data, at human-level accuracy. With DeepLabCut technology, scientists can delve deeper into the scientific understanding of motor control and behavior across animal species and timescales. + +Several areas of research, including neuroscience, medicine, and biomechanics, use data from tracking animal movement. DeepLabCut helps in understanding what humans and other animals are doing by parsing actions that have been recorded on film. Using automation for laborious tasks of tagging and monitoring, along with deep neural network based data analysis, DeepLabCut makes scientific studies involving observing animals, such as primates, mice, fish, flies etc., much faster and more accurate. + +{{< figure >}} +src = '/images/content_images/cs/race-horse.gif' title = 'Colored dots track the positions of a racehorse’s body part' alt = 'horserideranim' attribution = '(Source: Mackenzie Mathis)' +{{< /figure >}} + +DeepLabCut's non-invasive behavioral tracking of animals by extracting the poses of animals is crucial for scientific pursuits in domains such as biomechanics, genetics, ethology & neuroscience. Measuring animal poses non-invasively from video - without markers - in dynamically changing backgrounds is computationally challenging, both technically as well as in terms of resource needs and training data required. + +DeepLabCut allows researchers to estimate the pose of the subject, efficiently enabling them to quantify the behavior through a Python based software toolkit. With DeepLabCut, researchers can identify distinct frames from videos, digitally label specific body parts in a few dozen frames with a tailored GUI, and then the deep learning based pose estimation architectures in DeepLabCut learn how to pick out those same features in the rest of the video and in other similar videos of animals. It works across species of animals, from common laboratory animals such as flies and mice to more unusual animals like [cheetahs][cheetah-movement]. + +DeepLabCut uses a principle called [transfer learning](https://arxiv.org/pdf/1909.11229), which greatly reduces the amount of training data required and speeds up the convergence of the training period. Depending on the needs, users can pick different network architectures that provide faster inference (e.g. MobileNetV2), which can also be combined with real-time experimental feedback. DeepLabCut originally used the feature detectors from a top-performing human pose estimation architecture, called [DeeperCut](https://arxiv.org/abs/1605.03170), which inspired the name. The package now has been significantly changed to include additional architectures, augmentation methods, and a full front-end user experience. Furthermore, to support large-scale biological experiments DeepLabCut provides active learning capabilities so that users can increase the training set over time to cover edge cases and make their pose estimation algorithm robust within the specific context. + +Recently, the [DeepLabCut model zoo](http://www.mousemotorlab.org/dlc-modelzoo) was introduced, which provides pre-trained models for various species and experimental conditions from facial analysis in primates to dog posture. This can be run for instance in the cloud without any labeling of new data, or neural network training, and no programming experience is necessary. + +### Key Goals and Results + +* **Automation of animal pose analysis for scientific studies:** + + The primary objective of DeepLabCut technology is to measure and track posture of animals in a diverse settings. This data can be used, for example, in neuroscience studies to understand how the brain controls movement, or to elucidate how animals socially interact. Researchers have observed a [tenfold performance boost](https://www.biorxiv.org/content/10.1101/457242v1) with DeepLabCut. Poses can be inferred offline at up to 1200 frames per second (FPS). + +* **Creation of an easy-to-use Python toolkit for pose estimation:** + + DeepLabCut wanted to share their animal pose-estimation technology in the form of an easy to use tool that can be adopted by researchers easily. So they have created a complete, easy-to-use Python toolbox with project management features as well. These enable not only automation of pose-estimation but also managing the project end-to-end by helping the DeepLabCut Toolkit user right from the dataset collection stage to creating shareable and reusable analysis pipelines. + + Their [toolkit][DLCToolkit] is now available as open source. + + A typical DeepLabCut Workflow includes: + + - creation and refining of training sets via active learning + - creation of tailored neural networks for specific animals and scenarios + - code for large-scale inference on videos + - draw inferences using integrated visualization tools + +{{< figure >}} +src = '/images/content_images/cs/deeplabcut-toolkit-steps.png' title = 'Pose estimation steps with DeepLabCut' alt = 'dlcsteps' align = 'center' attribution = '(Source: DeepLabCut)' attributionlink = 'https://twitter.com/DeepLabCut/status/1198046918284210176/photo/1' +{{< /figure >}} + +### The Challenges + +* **Speed** + + Fast processing of animal behavior videos in order to measure their behavior and at the same time make scientific experiments more efficient, accurate. Extracting detailed animal poses for laboratory experiments, without markers, in dynamically changing backgrounds, can be challenging, both technically as well as in terms of resource needs and training data required. Coming up with a tool that is easy to use without the need for skills such as computer vision expertise that enables scientists to do research in more real-world contexts, is a non-trivial problem to solve. + +* **Combinatorics** + + Combinatorics involves assembly and integration of movement of multiple limbs into individual animal behavior. Assembling keypoints and their connections into individual animal movements and linking them across time is a complex process that requires heavy-duty numerical analysis, especially in case of multi-animal movement tracking in experiment videos. + +* **Data Processing** + + Last but not the least, array manipulation - processing large stacks of arrays corresponding to various images, target tensors and keypoints is fairly challenging. + +{{< figure >}} +src = '/images/content_images/cs/pose-estimation.png' title = 'Pose estimation variety and complexity' alt = 'challengesfig' align = 'center' attribution = '(Source: Mackenzie Mathis)' attributionlink = 'https://www.biorxiv.org/content/10.1101/476531v1.full.pdf' +{{< /figure >}} + +## NumPy's Role in meeting Pose Estimation Challenges + +NumPy addresses DeepLabCut technology's core need of numerical computations at high speed for behavioural analytics. Besides NumPy, DeepLabCut employs various Python software that utilize NumPy at their core, such as [SciPy](https://www.scipy.org), [Pandas](https://pandas.pydata.org), [matplotlib](https://matplotlib.org), [Tensorpack](https://github.com/tensorpack/tensorpack), [imgaug](https://github.com/aleju/imgaug), [scikit-learn](https://scikit-learn.org/stable/), [scikit-image](https://scikit-image.org) and [Tensorflow](https://www.tensorflow.org). + +The following features of NumPy played a key role in addressing the image processing, combinatorics requirements and need for fast computation in DeepLabCut pose estimation algorithms: + +* Vectorization +* Masked Array Operations +* Linear Algebra +* Random Sampling +* Reshaping of large arrays + +DeepLabCut utilizes NumPy’s array capabilities throughout the workflow offered by the toolkit. In particular, NumPy is used for sampling distinct frames for human annotation labeling, and for writing, editing and processing annotation data. Within TensorFlow the neural network is trained by DeepLabCut technology over thousands of iterations to predict the ground truth annotations from frames. For this purpose, target densities (scoremaps) are created to cast pose estimation as a image-to-image translation problem. To make the neural networks robust, data augmentation is employed, which requires the calculation of target scoremaps subject to various geometric and image processing steps. To make training fast, NumPy’s vectorization capabilities are leveraged. For inference, the most likely predictions from target scoremaps need to extracted and one needs to efficiently “link predictions to assemble individual animals”. + +{{< figure >}} +src = '/images/content_images/cs/deeplabcut-workflow.png' title = 'DeepLabCut Workflow' alt = 'workflow' attribution = '(Source: Mackenzie Mathis)' attributionlink = 'https://www.researchgate.net/figure/DeepLabCut-work-flow-The-diagram-delineates-the-work-flow-as-well-as-the-directory-and_fig1_329185962' +{{< /figure >}} + +## Summary + +Observing and efficiently describing behavior is a core tenant of modern ethology, neuroscience, medicine, and technology. [DeepLabCut](http://orga.cvss.cc/wp-content/uploads/2019/05/NathMathis2019.pdf) allows researchers to estimate the pose of the subject, efficiently enabling them to quantify the behavior. With only a small set of training images, the DeepLabCut Python toolbox allows training a neural network to within human level labeling accuracy, thus expanding its application to not only behavior analysis in the laboratory, but to potentially also in sports, gait analysis, medicine and rehabilitation studies. Complex combinatorics, data processing challenges faced by DeepLabCut algorithms are addressed through the use of NumPy's array manipulation capabilities. + +{{< figure >}} +src = '/images/content_images/cs/numpy_dlc_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} + +[cheetah-movement]: https://www.technologynetworks.com/neuroscience/articles/interview-a-deeper-cut-into-behavior-with-mackenzie-mathis-327618 + +[DLCToolkit]: https://github.com/DeepLabCut/DeepLabCut diff --git a/content/ar/case-studies/gw-discov.md b/content/ar/case-studies/gw-discov.md new file mode 100644 index 0000000000..08f54b7547 --- /dev/null +++ b/content/ar/case-studies/gw-discov.md @@ -0,0 +1,77 @@ +--- +title: "Case Study: Discovery of Gravitational Waves" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/gw_sxs_image.png' title = 'Gravitational Waves' alt = 'binary coalesce black hole generating gravitational waves' attribution = '(Image Credits: The Simulating eXtreme Spacetimes (SXS) Project at LIGO)' attributionlink = 'https://youtu.be/Zt8Z_uzG71o' +{{< /figure >}} + +{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="David Shoemaker, *LIGO Scientific Collaboration*" >}} The scientific Python ecosystem is critical infrastructure for the research done at LIGO. +{{< /blockquote >}} + +## About [Gravitational Waves](https://www.nationalgeographic.com/news/2017/10/what-are-gravitational-waves-ligo-astronomy-science/) and [LIGO](https://www.ligo.caltech.edu) + +Gravitational waves are ripples in the fabric of space and time, generated by cataclysmic events in the universe such as collision and merging of two black holes or coalescing binary stars or supernovae. Observing GW can not only help in studying gravity but also in understanding some of the obscure phenomena in the distant universe and its impact. + +The [Laser Interferometer Gravitational-Wave Observatory (LIGO)](https://www.ligo.caltech.edu) was designed to open the field of gravitational-wave astrophysics through the direct detection of gravitational waves predicted by Einstein’s General Theory of Relativity. It comprises two widely separated interferometers within the United States — one in Hanford, Washington and the other in Livingston, Louisiana — operated in unison to detect gravitational waves. Each of them has multi-kilometer-scale gravitational wave detectors that use laser interferometry. The LIGO Scientific Collaboration (LSC), is a group of more than 1000 scientists from universities around the United States and in 14 other countries supported by more than 90 universities and research institutes; approximately 250 students actively contributing to the collaboration. The new LIGO discovery is the first observation of gravitational waves themselves, made by measuring the tiny disturbances the waves make to space and time as they pass through the earth. It has opened up new astrophysical frontiers that explore the warped side of the universe—objects and phenomena that are made from warped spacetime. + + +### Key Objectives + +* Though its [mission](https://www.ligo.caltech.edu/page/what-is-ligo) is to detect gravitational waves from some of the most violent and energetic processes in the Universe, the data LIGO collects may have far-reaching effects on many areas of physics including gravitation, relativity, astrophysics, cosmology, particle physics, and nuclear physics. +* Crunch observed data via numerical relativity computations that involves complex maths in order to discern signal from noise, filter out relevant signal and statistically estimate significance of observed data +* Data visualization so that the binary / numerical results can be comprehended. + + + +### The Challenges + +* **Computation** + + Gravitational Waves are hard to detect as they produce a very small effect and have tiny interaction with matter. Processing and analyzing all of LIGO's data requires a vast computing infrastructure.After taking care of noise, which is billions of times of the signal, there is still very complex relativity equations and huge amounts of data which present a computational challenge: [O(10^7) CPU hrs needed for binary merger analyses](https://youtu.be/7mcHknWWzNI) spread on 6 dedicated LIGO clusters + +* **Data Deluge** + + As observational devices become more sensitive and reliable, the challenges posed by data deluge and finding a needle in a haystack rise multi-fold. LIGO generates terabytes of data every day! Making sense of this data requires an enormous effort for each and every detection. For example, the signals being collected by LIGO must be matched by supercomputers against hundreds of thousands of templates of possible gravitational-wave signatures. + +* **Visualization** + + Once the obstacles related to understanding Einstein’s equations well enough to solve them using supercomputers are taken care of, the next big challenge was making data comprehensible to the human brain. Simulation modeling as well as signal detection requires effective visualization techniques. Visualization also plays a role in lending more credibility to numerical relativity in the eyes of pure science aficionados, who did not give enough importance to numerical relativity until imaging and simulations made it easier to comprehend results for a larger audience. Speed of complex computations and rendering, re-rendering images and simulations using latest experimental inputs and insights can be a time consuming activity that challenges researchers in this domain. + +{{< figure >}} +src = '/images/content_images/cs/gw_strain_amplitude.png' alt = 'gravitational waves strain amplitude' title = 'Estimated gravitational-wave strain amplitude from GW150914' attribution = '(Graph Credits: Observation of Gravitational Waves from a Binary Black Hole Merger, ResearchGate Publication)' attributionlink = 'https://www.researchgate.net/publication/293886905_Observation_of_Gravitational_Waves_from_a_Binary_Black_Hole_Merger' +{{< /figure >}} + +## NumPy’s Role in the Detection of Gravitational Waves + +Gravitational waves emitted from the merger cannot be computed using any technique except brute force numerical relativity using supercomputers. The amount of data LIGO collects is as incomprehensibly large as gravitational wave signals are small. + +NumPy, the standard numerical analysis package for Python, was utilized by the software used for various tasks performed during the GW detection project at LIGO. NumPy helped in solving complex maths and data manipulation at high speed. Here are some examples: + +* [Signal Processing](https://www.uv.es/virgogroup/Denoising_ROF.html): Glitch detection, [Noise identification and Data Characterization](https://ep2016.europython.eu/media/conference/slides/pyhton-in-gravitational-waves-research-communities.pdf) (NumPy, scikit-learn, scipy, matplotlib, pandas, pyCharm) +* Data retrieval: Deciding which data can be analyzed, figuring out whether it contains a signal - needle in a haystack +* Statistical analysis: estimate the statistical significance of observational data, estimating the signal parameters (e.g. masses of stars, spin velocity, and distance) by comparison with a model. +* Visualization of data + - Time series + - Spectrograms +* Compute Correlations +* Key [Software](https://github.com/lscsoft) developed in GW data analysis such as [GwPy](https://gwpy.github.io/docs/stable/overview.html) and [PyCBC](https://pycbc.org) uses NumPy and AstroPy under the hood for providing object based interfaces to utilities, tools, and methods for studying data from gravitational-wave detectors. + +{{< figure >}} +src = '/images/content_images/cs/gwpy-numpy-dep-graph.png' alt = 'gwpy-numpy depgraph' title = 'Dependency graph showing how GwPy package depends on NumPy' +{{< /figure >}} + +---- + +{{< figure >}} +src = '/images/content_images/cs/PyCBC-numpy-dep-graph.png' alt = 'PyCBC-numpy depgraph' title = 'Dependency graph showing how PyCBC package depends on NumPy' +{{< /figure >}} + +## Summary + +GW detection has enabled researchers to discover entirely unexpected phenomena while providing new insight into many of the most profound astrophysical phenomena known. Number crunching and data visualization is a crucial step that helps scientists gain insights into data gathered from the scientific observations and understand the results. The computations are complex and cannot be comprehended by humans unless it is visualized using computer simulations that are fed with the real observed data and analysis. NumPy along with other Python packages such as matplotlib, pandas, and scikit-learn is [enabling researchers](https://www.gw-openscience.org/events/GW150914/) to answer complex questions and discover new horizons in our understanding of the universe. + +{{< figure >}} +src = '/images/content_images/cs/numpy_gw_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} diff --git a/content/ar/citing-numpy.md b/content/ar/citing-numpy.md new file mode 100644 index 0000000000..5bb5d791b4 --- /dev/null +++ b/content/ar/citing-numpy.md @@ -0,0 +1,35 @@ +--- +title: Citing NumPy +sidebar: false +--- + +If NumPy has been significant in your research, and you would like to acknowledge the project in your academic publication, we suggest citing the following paper: + +* Harris, C.R., Millman, K.J., van der Walt, S.J. et al. _Array programming with NumPy_. Nature 585, 357–362 (2020). DOI: [10.1038/s41586-020-2649-2](https://doi.org/10.1038/s41586-020-2649-2). ([Publisher link](https://www.nature.com/articles/s41586-020-2649-2)). + +_In BibTeX format:_ + + ``` +@Article{ harris2020array, + title = {Array programming with {NumPy}}, + author = {Charles R. Harris and K. Jarrod Millman and St{\'{e}}fan J. + van der Walt and Ralf Gommers and Pauli Virtanen and David + Cournapeau and Eric Wieser and Julian Taylor and Sebastian + Berg and Nathaniel J. Smith and Robert Kern and Matti Picus + and Stephan Hoyer and Marten H. van Kerkwijk and Matthew + Brett and Allan Haldane and Jaime Fern{\'{a}}ndez del + R{\'{i}}o and Mark Wiebe and Pearu Peterson and Pierre + G{\'{e}}rard-Marchant and Kevin Sheppard and Tyler Reddy and + Warren Weckesser and Hameer Abbasi and Christoph Gohlke and + Travis E. Oliphant}, + year = {2020}, + month = sep, + journal = {Nature}, + volume = {585}, + number = {7825}, + pages = {357--362}, + doi = {10.1038/s41586-020-2649-2}, + publisher = {Springer Science and Business Media {LLC}}, + url = {https://doi.org/10.1038/s41586-020-2649-2} +} +``` diff --git a/content/ar/code-of-conduct.md b/content/ar/code-of-conduct.md new file mode 100644 index 0000000000..5ee1f4bcbe --- /dev/null +++ b/content/ar/code-of-conduct.md @@ -0,0 +1,83 @@ +--- +title: NumPy Code of Conduct +sidebar: false +aliases: + - /conduct.html +--- + +### Introduction + +This Code of Conduct applies to all spaces managed by the NumPy project, including all public and private mailing lists, issue trackers, wikis, blogs, Twitter, and any other communication channel used by our community. The NumPy project does not organise in-person events, however events related to our community should have a code of conduct similar in spirit to this one. + +This Code of Conduct should be honored by everyone who participates in the NumPy community formally or informally, or claims any affiliation with the project, in any project-related activities and especially when representing the project, in any role. + +This code is not exhaustive or complete. It serves to distill our common understanding of a collaborative, shared environment and goals. Please try to follow this code in spirit as much as in letter, to create a friendly and productive environment that enriches the surrounding community. + +### Specific Guidelines + +We strive to: + +1. Be open. We invite anyone to participate in our community. We prefer to use public methods of communication for project-related messages, unless discussing something sensitive. This applies to messages for help or project-related support, too; not only is a public support request much more likely to result in an answer to a question, it also ensures that any inadvertent mistakes in answering are more easily detected and corrected. +2. Be empathetic, welcoming, friendly, and patient. We work together to resolve conflict, and assume good intentions. We may all experience some frustration from time to time, but we do not allow frustration to turn into a personal attack. A community where people feel uncomfortable or threatened is not a productive one. +3. Be collaborative. Our work will be used by other people, and in turn we will depend on the work of others. When we make something for the benefit of the project, we are willing to explain to others how it works, so that they can build on the work to make it even better. Any decision we make will affect users and colleagues, and we take those consequences seriously when making decisions. +4. Be inquisitive. Nobody knows everything! Asking questions early avoids many problems later, so we encourage questions, although we may direct them to the appropriate forum. We will try hard to be responsive and helpful. +5. Be careful in the words that we choose. We are careful and respectful in our communication, and we take responsibility for our own speech. Be kind to others. Do not insult or put down other participants. We will not accept harassment or other exclusionary behaviour, such as: + * Violent threats or language directed against another person. + * Sexist, racist, or otherwise discriminatory jokes and language. + * Posting sexually explicit or violent material. + * Posting (or threatening to post) other people’s personally identifying information (“doxing”). + * Sharing private content, such as emails sent privately or non-publicly, or unlogged forums such as IRC channel history, without the sender’s consent. + * Personal insults, especially those using racist or sexist terms. + * Unwelcome sexual attention. + * Excessive profanity. Please avoid swearwords; people differ greatly in their sensitivity to swearing. + * Repeated harassment of others. In general, if someone asks you to stop, then stop. + * Advocating for, or encouraging, any of the above behaviour. + +### Diversity Statement + +The NumPy project welcomes and encourages participation by everyone. We are committed to being a community that everyone enjoys being part of. Although we may not always be able to accommodate each individual’s preferences, we try our best to treat everyone kindly. + +No matter how you identify yourself or how others perceive you: we welcome you. Though no list can hope to be comprehensive, we explicitly honour diversity in: age, culture, ethnicity, genotype, gender identity or expression, language, national origin, neurotype, phenotype, political beliefs, profession, race, religion, sexual orientation, socioeconomic status, subculture and technical ability, to the extent that these do not conflict with this code of conduct. + +Though we welcome people fluent in all languages, NumPy development is conducted in English. + +Standards for behaviour in the NumPy community are detailed in the Code of Conduct above. Participants in our community should uphold these standards in all their interactions and help others to do so as well (see next section). + +### Reporting Guidelines + +We know that it is painfully common for internet communication to start at or devolve into obvious and flagrant abuse. We also recognize that sometimes people may have a bad day, or be unaware of some of the guidelines in this Code of Conduct. Please keep this in mind when deciding on how to respond to a breach of this Code. + +For clearly intentional breaches, report those to the Code of Conduct Committee (see below). For possibly unintentional breaches, you may reply to the person and point out this code of conduct (either in public or in private, whatever is most appropriate). If you would prefer not to do that, please feel free to report to the Code of Conduct Committee directly, or ask the Committee for advice, in confidence. + +You can report issues to the NumPy Code of Conduct Committee at numpy-conduct@googlegroups.com. + +Currently, the Committee consists of: + +* Stefan van der Walt +* Melissa Weber Mendonça +* Rohit Goswami + +If your report involves any members of the Committee, or if they feel they have a conflict of interest in handling it, then they will recuse themselves from considering your report. Alternatively, if for any reason you feel uncomfortable making a report to the Committee, then you can also contact senior NumFOCUS staff at [conduct@numfocus.org](https://numfocus.org/code-of-conduct#persons-responsible). + +### Incident reporting resolution & Code of Conduct enforcement + +_This section summarizes the most important points, more details can be found in_ [NumPy Code of Conduct - How to follow up on a report](report-handling-manual). + +We will investigate and respond to all complaints. The NumPy Code of Conduct Committee and the NumPy Steering Committee (if involved) will protect the identity of the reporter, and treat the content of complaints as confidential (unless the reporter agrees otherwise). + +In case of severe and obvious breaches, e.g. personal threat or violent, sexist or racist language, we will immediately disconnect the originator from NumPy communication channels; please see the manual for details. + +In cases not involving clear severe and obvious breaches of this Code of Conduct the process for acting on any received Code of Conduct violation report will be: + +1. acknowledge report is received, +2. reasonable discussion/feedback, +3. mediation (if feedback didn’t help, and only if both reporter and reportee agree to this), +4. enforcement via transparent decision (see [Resolutions](report-handling-manual/#resolutions)) by the Code of Conduct Committee. + +The Committee will respond to any report as soon as possible, and at most within 72 hours. + +### Endnotes + +We are thankful to the groups behind the following documents, from which we drew content and inspiration: + +- [The SciPy Code of Conduct](https://docs.scipy.org/doc/scipy/dev/conduct/code_of_conduct.html) diff --git a/content/ar/community.md b/content/ar/community.md new file mode 100644 index 0000000000..5034fba239 --- /dev/null +++ b/content/ar/community.md @@ -0,0 +1,66 @@ +--- +title: Community +sidebar: false +--- + +NumPy is a community-driven open source project developed by a diverse group of [contributors](/teams/). The NumPy leadership has made a strong commitment to creating an open, inclusive, and positive community. Please read the [NumPy Code of Conduct](/code-of-conduct) for guidance on how to interact with others in a way that makes the community thrive. + +We offer several communication channels to learn, share your knowledge and connect with others within the NumPy community. + + +## Participate online + +The following are ways to engage directly with the NumPy project and community. _Please note that we encourage users and community members to support each other for usage questions - see [Get Help](/gethelp)._ + + +### [NumPy mailing list](https://mail.python.org/mailman/listinfo/numpy-discussion) + +This list is the main forum for longer-form discussions, like adding new features to NumPy, making changes to the NumPy Roadmap, and all kinds of project-wide decision making. Announcements about NumPy, such as for releases, developer meetings, sprints or conference talks are also made on this list. + +On this list please use bottom posting, reply to the list (rather than to another sender), and don't reply to digests. A searchable archive of this list is available [here](https://mail.python.org/archives/list/numpy-discussion@python.org/). + +*** + +### [GitHub issue tracker](https://github.com/numpy/numpy/issues) + +- For bug reports (e.g. "`np.arange(3).shape` returns `(5,)`, when it should return `(3,)`"); +- documentation issues (e.g. "I found this section unclear"); +- and feature requests (e.g. "I would like to have a new interpolation method in `np.percentile`"). + +_Please note that GitHub is not the right place to report a security vulnerability. If you think you have found a security vulnerability in NumPy, please report it [here](https://tidelift.com/docs/security)._ + +*** + +### [Slack](https://numpy-team.slack.com) + +A real-time chat room to ask questions about _contributing_ to NumPy. This is a private space, specifically meant for people who are hesitant to bring up their questions or ideas on a large public mailing list or GitHub. Please see [here](https://numpy.org/devdocs/dev/index.html#contributing-to-numpy) for more details and how to get an invite. + + +## Study Groups and Meetups + +If you would like to find a local meetup or study group to learn more about NumPy and the wider ecosystem of Python packages for data science and scientific computing, we recommend exploring the [PyData meetups](https://www.meetup.com/pro/pydata/) (150+ meetups, 100,000+ members). + +NumPy also organizes in-person sprints for its team and interested contributors occasionally. These are typically planned several months in advance and will be announced on the [mailing list](https://mail.python.org/mailman/listinfo/numpy-discussion) and [Twitter](https://twitter.com/numpy_team). + + +## Conferences + +The NumPy project doesn't organize its own conferences. The conferences that have traditionally been most popular with NumPy maintainers, contributors and users are the SciPy and PyData conference series: + +- [SciPy US](https://conference.scipy.org) +- [EuroSciPy](https://www.euroscipy.org) +- [SciPy Latin America](https://www.scipyla.org) +- [SciPy India](https://scipy.in) +- [SciPy Japan](https://conference.scipy.org) +- [PyData conferences](https://pydata.org/event-schedule/) (15-20 events a year spread over many countries) + +Many of these conferences include tutorial days that cover NumPy and/or sprints where you can learn how to contribute to NumPy or related open source projects. + + +## Join the NumPy community + +To thrive, the NumPy project needs your expertise and enthusiasm. Not a coder? Not a problem! There are many ways to contribute to NumPy. + +If you are interested in becoming a NumPy contributor (yay!) we recommend checking out our [Contribute](/contribute) page. + +Also, feel free to stop by and say hi at one of our community meetings. To keep track of them, check out our events calendar [here](https://scientific-python.org/calendars/). diff --git a/content/ar/config.yaml b/content/ar/config.yaml new file mode 100644 index 0000000000..4aaf75b2e6 --- /dev/null +++ b/content/ar/config.yaml @@ -0,0 +1,140 @@ +languageName: English +params: + description: Why NumPy? Powerful n-dimensional arrays. Numerical computing tools. Interoperable. Performant. Open source. + navbarlogo: + image: logo.svg + text: NumPy + link: / + hero: + #Main hero title + title: NumPy + #Hero subtitle (optional) + subtitle: The fundamental package for scientific computing with Python + #Button text + buttontext: "Latest release: NumPy 1.26. View all releases" + #Where the main hero button links to + buttonlink: "/news/#releases" + #Hero image (from static/images/___) + image: logo.svg + shell: + title: placeholder + intro: + - + title: Try NumPy + text: Use the interactive shell to try NumPy in the browser + docslink: Don't forget to check out the docs. + casestudies: + title: CASE STUDIES + features: + - + title: First Image of a Black Hole + text: How NumPy, together with libraries like SciPy and Matplotlib that depend on NumPy, enabled the Event Horizon Telescope to produce the first ever image of a black hole + img: /images/content_images/case_studies/blackhole.png + alttext: First image of a black hole. It is an orange circle in a black background. + url: /case-studies/blackhole-image + - + title: Detection of Gravitational Waves + text: In 1916, Albert Einstein predicted gravitational waves; 100 years later their existence was confirmed by LIGO scientists using NumPy. + img: /images/content_images/case_studies/gravitional.png + alttext: Two orbs orbiting each other. They are displacing gravity around them. + url: /case-studies/gw-discov + - + title: Sports Analytics + text: Cricket Analytics is changing the game by improving player and team performance through statistical modelling and predictive analytics. NumPy enables many of these analyses. + img: /images/content_images/case_studies/sports.jpg + alttext: Cricket ball on green field. + url: /case-studies/cricket-analytics + - + title: Pose Estimation using deep learning + text: DeepLabCut uses NumPy for accelerating scientific studies that involve observing animal behavior for better understanding of motor control, across species and timescales. + img: /images/content_images/case_studies/deeplabcut.png + alttext: Cheetah pose analysis + url: /case-studies/deeplabcut-dnn + tabs: + title: ECOSYSTEM + section5: false + navbar: + - + title: Install + url: /install + - + title: Documentation + url: https://numpy.org/doc/stable + - + title: Learn + url: /learn + - + title: Community + url: /community + - + title: About Us + url: /about + - + title: News + url: /news + - + title: Contribute + url: /contribute + footer: + logo: logo.svg + socialmediatitle: "" + socialmedia: + - + link: https://github.com/numpy/numpy + icon: github + - + link: https://www.youtube.com/channel/UCguIL9NZ7ybWK5WQ53qbHng + icon: youtube + - + link: https://twitter.com/numpy_team + icon: twitter + quicklinks: + column1: + title: "" + links: + - + text: Install + link: /install + - + text: Documentation + link: https://numpy.org/doc/stable + - + text: Learn + link: /learn + - + text: Citing Numpy + link: /citing-numpy + - + text: Roadmap + link: https://numpy.org/neps/roadmap.html + column2: + links: + - + text: About us + link: /about + - + text: Community + link: /community + - + text: User surveys + link: /user-surveys + - + text: Contribute + link: /contribute + - + text: Code of conduct + link: /code-of-conduct + column3: + links: + - + text: Get help + link: /gethelp + - + text: Terms of use + link: /terms + - + text: Privacy + link: /privacy + - + text: Press kit + link: /press-kit diff --git a/content/ar/contribute.md b/content/ar/contribute.md new file mode 100644 index 0000000000..6efff53624 --- /dev/null +++ b/content/ar/contribute.md @@ -0,0 +1,66 @@ +--- +title: Contribute to NumPy +sidebar: false +--- + +The NumPy project welcomes your expertise and enthusiasm! Your choices aren't limited to programming, as you can see below there are many areas where we need **your** help. + +If you're unsure where to start or how your skills fit in, _reach out!_ You can ask on the [mailing list](https://mail.python.org/mailman/listinfo/numpy-discussion) or [GitHub](http://github.com/numpy/numpy) (open an [issue](https://github.com/numpy/numpy/issues) or comment on a relevant issue). + +Those are our preferred channels (open source is open by nature), but if you prefer to talk privately, contact our community coordinators at or on [Slack](https://numpy-team.slack.com) (write for an invite). + +We also have a biweekly _community call_, details of which are announced on the [mailing list](https://mail.python.org/mailman/listinfo/numpy-discussion). You are very welcome to join. If you are new to contributing to open source, we also highly recommend reading [this guide](https://opensource.guide/how-to-contribute/). + +Our community aspires to treat everyone equally and to value all contributions. We have a [Code of Conduct](/code-of-conduct) to foster an open and welcoming environment. + +### Writing code + +Programmers, this [guide](https://numpy.org/devdocs/dev/index.html#development-process-summary) explains how to contribute to the NumPy codebase.
Check out also our [YouTube channel](https://www.youtube.com/playlist?list=PLCK6zCrcN3GXBUUzDr9L4__LnXZVtaIzS) for additional advice. + + +### Reviewing pull requests +The project has more than 250 open pull requests -- meaning many potential improvements and many open-source contributors waiting for feedback. If you're a developer who knows NumPy, you can help even if you're not familiar with the codebase. You can: +* summarize a long-running discussion +* triage documentation PRs +* test proposed changes + + +### Developing educational materials + +NumPy's [User Guide](https://numpy.org/devdocs) is undergoing rehabilitation. We're in need of new tutorials, how-to's, and deep-dive explanations, and the site needs restructuring. Opportunities aren't limited to writers. We'd also welcome worked examples, notebooks, and videos. [NEP 44 — Restructuring the NumPyDocumentation](https://numpy.org/neps/nep-0044-restructuring-numpy-docs.html) lays out our ideas -- and you may have others. + + +### Issue triaging + +The [NumPy issue tracker](https://github.com/numpy/numpy/issues) has a _lot_ of open issues. Some are no longer valid, some should be prioritized, and some would make good issues for new contributors. You can: + +* check if older bugs are still present +* find duplicate issues and link related ones +* add good self-contained reproducers to issues +* label issues correctly (this requires triage rights -- just ask) + +Please just dive in. + + +### Website development + +We've just revamped our website, but we're far from done. If you love web development, these [issues](https://github.com/numpy/numpy.org/issues?q=is%3Aissue+is%3Aopen+label%3Adesign) list some of our unmet needs -- and feel free to share your own ideas. + + +### Graphic design + +We can barely begin to list the contributions a graphic designer can make here. Our docs are parched for illustration; our growing website craves images -- opportunities abound. + + +### Translating website content + +We plan multiple translations of [numpy.org](https://numpy.org) to make NumPy accessible to users in their native language. Volunteer translators are at the heart of this effort. See [here](https://numpy.org/neps/nep-0028-website-redesign.html#translation-multilingual-i18n) for background; comment on [this GitHub issue](https://github.com/numpy/numpy.org/issues/55) to sign up. + + +### Community coordination and outreach + +Through community contact we share our work more widely and learn where we're falling short. We're eager to get more people involved in efforts like our [Twitter](https://twitter.com/numpy_team) account, organizing NumPy [code sprints](https://scisprints.github.io/), a newsletter, and perhaps a blog. + +### Fundraising + +NumPy was all-volunteer for many years, but as its importance grew it became clear that to ensure stability and growth we'd need financial support. [This SciPy'19 talk](https://www.youtube.com/watch?v=dBTJD_FDVjU) explains how much difference that support has made. Like all the nonprofit world, we're constantly searching for grants, sponsorships, and other kinds of support. We have a number of ideas and of course we welcome more. Fundraising is a scarce skill here -- we'd appreciate your help. diff --git a/content/ar/gethelp.md b/content/ar/gethelp.md new file mode 100644 index 0000000000..da3bff67a5 --- /dev/null +++ b/content/ar/gethelp.md @@ -0,0 +1,20 @@ +--- +title: Get Help +sidebar: false +--- + +**Development issues:** For NumPy development-related matters (e.g., bug reports), please see [Community](/community). + +**User questions:** The best way to get help is to post your question to a site like [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) or [Reddit](https://www.reddit.com/r/Numpy/). We wish we could keep an eye on these sites, or answer questions directly, but the volume is a little overwhelming! + +### [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) + +A forum for asking usage questions, e.g. "How do I do X in NumPy?”. Please [use the `#numpy` tag](https://stackoverflow.com/help/tagging) + +*** + +### [Reddit](https://www.reddit.com/r/Numpy/) + +Another forum for usage questions. + +*** diff --git a/content/ar/history.md b/content/ar/history.md new file mode 100644 index 0000000000..aa669d375b --- /dev/null +++ b/content/ar/history.md @@ -0,0 +1,21 @@ +--- +title: History of NumPy +sidebar: false +--- + +NumPy is a foundational Python library that provides array data structures and related fast numerical routines. When started, the library had little funding, and was written mainly by graduate students—many of them without computer science education, and often without a blessing of their advisors. To even imagine that a small group of “rogue” student programmers could upend the already well-established ecosystem of research software—backed by millions in funding and many hundreds of highly qualified engineers — was preposterous. Yet, the philosophical motivations behind a fully open tool stack, in combination with the excited, friendly community with a singular focus, have proven auspicious in the long run. Nowadays, NumPy is relied upon by scientists, engineers, and many other professionals around the world. For example, the published scripts used in the analysis of gravitational waves import NumPy, and the M87 black hole imaging project directly cites NumPy. + +For the in-depth account on milestones in the development of NumPy and related libraries please see [arxiv.org](https://arxiv.org/abs/1907.10121). + +If you’d like to obtain a copy of the original Numeric and Numarray libraries, follow the links below: + +[Download Page for *Numeric*](https://sourceforge.net/projects/numpy/files/Old%20Numeric/)* + +[Download Page for *Numarray*](https://sourceforge.net/projects/numpy/files/Old%20Numarray/)* + +*Please note that these older array packages are no longer maintained, and users are strongly advised to use NumPy for any array-related purposes or refactor any pre-existing code to utilize the NumPy library. + +### Historic Documentation + +[Download *`Numeric'* Manual](static/numeric-manual.pdf) + diff --git a/content/ar/install.md b/content/ar/install.md new file mode 100644 index 0000000000..f64a17841d --- /dev/null +++ b/content/ar/install.md @@ -0,0 +1,126 @@ +--- +title: Installing NumPy +sidebar: false +--- + +The only prerequisite for installing NumPy is Python itself. If you don't have Python yet and want the simplest way to get started, we recommend you use the [Anaconda Distribution](https://www.anaconda.com/download) - it includes Python, NumPy, and many other commonly used packages for scientific computing and data science. + +NumPy can be installed with `conda`, with `pip`, with a package manager on macOS and Linux, or [from source](https://numpy.org/devdocs/building). For more detailed instructions, consult our [Python and NumPy installation guide](#python-numpy-install-guide) below. + +**CONDA** + +If you use `conda`, you can install NumPy from the `defaults` or `conda-forge` channels: + +```bash +# Best practice, use an environment rather than install in the base env +conda create -n my-env +conda activate my-env +# If you want to install from conda-forge +conda config --env --add channels conda-forge +# The actual install command +conda install numpy +``` + +**PIP** + +If you use `pip`, you can install NumPy with: + +```bash +pip install numpy +``` +Also when using pip, it's good practice to use a virtual environment - see [Reproducible Installs](#reproducible-installs) below for why, and [this guide](https://dev.to/bowmanjd/python-tools-for-managing-virtual-environments-3bko#howto) for details on using virtual environments. + + + + +# Python and NumPy installation guide + +Installing and managing packages in Python is complicated, there are a number of alternative solutions for most tasks. This guide tries to give the reader a sense of the best (or most popular) solutions, and give clear recommendations. It focuses on users of Python, NumPy, and the PyData (or numerical computing) stack on common operating systems and hardware. + +## Recommendations + +We'll start with recommendations based on the user's experience level and operating system of interest. If you're in between "beginning" and "advanced", please go with "beginning" if you want to keep things simple, and with "advanced" if you want to work according to best practices that go a longer way in the future. + +### Beginning users + +On all of Windows, macOS, and Linux: + +- Install [Anaconda](https://www.anaconda.com/download) (it installs all packages you need and all other tools mentioned below). +- For writing and executing code, use notebooks in [JupyterLab](https://jupyterlab.readthedocs.io/en/stable/index.html) for exploratory and interactive computing, and [Spyder](https://www.spyder-ide.org/) or [Visual Studio Code](https://code.visualstudio.com/) for writing scripts and packages. +- Use [Anaconda Navigator](https://docs.anaconda.com/anaconda/navigator/) to manage your packages and start JupyterLab, Spyder, or Visual Studio Code. + + +### Advanced users + +#### Conda + +- Install [Miniforge](https://github.com/conda-forge/miniforge). +- Keep the `base` conda environment minimal, and use one or more [conda environments](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html) to install the package you need for the task or project you're working on. + +#### Alternative if you prefer pip/PyPI + +For users who know, from personal preference or reading about the main differences between conda and pip below, they prefer a pip/PyPI-based solution, we recommend: +- Install Python from [python.org](https://www.python.org/downloads/), [Homebrew](https://brew.sh/), or your Linux package manager. +- Use [Poetry](https://python-poetry.org/) as the most well-maintained tool that provides a dependency resolver and environment management capabilities in a similar fashion as conda does. + + +## Python package management + +Managing packages is a challenging problem, and, as a result, there are lots of tools. For web and general purpose Python development there's a whole [host of tools](https://packaging.python.org/guides/tool-recommendations/) complementary with pip. For high-performance computing (HPC), [Spack](https://github.com/spack/spack) is worth considering. For most NumPy users though, [conda](https://conda.io/en/latest/) and [pip](https://pip.pypa.io/en/stable/) are the two most popular tools. + + +### Pip & conda + +The two main tools that install Python packages are `pip` and `conda`. Their functionality partially overlaps (e.g. both can install `numpy`), however, they can also work together. We'll discuss the major differences between pip and conda here - this is important to understand if you want to manage packages effectively. + +The first difference is that conda is cross-language and it can install Python, while pip is installed for a particular Python on your system and installs other packages to that same Python install only. This also means conda can install non-Python libraries and tools you may need (e.g. compilers, CUDA, HDF5), while pip can't. + +The second difference is that pip installs from the Python Packaging Index (PyPI), while conda installs from its own channels (typically "defaults" or "conda-forge"). PyPI is the largest collection of packages by far, however, all popular packages are available for conda as well. + +The third difference is that conda is an integrated solution for managing packages, dependencies and environments, while with pip you may need another tool (there are many!) for dealing with environments or complex dependencies. + + + +### Reproducible installs + +As libraries get updated, results from running your code can change, or your code can break completely. It's important to be able to reconstruct the set of packages and versions you're using. Best practice is to: + +1. use a different environment per project you're working on, +2. record package names and versions using your package installer; each has its own metadata format for this: + - Conda: [conda environments and environment.yml](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html) + - Pip: [virtual environments](https://docs.python.org/3/tutorial/venv.html) and [requirements.txt](https://pip.readthedocs.io/en/latest/user_guide/#requirements-files) + - Poetry: [virtual environments and pyproject.toml](https://python-poetry.org/docs/basic-usage/) + + + +## NumPy packages & accelerated linear algebra libraries + +NumPy doesn't depend on any other Python packages, however, it does depend on an accelerated linear algebra library - typically [Intel MKL](https://software.intel.com/en-us/mkl) or [OpenBLAS](https://www.openblas.net/). Users don't have to worry about installing those (they're automatically included in all NumPy install methods). Power users may still want to know the details, because the used BLAS can affect performance, behavior and size on disk: + +- The NumPy wheels on PyPI, which is what pip installs, are built with OpenBLAS. The OpenBLAS libraries are included in the wheel. This makes the wheel larger, and if a user installs (for example) SciPy as well, they will now have two copies of OpenBLAS on disk. + +- In the conda defaults channel, NumPy is built against Intel MKL. MKL is a separate package that will be installed in the users' environment when they install NumPy. + +- In the conda-forge channel, NumPy is built against a dummy "BLAS" package. When a user installs NumPy from conda-forge, that BLAS package then gets installed together with the actual library - this defaults to OpenBLAS, but it can also be MKL (from the defaults channel), or even [BLIS](https://github.com/flame/blis) or reference BLAS. + +- The MKL package is a lot larger than OpenBLAS, it's about 700 MB on disk while OpenBLAS is about 30 MB. + +- MKL is typically a little faster and more robust than OpenBLAS. + +Besides install sizes, performance and robustness, there are two more things to consider: + +- Intel MKL is not open source. For normal use this is not a problem, but if a user needs to redistribute an application built with NumPy, this could be an issue. +- Both MKL and OpenBLAS will use multi-threading for function calls like `np.dot`, with the number of threads being determined by both a build-time option and an environment variable. Often all CPU cores will be used. This is sometimes unexpected for users; NumPy itself doesn't auto-parallelize any function calls. It typically yields better performance, but can also be harmful - for example when using another level of parallelization with Dask, scikit-learn or multiprocessing. + + +## Troubleshooting + +If your installation fails with the message below, see [Troubleshooting ImportError](https://numpy.org/doc/stable/user/troubleshooting-importerror.html). + +``` +IMPORTANT: PLEASE READ THIS FOR ADVICE ON HOW TO SOLVE THIS ISSUE! + +Importing the numpy c-extensions failed. This error can happen for +different reasons, often due to issues with your setup. +``` + diff --git a/content/ar/learn.md b/content/ar/learn.md new file mode 100644 index 0000000000..4f9fa53ae3 --- /dev/null +++ b/content/ar/learn.md @@ -0,0 +1,76 @@ +--- +title: Learn +sidebar: false +--- + +For the **official NumPy documentation** visit [numpy.org/doc/stable](https://numpy.org/doc/stable). + +*** + +Below is a curated collection of educational resources, both for self-learning and teaching others, developed by NumPy contributors and vetted by the community. + +## Beginners + +There's a ton of information about NumPy out there. If you are just starting, we'd strongly recommend the following: + + **Tutorials** + +* [NumPy Quickstart Tutorial](https://numpy.org/devdocs/user/quickstart.html) +* [NumPy Tutorials](https://numpy.org/numpy-tutorials) A collection of tutorials and educational materials in the format of Jupyter Notebooks developed and maintained by the NumPy Documentation team. To submit your own content, visit the [numpy-tutorials repository on GitHub](https://github.com/numpy/numpy-tutorials). +* [NumPy Illustrated: The Visual Guide to NumPy *by Lev Maximov*](https://betterprogramming.pub/3b1d4976de1d?sk=57b908a77aa44075a49293fa1631dd9b) +* [Scientific Python Lectures](https://lectures.scientific-python.org/) Besides covering NumPy, these lectures offer a broader introduction to the scientific Python ecosystem. +* [NumPy: the absolute basics for beginners](https://numpy.org/devdocs/user/absolute_beginners.html) +* [NumPy tutorial *by Nicolas Rougier*](https://github.com/rougier/numpy-tutorial) +* [Stanford CS231 *by Justin Johnson*](http://cs231n.github.io/python-numpy-tutorial/) +* [NumPy User Guide](https://numpy.org/devdocs) + + **Books** + +* [Guide to NumPy *by Travis E. Oliphant*](https://web.mit.edu/dvp/Public/numpybook.pdf) This is a free version 1 from 2006. For the latest copy (2015) see [here](https://www.barnesandnoble.com/w/guide-to-numpy-travis-e-oliphant-phd/1144670472). +* [From Python to NumPy *by Nicolas P. Rougier*](https://www.labri.fr/perso/nrougier/from-python-to-numpy/) +* [Elegant SciPy](https://www.amazon.com/Elegant-SciPy-Art-Scientific-Python/dp/1491922877) *by Juan Nunez-Iglesias, Stefan van der Walt, and Harriet Dashnow* + +You may also want to check out the [Goodreads list](https://www.goodreads.com/shelf/show/python-scipy) on the subject of "Python+SciPy." Most books there are about the "SciPy ecosystem," which has NumPy at its core. + + **Videos** + +* [Introduction to Numerical Computing with NumPy](http://youtu.be/ZB7BZMhfPgk) *by Alex Chabot-Leclerc* + +*** + +## Advanced + +Try these advanced resources for a better understanding of NumPy concepts like advanced indexing, splitting, stacking, linear algebra, and more. + + **Tutorials** + +* [100 NumPy Exercises](http://www.labri.fr/perso/nrougier/teaching/numpy.100/index.html) *by Nicolas P. Rougier* +* [An Introduction to NumPy and Scipy](https://engineering.ucsb.edu/~shell/che210d/numpy.pdf) *by M. Scott Shell* +* [Numpy Medkits](http://mentat.za.net/numpy/numpy_advanced_slides/) *by Stéfan van der Walt* +* [NumPy Tutorials](https://numpy.org/numpy-tutorials) A collection of tutorials and educational materials in the format of Jupyter Notebooks developed and maintained by the NumPy Documentation team. To submit your own content, visit the [numpy-tutorials repository on GitHub](https://github.com/numpy/numpy-tutorials). + + **Books** + +* [Python Data Science Handbook](https://www.amazon.com/Python-Data-Science-Handbook-Essential/dp/1098121228) *by Jake Vanderplas* +* [Python for Data Analysis](https://www.amazon.com/Python-Data-Analysis-Wrangling-Jupyter/dp/109810403X) *by Wes McKinney* +* [Numerical Python: Scientific Computing and Data Science Applications with Numpy, SciPy, and Matplotlib](https://www.amazon.com/Numerical-Python-Scientific-Applications-Matplotlib/dp/1484242459) *by Robert Johansson* + + **Videos** + +* [Advanced NumPy - broadcasting rules, strides, and advanced indexing](https://www.youtube.com/watch?v=cYugp9IN1-Q) *by Juan Nunez-Iglesias* + +*** + +## NumPy Talks + +* [The Future of NumPy Indexing](https://www.youtube.com/watch?v=o0EacbIbf58) *by Jaime Fernández* (2016) +* [Evolution of Array Computing in Python](https://www.youtube.com/watch?v=HVLPJnvInzM&t=10s) *by Ralf Gommers* (2019) +* [NumPy: what has changed and what is going to change?](https://www.youtube.com/watch?v=YFLVQFjRmPY) *by Matti Picus* (2019) +* [Inside NumPy](https://www.youtube.com/watch?v=dBTJD_FDVjU) *by Ralf Gommers, Sebastian Berg, Matti Picus, Tyler Reddy, Stefan van der Walt, Charles Harris* (2019) +* [Brief Review of Array Computing in Python](https://www.youtube.com/watch?v=f176j2g2eNc) *by Travis Oliphant* (2019) + +*** + +## Citing NumPy + +If NumPy has been significant in your research, and you would like to acknowledge the project in your academic publication, please see [this citation information](/citing-numpy). diff --git a/content/ar/news.md b/content/ar/news.md new file mode 100644 index 0000000000..793619c0d1 --- /dev/null +++ b/content/ar/news.md @@ -0,0 +1,279 @@ +--- +title: News +sidebar: false +newsHeader: "NumPy 2.0 release date: June 16" +date: 2024-05-23 +--- + +### NumPy 2.0 release date: June 16 + +_23 May, 2024_ -- We are excited to announce that NumPy 2.0 is planned to be released on June 16, 2024. This release has been over a year in the making, and is the first major release since 2006. Importantly, in addition to many new features and performance improvement, it contains **breaking changes** to the ABI as well as the Python and C APIs. It is likely that downstream packages and end user code needs to be adapted - if you can, please verify whether your code works with NumPy `2.0.0rc2`. **Please see the following for more details:** + +- The [NumPy 2.0 migration guide](https://numpy.org/devdocs/numpy_2_0_migration_guide.html) +- The [2.0.0 release notes](https://numpy.org/devdocs/release/2.0.0-notes.html) +- Announcement issue for status updates: [numpy#24300](https://github.com/numpy/numpy/issues/24300) + + +### NumFOCUS end of the year fundraiser +_Dec 19, 2023_ -- NumFOCUS has teamed up with PyCharm during their EOY campaign to offer a 30% discount on first-time PyCharm licenses. All year-one revenue from PyCharm purchases from now until December 23rd, 2023 will go directly to the NumFOCUS programs. + +Use unique URL that will allow to track purchases https://lp.jetbrains.com/support-data-science/ or a coupon code ISUPPORTDATASCIENCE  + +### NumPy 1.26.0 released + +_Sep 16, 2023_ -- [NumPy 1.26.0](https://numpy.org/doc/stable/release/1.26.0-notes.html) is now available. The highlights of the release are: + +* Python 3.12.0 support. +* Cython 3.0.0 compatibility. +* Use of the Meson build system +* Updated SIMD support +* f2py fixes, meson and bind(x) support +* Support for the updated Accelerate BLAS/LAPACK library + +The NumPy 1.26.0 release is a continuation of the 1.25.x series that marks the transition to the Meson build system and provision of support for Cython 3.0.0. A total of 20 people contributed to this release and 59 pull requests were merged. + +The Python versions supported by this release are 3.9-3.12. + +### numpy.org is now available in Japanese and Portuguese + +_Aug 2, 2023_ -- numpy.org is now available in 2 additional languages: Japanese and Portuguese. This wouldn’t be possible without our dedicated volunteers: + +_Portuguese:_ +* Melissa Weber Mendonça (melissawm) +* Ricardo Prins (ricardoprins) +* Getúlio Silva (getuliosilva) +* Julio Batista Silva (jbsilva) +* Alexandre de Siqueira (alexdesiqueira) +* Alexandre B A Villares (villares) +* Vini Salazar (vinisalazar) + +_Japanese:_ +* Atsushi Sakai (AtsushiSakai) +* KKunai +* Tom Kelly (TomKellyGenetics) +* Yuji Kanagawa (kngwyu) +* Tetsuo Koyama (tkoyama010) + +The work on the translation infrastructure is supported with funding from CZI. + +Looking ahead, we’d love to translate the website into more languages. If you’d like to help, please connect with the NumPy Translations Team on Slack: https://join.slack.com/t/numpy-team/shared_invite/zt-1gokbq56s-bvEpo10Ef7aHbVtVFeZv2w. (Look for the #translations channel.) We are also building a Translations Team who will be working on localizing documentation and educational content across the Scientific Python ecosystem. If this piqued your interest, join us on the Scientific Python Discord: https://discord.gg/khWtqY6RKr. (Look for the #translation channel.) + +### NumPy 1.25.0 released + +_Jun 17, 2023_ -- [NumPy 1.25.0](https://numpy.org/doc/stable/release/1.25.0-notes.html) is now available. The highlights of the release are: + +* Support for MUSL, there are now MUSL wheels. +* Support for the Fujitsu C/C++ compiler. +* Object arrays are now supported in einsum. +* Support for the inplace matrix multiplication (`@=`). + +The NumPy 1.25.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase the execution speed, and clarify the documentation. There has also been preparatory work for the future NumPy 2.0.0, resulting in a large number of new and expired deprecations. + +A total of 148 people contributed to this release and 530 pull requests were merged. + +The Python versions supported by this release are 3.9-3.11. + +### Fostering an Inclusive Culture: Call for Participation + +_May 10, 2023_ -- Fostering an Inclusive Culture: Call for Participation + +How can we be better when it comes to diversity and inclusion? Read the report and find out how to get involved [here](https://contributor-experience.org/docs/posts/dei-report/). + +### NumPy documentation team leadership transition + +_Jan 6, 2023_ –- Mukulika Pahari and Ross Barnowski are appointed as the new NumPy documentation team leads replacing Melissa Mendonça. We thank Melissa for all her contributions to the NumPy official documentation and educational materials, and Mukulika and Ross for stepping up. + +### NumPy 1.24.0 released + +_Dec 18, 2022_ -- [NumPy 1.24.0](https://numpy.org/doc/stable/release/1.24.0-notes.html) is now available. The highlights of the release are: + +* New "dtype" and "casting" keywords for stacking functions. +* New F2PY features and fixes. +* Many new deprecations, check them out. +* Many expired deprecations, + +The NumPy 1.24.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase execution speed, and clarify the documentation. There are a large number of new and expired deprecations due to changes in dtype promotion and cleanups. It is the work of 177 contributors spread over 444 pull requests. The supported Python versions are 3.8-3.11. + +### Numpy 1.23.0 released + +_Jun 22, 2022_ -- [NumPy 1.23.0](https://numpy.org/doc/stable/release/1.23.0-notes.html) is now available. The highlights of the release are: + +* Implementation of `loadtxt` in C, greatly improving its performance. +* Exposure of DLPack at the Python level for easy data exchange. +* Changes to the promotion and comparisons of structured dtypes. +* Improvements to f2py. + +The NumPy 1.23.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase the execution speed, clarify the documentation, and expire old deprecations. It is the work of 151 contributors spread over 494 pull requests. The Python versions supported by this release 3.8-3.10. Python 3.11 will be supported when it reaches the rc stage. + +### NumFOCUS DEI research study: call for participation + +_Apr 13, 2022_ -- NumPy is working with [NumFOCUS](http://numfocus.org/) on a [research project](https://numfocus.org/diversity-inclusion-disc/a-pivotal-time-in-numfocuss-project-aimed-dei-efforts?eType=EmailBlastContent&eId=f41a86c3-60d4-4cf9-86cf-58eb49dc968c) funded by the [Gordon & Betty Moore Foundation](https://www.moore.org/) to understand the barriers to participation that contributors, particularly those from historically underrepresented groups, face in the open-source software community. The research team would like to talk to new contributors, project developers and maintainers, and those who have contributed in the past about their experiences joining and contributing to NumPy. + +**Interested in sharing your experiences?** + +Please complete this brief [“Participant Interest” form](https://numfocus.typeform.com/to/WBWVJSqe) which contains additional information on the research goals, privacy, and confidentiality considerations. Your participation will be valuable to the growth and sustainability of diverse and inclusive open-source software communities. Accepted participants will participate in a 30-minute interview with a research team member. + +### Numpy 1.22.0 release + +_Dec 31, 2021_ -- [NumPy 1.22.0](https://numpy.org/doc/stable/release/1.22.0-notes.html) is now available. The highlights of the release are: + +* Type annotations of the main namespace are essentially complete. Upstream is a moving target, so there will likely be further improvements, but the major work is done. This is probably the most user visible enhancement in this release. +* A preliminary version of the proposed [array API Standard](https://data-apis.org/array-api/latest/) is provided (see [NEP 47](https://numpy.org/neps/nep-0047-array-api-standard.html)). This is a step in creating a standard collection of functions that can be used across libraries such as CuPy and JAX. +* NumPy now has a DLPack backend. DLPack provides a common interchange format for array (tensor) data. +* New methods for `quantile`, `percentile`, and related functions. The new methods provide a complete set of the methods commonly found in the literature. +* The universal functions have been refactored to implement most of [NEP 43](https://numpy.org/neps/nep-0043-extensible-ufuncs.html). This also unlocks the ability to experiment with the future DType API. +* A new configurable memory allocator for use by downstream projects. + +NumPy 1.22.0 is a big release featuring the work of 153 contributors spread over 609 pull requests. The Python versions supported by this release are 3.8-3.10. + +### Advancing an inclusive culture in the scientific Python ecosystem + +_August 31, 2021_ -- We are happy to announce the Chan Zuckerberg Initiative has [awarded a grant](https://chanzuckerberg.com/newsroom/czi-awards-16-million-for-foundational-open-source-software-tools-essential-to-biomedicine/) to support the onboarding, inclusion, and retention of people from historically marginalized groups on scientific Python projects, and to structurally improve the community dynamics for NumPy, SciPy, Matplotlib, and Pandas. + +As a part of [CZI's Essential Open Source Software for Science program](https://chanzuckerberg.com/eoss/), this [Diversity & Inclusion supplemental grant](https://cziscience.medium.com/advancing-diversity-and-inclusion-in-scientific-open-source-eaabe6a5488b) will support the creation of dedicated Contributor Experience Lead positions to identify, document, and implement practices to foster inclusive open-source communities. This project will be led by Melissa Mendonça (NumPy), with additional mentorship and guidance provided by Ralf Gommers (NumPy, SciPy), Hannah Aizenman and Thomas Caswell (Matplotlib), Matt Haberland (SciPy), and Joris Van den Bossche (Pandas). + +This is an ambitious project aiming to discover and implement activities that should structurally improve the community dynamics of our projects. By establishing these new cross-project roles, we hope to introduce a new collaboration model to the Scientific Python communities, allowing community-building work within the ecosystem to be done more efficiently and with greater outcomes. We also expect to develop a clearer picture of what works and what doesn't in our projects to engage and retain new contributors, especially from historically underrepresented groups. Finally, we plan on producing detailed reports on the actions executed, explaining how they have impacted our projects in terms of representation and interaction with our communities. + +The two-year project is expected to start by November 2021, and we are excited to see the results from this work! [You can read the full proposal here](https://figshare.com/articles/online_resource/Advancing_an_inclusive_culture_in_the_scientific_Python_ecosystem/16548063). + +### 2021 NumPy survey + +_July 12, 2021_ -- At NumPy, we believe in the power of our community. 1,236 NumPy users from 75 countries participated in our inaugural survey last year. The survey findings gave us a very good understanding of what we should focus on for the next 12 months. + +It’s time for another survey, and we are counting on you once again. It will take about 15 minutes of your time. Besides English, the survey questionnaire is available in 8 additional languages: Bangla, French, Hindi, Japanese, Mandarin, Portuguese, Russian, and Spanish. + +Follow the link to get started: https://berkeley.qualtrics.com/jfe/form/SV_aaOONjgcBXDSl4q. + + +### Numpy 1.21.0 release + +_Jun 23, 2021_ -- [NumPy 1.21.0](https://numpy.org/doc/stable/release/1.21.0-notes.html) is now available. The highlights of the release are: + +- continued SIMD work covering more functions and platforms, +- initial work on the new dtype infrastructure and casting, +- universal2 wheels for Python 3.8 and Python 3.9 on Mac, +- improved documentation, +- improved annotations, +- new `PCG64DXSM` bitgenerator for random numbers. + +This NumPy release is the result of 581 merged pull requests contributed by 175 people. The Python versions supported for this release are 3.7-3.9, support for Python 3.10 will be added after Python 3.10 is released. + + +### 2020 NumPy survey results + +_Jun 22, 2021_ -- In 2020, the NumPy survey team in partnership with students and faculty from the University of Michigan and the University of Maryland conducted the first official NumPy community survey. Find the survey results here: https://numpy.org/user-survey-2020/. + + +### Numpy 1.20.0 release + +_Jan 30, 2021_ -- [NumPy 1.20.0](https://numpy.org/doc/stable/release/1.20.0-notes.html) is now available. This is the largest NumPy release to date, thanks to 180+ contributors. The two most exciting new features are: +- Type annotations for large parts of NumPy, and a new `numpy.typing` submodule containing `ArrayLike` and `DtypeLike` aliases that users and downstream libraries can use when adding type annotations in their own code. +- Multi-platform SIMD compiler optimizations, with support for x86 (SSE, AVX), ARM64 (Neon), and PowerPC (VSX) instructions. This yielded significant performance improvements for many functions (examples: [sin/cos](https://github.com/numpy/numpy/pull/17587), [einsum](https://github.com/numpy/numpy/pull/18194)). + +### Diversity in the NumPy project + +_Sep 20, 2020_ -- We wrote a [statement on the state of, and discussion on social media around, diversity and inclusion in the NumPy project](/diversity_sep2020). + + +### First official NumPy paper published in Nature! + +_Sep 16, 2020_ -- We are pleased to announce the publication of [the first official paper on NumPy](https://www.nature.com/articles/s41586-020-2649-2) as a review article in Nature. This comes 14 years after the release of NumPy 1.0. The paper covers applications and fundamental concepts of array programming, the rich scientific Python ecosystem built on top of NumPy, and the recently added array protocols to facilitate interoperability with external array and tensor libraries like CuPy, Dask, and JAX. + + +### Python 3.9 is coming, when will NumPy release binary wheels? + +_Sept 14, 2020_ -- Python 3.9 will be released in a few weeks. If you are an early adopter of Python versions, you may be dissapointed to find that NumPy (and other binary packages like SciPy) will not have binary wheels ready on the day of the release. It is a major effort to adapt the build infrastructure to a new Python version and it typically takes a few weeks for the packages to appear on PyPI and conda-forge. In preparation for this event, please make sure to +- update your `pip` to version 20.1 at least to support `manylinux2010` and `manylinux2014` +- use [`--only-binary=numpy`](https://pip.pypa.io/en/stable/reference/pip_install/#cmdoption-only-binary) or `--only-binary=:all:` to prevent `pip` from trying to build from source. + + +### Numpy 1.19.2 release + +_Sep 10, 2020_ -- [NumPy 1.19.2](https://numpy.org/devdocs/release/1.19.2-notes.html) is now available. This latest release in the 1.19 series fixes several bugs, prepares for the [upcoming Cython 3.x release](http://docs.cython.org/en/latest/src/changes.html) and pins setuptools to keep distutils working while upstream modifications are ongoing. The aarch64 wheels are built with the latest manylinux2014 release that fixes the problem of differing page sizes used by different linux distros. + +### The inaugural NumPy survey is live! + +_Jul 2, 2020_ -- This survey is meant to guide and set priorities for decision-making about the development of NumPy as software and as a community. The survey is available in 8 additional languages besides English: Bangla, Hindi, Japanese, Mandarin, Portuguese, Russian, Spanish and French. + +Please help us make NumPy better and take the survey [here](https://umdsurvey.umd.edu/jfe/form/SV_8bJrXjbhXf7saAl). + + +### NumPy has a new logo! + +_Jun 24, 2020_ -- NumPy now has a new logo: + +NumPy logo + +The logo is a modern take on the old one, with a cleaner design. Thanks to Isabela Presedo-Floyd for designing the new logo, as well as to Travis Vaught for the old logo that served us well for 15+ years. + + +### NumPy 1.19.0 release + +_Jun 20, 2020_ -- NumPy 1.19.0 is now available. This is the first release without Python 2 support, hence it was a "clean-up release". The minimum supported Python version is now Python 3.6. An important new feature is that the random number generation infrastructure that was introduced in NumPy 1.17.0 is now accessible from Cython. + + +### Season of Docs acceptance + +_May 11, 2020_ -- NumPy has been accepted as one of the mentor organizations for the Google Season of Docs program. We are excited about the opportunity to work with a technical writer to improve NumPy's documentation once again! For more details, please see [the official Season of Docs site](https://developers.google.com/season-of-docs/) and our [ideas page](https://github.com/numpy/numpy/wiki/Google-Season-of-Docs-2020-Project-Ideas). + + +### NumPy 1.18.0 release + +_Dec 22, 2019_ -- NumPy 1.18.0 is now available. After the major changes in 1.17.0, this is a consolidation release. It is the last minor release that will support Python 3.5. Highlights of the release includes the addition of basic infrastructure for linking with 64-bit BLAS and LAPACK libraries, and a new C-API for `numpy.random`. + +Please see the [release notes](https://github.com/numpy/numpy/releases/tag/v1.18.0) for more details. + + +### NumPy receives a grant from the Chan Zuckerberg Initiative + +_Nov 15, 2019_ -- We are pleased to announce that NumPy and OpenBLAS, one of NumPy's key dependencies, have received a joint grant for $195,000 from the Chan Zuckerberg Initiative through their [Essential Open Source Software for Science program](https://chanzuckerberg.com/eoss/) that supports software maintenance, growth, development, and community engagement for open source tools critical to science. + +This grant will be used to ramp up the efforts in improving NumPy documentation, website redesign, and community development to better serve our large and rapidly growing user base, and ensure the long-term sustainability of the project. While the OpenBLAS team will focus on addressing sets of key technical issues, in particular thread-safety, AVX-512, and thread-local storage (TLS) issues, as well as algorithmic improvements in ReLAPACK (Recursive LAPACK) on which OpenBLAS depends. + +More details on our proposed initiatives and deliverables can be found in the [full grant proposal](https://figshare.com/articles/Proposal_NumPy_OpenBLAS_for_Chan_Zuckerberg_Initiative_EOSS_2019_round_1/10302167). The work is scheduled to start on Dec 1st, 2019 and continue for the next 12 months. + + + + +## Releases + +Here is a list of NumPy releases, with links to release notes. Bugfix releases (only the `z` changes in the `x.y.z` version number) have no new features; minor releases (the `y` increases) do. + +- NumPy 1.26.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.4)) -- _5 Feb 2024_. +- NumPy 1.26.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.3)) -- _2 Jan 2024_. +- NumPy 1.26.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.2)) -- _12 Nov 2023_. +- NumPy 1.26.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.1)) -- _14 Oct 2023_. +- NumPy 1.26.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.0)) -- _16 Sep 2023_. +- NumPy 1.25.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.25.2)) -- _31 Jul 2023_. +- NumPy 1.25.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.25.1)) -- _8 Jul 2023_. +- NumPy 1.24.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.4)) -- _26 Jun 2023_. +- NumPy 1.25.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.25.0)) -- _17 Jun 2023_. +- NumPy 1.24.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.3)) -- _22 Apr 2023_. +- NumPy 1.24.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.2)) -- _5 Feb 2023_. +- NumPy 1.24.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.1)) -- _26 Dec 2022_. +- NumPy 1.24.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.0)) -- _18 Dec 2022_. +- NumPy 1.23.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.5)) -- _19 Nov 2022_. +- NumPy 1.23.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.4)) -- _12 Oct 2022_. +- NumPy 1.23.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.3)) -- _9 Sep 2022_. +- NumPy 1.23.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.2)) -- _14 Aug 2022_. +- NumPy 1.23.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.1)) -- _8 Jul 2022_. +- NumPy 1.23.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.0)) -- _22 Jun 2022_. +- NumPy 1.22.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.4)) -- _20 May 2022_. +- NumPy 1.21.6 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.21.6)) -- _12 Apr 2022_. +- NumPy 1.22.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.3)) -- _7 Mar 2022_. +- NumPy 1.22.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.2)) -- _3 Feb 2022_. +- NumPy 1.22.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.1)) -- _14 Jan 2022_. +- NumPy 1.22.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.0)) -- _31 Dec 2021_. +- NumPy 1.21.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.21.5)) -- _19 Dec 2021_. +- NumPy 1.21.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.21.0)) -- _22 Jun 2021_. +- NumPy 1.20.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.20.3)) -- _10 May 2021_. +- NumPy 1.20.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.20.0)) -- _30 Jan 2021_. +- NumPy 1.19.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.19.5)) -- _5 Jan 2021_. +- NumPy 1.19.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.19.0)) -- _20 Jun 2020_. +- NumPy 1.18.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.18.4)) -- _3 May 2020_. +- NumPy 1.17.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.17.5)) -- _1 Jan 2020_. +- NumPy 1.18.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.18.0)) -- _22 Dec 2019_. +- NumPy 1.17.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.17.0)) -- _26 Jul 2019_. +- NumPy 1.16.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.16.0)) -- _14 Jan 2019_. +- NumPy 1.15.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.15.0)) -- _23 Jul 2018_. +- NumPy 1.14.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.14.0)) -- _7 Jan 2018_. diff --git a/content/ar/press-kit.md b/content/ar/press-kit.md new file mode 100644 index 0000000000..2c8970bb29 --- /dev/null +++ b/content/ar/press-kit.md @@ -0,0 +1,8 @@ +--- +title: Press kit +sidebar: false +--- + +We would like to make it easy for you to include the NumPy project identity in your next academic paper, course materials, or presentation. + +You will find several high-resolution versions of the NumPy logo [here](https://github.com/numpy/numpy/tree/main/branding/logo). Note that by using the numpy.org resources, you accept the [NumPy Code of Conduct](/code-of-conduct). diff --git a/content/ar/privacy.md b/content/ar/privacy.md new file mode 100644 index 0000000000..6064e4c4f1 --- /dev/null +++ b/content/ar/privacy.md @@ -0,0 +1,8 @@ +--- +title: Privacy Policy +sidebar: false +--- + +**numpy.org** is operated by [NumFOCUS, Inc.](https://numfocus.org), the fiscal sponsor of the NumPy project. For the Privacy Policy of this website please refer to https://numfocus.org/privacy-policy. + +If you have any questions about the policy or NumFOCUS’s data collection, use, and disclosure practices, please contact the NumFOCUS staff at privacy@numfocus.org. diff --git a/content/ar/report-handling-manual.md b/content/ar/report-handling-manual.md new file mode 100644 index 0000000000..5586668cba --- /dev/null +++ b/content/ar/report-handling-manual.md @@ -0,0 +1,95 @@ +--- +title: NumPy Code of Conduct - How to follow up on a report +sidebar: false +--- + +This is the manual followed by NumPy’s Code of Conduct Committee. It’s used when we respond to an issue to make sure we’re consistent and fair. + +Enforcing the [Code of Conduct](/code-of-conduct) impacts our community today and for the future. It’s an action that we do not take lightly. When reviewing enforcement measures, the Code of Conduct Committee will keep the following values and guidelines in mind: + +* Act in a personal manner rather than impersonal. The Committee can engage the parties to understand the situation while respecting the privacy and any necessary confidentiality of reporters. However, sometimes it is necessary to communicate with one or more individuals directly: the Committee’s goal is to improve the health of our community rather than only produce a formal decision. +* Emphasize empathy for individuals rather than judging behavior, avoiding binary labels of “good” and “bad/evil”. Overt, clear-cut aggression and harassment exist, and we will address them firmly. But many scenarios that can prove challenging to resolve are those where normal disagreements devolve into unhelpful or harmful behavior from multiple parties. Understanding the full context and finding a path that re-engages all is hard, but ultimately the most productive for our community. +* We understand that email is a difficult medium and can be isolating. Receiving criticism over email, without personal contact, can be particularly painful. This makes it especially important to keep an atmosphere of open-minded respect for the views of others. It also means that we must be transparent in our actions, and that we will do everything in our power to make sure that all our members are treated fairly and with sympathy. +* Discrimination can be subtle and it can be unconscious. It can show itself as unfairness and hostility in otherwise ordinary interactions. We know that this does occur, and we will take care to look out for it. We would very much like to hear from you if you feel you have been treated unfairly, and we will use these procedures to make sure that your complaint is heard and addressed. +* Help increase engagement in good discussion practice: try to identify where discussion may have broken down, and provide actionable information, pointers, and resources that can lead to positive change on these points. +* Be mindful of the needs of new members: provide them with explicit support and consideration, with the aim of increasing participation from underrepresented groups in particular. +* Individuals come from different cultural backgrounds and native languages. Try to identify any honest misunderstandings caused by a non-native speaker and help them understand the issue and what they can change to avoid causing offence. Complex discussion in a foreign language can be very intimidating, and we want to grow our diversity also across nationalities and cultures. + + +## Mediation + +Voluntary informal mediation is a tool at our disposal. In contexts such as when two or more parties have all escalated to the point of inappropriate behavior (something sadly common in human conflict), it may be useful to facilitate a mediation process. This is only an example: the Committee can consider mediation in any case, mindful that the process is meant to be strictly voluntary and no party can be pressured to participate. If the Committee suggests mediation, it should: + +* Find a candidate who can serve as a mediator. +* Obtain the agreement of the reporter(s). The reporter(s) have complete freedom to decline the mediation idea or to propose an alternate mediator. +* Obtain the agreement of the reported person(s). +* Settle on the mediator: while parties can propose a different mediator than the suggested candidate, only if a common agreement is reached on all terms can the process move forward. +* Establish a timeline for mediation to complete, ideally within two weeks. + +The mediator will engage with all the parties and seek a resolution that is satisfactory to all. Upon completion, the mediator will provide a report (vetted by all parties to the process) to the Committee, with recommendations on further steps. The Committee will then evaluate these results (whether a satisfactory resolution was achieved or not) and decide on any additional action deemed necessary. + + +## How the Committee will respond to reports + +When the Committee (or a Committee member) receives a report, they will first determine whether the report is about a clear and severe breach (as defined below). If so, immediate action needs to be taken in addition to the regular report handling process. + + +## Clear and severe breach actions + +We know that it is painfully common for internet communication to start at or devolve into obvious and flagrant abuse. We will deal quickly with clear and severe breaches like personal threats, violent, sexist or racist language. + +When a member of the Code of Conduct Committee becomes aware of a clear and severe breach, they will do the following: + +* Immediately disconnect the originator from all NumPy communication channels. +* Reply to the reporter that their report has been received and that the originator has been disconnected. +* In every case, the moderator should make a reasonable effort to contact the originator, and tell them specifically how their language or actions qualify as a “clear and severe breach”. The moderator should also say that, if the originator believes this is unfair or they want to be reconnected to NumPy, they have the right to ask for a review, as below, by the Code of Conduct Committee. The moderator should copy this explanation to the Code of Conduct Committee. +* The Code of Conduct Committee will formally review and sign off on all cases where this mechanism has been applied to make sure it is not being used to control ordinary heated disagreement. + + +## Report handling + +When a report is sent to the Committee they will immediately reply to the reporter to confirm receipt. This reply must be sent within 72 hours, and the group should strive to respond much quicker than that. + +If a report doesn’t contain enough information, the Committee will obtain all relevant data before acting. The Committee is empowered to act on the Steering Council’s behalf in contacting any individuals involved to get a more complete account of events. + +The Committee will then review the incident and determine, to the best of their ability: + +* What happened. +* Whether this event constitutes a Code of Conduct violation. +* Who are the responsible party(ies). +* Whether this is an ongoing situation, and there is a threat to anyone’s physical safety. + +This information will be collected in writing, and whenever possible the group’s deliberations will be recorded and retained (i.e. chat transcripts, email discussions, recorded conference calls, summaries of voice conversations, etc). + +It is important to retain an archive of all activities of this Committee to ensure consistency in behavior and provide institutional memory for the project. To assist in this, the default channel of discussion for this Committee will be a private mailing list accessible to current and future members of the Committee as well as members of the Steering Council upon justified request. If the Committee finds the need to use off-list communications (e.g. phone calls for early/rapid response), it should in all cases summarize these back to the list so there’s a good record of the process. + +The Code of Conduct Committee should aim to have a resolution agreed upon within two weeks. In the event that a resolution can’t be determined in that time, the Committee will respond to the reporter(s) with an update and projected timeline for resolution. + + +## Resolutions + +The Committee must agree on a resolution by consensus. If the group cannot reach consensus and deadlocks for over a week, the group will turn the matter over to the Steering Council for resolution. + +Possible responses may include: + +* Taking no further action: + - if we determine no violations have occurred; + - if the matter has been resolved publicly while the Committee was considering responses. +* Coordinating voluntary mediation: if all involved parties agree, the Committee may facilitate a mediation process as detailed above. +* Remind publicly, and point out that some behavior/actions/language have been judged inappropriate and why in the current context, or can but hurtful to some people, requesting the community to self-adjust. +* A private reprimand from the Committee to the individual(s) involved. In this case, the group chair will deliver that reprimand to the individual(s) over email, cc’ing the group. +* A public reprimand. In this case, the Committee chair will deliver that reprimand in the same venue that the violation occurred, within the limits of practicality. E.g., the original mailing list for an email violation, but for a chat room discussion where the person/context may be gone, they can be reached by other means. The group may choose to publish this message elsewhere for documentation purposes. +* A request for a public or private apology, assuming the reporter agrees to this idea: they may at their discretion refuse further contact with the violator. The chair will deliver this request. The Committee may, if it chooses, attach “strings” to this request: for example, the group may ask a violator to apologize in order to retain one’s membership on a mailing list. +* A “mutually agreed upon hiatus” where the Committee asks the individual to temporarily refrain from community participation. If the individual chooses not to take a temporary break voluntarily, the Committee may issue a “mandatory cooling off period”. +* A permanent or temporary ban from some or all NumPy spaces (mailing lists, gitter.im, etc.). The group will maintain records of all such bans so that they may be reviewed in the future or otherwise maintained. + +Once a resolution is agreed upon, but before it is enacted, the Committee will contact the original reporter and any other affected parties and explain the proposed resolution. The Committee will ask if this resolution is acceptable, and must note feedback for the record. + +Finally, the Committee will make a report to the NumPy Steering Council (as well as the NumPy core team in the event of an ongoing resolution, such as a ban). + +The Committee will never publicly discuss the issue; all public statements will be made by the chair of the Code of Conduct Committee or the NumPy Steering Council. + + +## Conflicts of Interest + +In the event of any conflict of interest, a Committee member must immediately notify the other members, and recuse themselves if necessary. diff --git a/content/ar/tabcontents.yaml b/content/ar/tabcontents.yaml new file mode 100644 index 0000000000..d74cba9bce --- /dev/null +++ b/content/ar/tabcontents.yaml @@ -0,0 +1,373 @@ +params: + machinelearning: + paras: + - + para1: NumPy forms the basis of powerful machine learning libraries like [scikit-learn](https://scikit-learn.org) and [SciPy](https://www.scipy.org). As machine learning grows, so does the list of libraries built on NumPy. [TensorFlow’s](https://www.tensorflow.org) deep learning capabilities have broad applications — among them speech and image recognition, text-based applications, time-series analysis, and video detection. [PyTorch](https://pytorch.org), another deep learning library, is popular among researchers in computer vision and natural language processing. + para2: Statistical techniques called [ensemble](https://towardsdatascience.com/ensemble-methods-bagging-boosting-and-stacking-c9214a10a205) methods such as binning, bagging, stacking, and boosting are among the ML algorithms implemented by tools such as [XGBoost](https://xgboost.readthedocs.io/), [LightGBM](https://lightgbm.readthedocs.io/en/latest/), and [CatBoost](https://catboost.ai) — one of the fastest inference engines. [Yellowbrick](https://www.scikit-yb.org/en/latest/) and [Eli5](https://eli5.readthedocs.io/en/latest/) offer machine learning visualizations. + arraylibraries: + intro: + - + text: NumPy's API is the starting point when libraries are written to exploit innovative hardware, create specialized array types, or add capabilities beyond what NumPy provides. + headers: + - + text: Array Library + - + text: Capabilities & Application areas + libraries: + - + title: Dask + text: Distributed arrays and advanced parallelism for analytics, enabling performance at scale. + img: /images/content_images/arlib/dask.png + alttext: Dask + url: https://dask.org/ + - + title: CuPy + text: NumPy-compatible array library for GPU-accelerated computing with Python. + img: /images/content_images/arlib/cupy.png + alttext: CuPy + url: https://cupy.dev + - + title: JAX + text: "Composable transformations of NumPy programs: differentiate, vectorize, just-in-time compilation to GPU/TPU." + img: /images/content_images/arlib/jax_logo_250px.png + alttext: JAX + url: https://jax.readthedocs.io/ + - + title: Xarray + text: Labeled, indexed multi-dimensional arrays for advanced analytics and visualization. + img: /images/content_images/arlib/xarray.png + alttext: xarray + url: https://xarray.pydata.org/en/stable/index.html + - + title: Sparse + text: NumPy-compatible sparse array library that integrates with Dask and SciPy's sparse linear algebra. + img: /images/content_images/arlib/sparse.png + alttext: sparse + url: https://sparse.pydata.org/en/latest/ + - + title: PyTorch + text: Deep learning framework that accelerates the path from research prototyping to production deployment. + img: /images/content_images/arlib/pytorch-logo-dark.svg + alttext: PyTorch + url: https://pytorch.org/ + - + title: TensorFlow + text: An end-to-end platform for machine learning to easily build and deploy ML powered applications. + img: /images/content_images/arlib/tensorflow-logo.svg + alttext: TensorFlow + url: https://www.tensorflow.org + - + title: Arrow + text: A cross-language development platform for columnar in-memory data and analytics. + img: /images/content_images/arlib/arrow.png + alttext: arrow + url: https://arrow.apache.org/ + - + title: xtensor + text: Multi-dimensional arrays with broadcasting and lazy computing for numerical analysis. + img: /images/content_images/arlib/xtensor.png + alttext: xtensor + url: https://github.com/xtensor-stack/xtensor-python + - + title: Awkward Array + text: Manipulate JSON-like data with NumPy-like idioms. + img: /images/content_images/arlib/awkward.svg + alttext: awkward + url: https://awkward-array.org/ + - + title: uarray + text: Python backend system that decouples API from implementation; unumpy provides a NumPy API. + img: /images/content_images/arlib/uarray.png + alttext: uarray + url: https://uarray.org/en/latest/ + - + title: tensorly + text: Tensor learning, algebra and backends to seamlessly use NumPy, PyTorch, TensorFlow or CuPy. + img: /images/content_images/arlib/tensorly.png + alttext: tensorly + url: http://tensorly.org/stable/home.html + scientificdomains: + intro: + - + text: Nearly every scientist working in Python draws on the power of NumPy. + - + text: "NumPy brings the computational power of languages like C and Fortran to Python, a language much easier to learn and use. With this power comes simplicity: a solution in NumPy is often clear and elegant." + libraries: + - + title: Quantum Computing + alttext: A computer chip. + img: /images/content_images/sc_dom_img/quantum_computing.svg + links: + - + url: http://qutip.org + label: QuTiP + - + url: https://pyquil-docs.rigetti.com/en/stable + label: PyQuil + - + url: https://qiskit.org + label: Qiskit + - + url: https://pennylane.ai + label: PennyLane + - + title: Statistical Computing + alttext: A line graph with the line moving up. + img: /images/content_images/sc_dom_img/statistical_computing.svg + links: + - + url: https://pandas.pydata.org/ + label: Pandas + - + url: https://www.statsmodels.org/ + label: statsmodels + - + url: https://xarray.pydata.org/en/stable/ + label: Xarray + - + url: https://seaborn.pydata.org/ + label: Seaborn + - + title: Signal Processing + alttext: A bar chart with positive and negative values. + img: /images/content_images/sc_dom_img/signal_processing.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://pywavelets.readthedocs.io/ + label: PyWavelets + - + url: https://python-control.org/ + label: python-control + - + url: https://hyperspy.org/ + label: HyperSpy + - + title: Image Processing + alttext: An photograph of the mountains. + img: /images/content_images/sc_dom_img/image_processing.svg + links: + - + url: https://scikit-image.org/ + label: Scikit-image + - + url: https://opencv.org/ + label: OpenCV + - + url: https://mahotas.rtfd.io/ + label: Mahotas + - + title: Graphs and Networks + alttext: A simple graph. + img: /images/content_images/sc_dom_img/sd6.svg + links: + - + url: https://networkx.org/ + label: NetworkX + - + url: https://graph-tool.skewed.de/ + label: graph-tool + - + url: https://igraph.org/python/ + label: igraph + - + url: https://pygsp.rtfd.io/ + label: PyGSP + - + title: Astronomy + alttext: A telescope. + img: /images/content_images/sc_dom_img/astronomy_processes.svg + links: + - + url: https://www.astropy.org/ + label: AstroPy + - + url: https://sunpy.org/ + label: SunPy + - + url: https://spacepy.github.io/ + label: SpacePy + - + title: Cognitive Psychology + alttext: A human head with gears. + img: /images/content_images/sc_dom_img/cognitive_psychology.svg + links: + - + url: https://www.psychopy.org/ + label: PsychoPy + - + title: Bioinformatics + alttext: A strand of DNA. + img: /images/content_images/sc_dom_img/bioinformatics.svg + links: + - + url: https://biopython.org/ + label: BioPython + - + url: http://scikit-bio.org/ + label: Scikit-Bio + - + url: https://github.com/openvax/pyensembl + label: PyEnsembl + - + url: http://etetoolkit.org/ + label: ETE + - + title: Bayesian Inference + alttext: A graph with a bell-shaped curve. + img: /images/content_images/sc_dom_img/bayesian_inference.svg + links: + - + url: https://pystan.readthedocs.io/en/latest/ + label: PyStan + - + url: https://docs.pymc.io/ + label: PyMC3 + - + url: https://arviz-devs.github.io/arviz/ + label: ArviZ + - + url: https://emcee.readthedocs.io/ + label: emcee + - + title: Mathematical Analysis + alttext: Four mathematical symbols. + img: /images/content_images/sc_dom_img/mathematical_analysis.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://www.sympy.org/ + label: SymPy + - + url: https://www.cvxpy.org/ + label: cvxpy + - + url: https://fenicsproject.org/ + label: FEniCS + - + title: Chemistry + alttext: A test tube. + img: /images/content_images/sc_dom_img/chemistry.svg + links: + - + url: https://cantera.org/ + label: Cantera + - + url: https://www.mdanalysis.org/ + label: MDAnalysis + - + url: https://github.com/rdkit/rdkit + label: RDKit + - + url: https://www.pybamm.org/ + label: PyBaMM + - + title: Geoscience + alttext: The Earth. + img: /images/content_images/sc_dom_img/geoscience.svg + links: + - + url: https://pangeo.io/ + label: Pangeo + - + url: https://simpeg.xyz/ + label: Simpeg + - + url: https://github.com/obspy/obspy/wiki + label: ObsPy + - + url: https://www.fatiando.org/ + label: Fatiando a Terra + - + title: Geographic Processing + alttext: A map. + img: /images/content_images/sc_dom_img/GIS.svg + links: + - + url: https://shapely.readthedocs.io/ + label: Shapely + - + url: https://geopandas.org/ + label: GeoPandas + - + url: https://python-visualization.github.io/folium + label: Folium + - + title: Architecture & Engineering + alttext: A microprocessor development board. + img: /images/content_images/sc_dom_img/robotics.svg + links: + - + url: https://compas.dev/ + label: COMPAS + - + url: https://cityenergyanalyst.com/ + label: City Energy Analyst + - + url: https://nortikin.github.io/sverchok/ + label: Sverchok + datascience: + intro: "NumPy lies at the core of a rich ecosystem of data science libraries. A typical exploratory data science workflow might look like:" + image1: + - + img: /images/content_images/ds-landscape.png + alttext: Diagram of Python Libraries. The five catagories are 'Extract, Transform, Load', 'Data Exploration', 'Data Modeling', 'Data Evaluation' and 'Data Presentation'. + image2: + - + img: /images/content_images/data-science.png + alttext: Diagram of three overlapping circles. The circles are labeled 'Mathematics', 'Computer Science' and 'Domain Expertise'. In the middle of the diagram, which has the three circles overlapping it, is an area labeled 'Data Science'. + examples: + - + text: "Extract, Transform, Load: [Pandas](https://pandas.pydata.org), [Intake](https://intake.readthedocs.io), [PyJanitor](https://pyjanitor-devs.github.io/pyjanitor/)" + - + text: "Exploratory analysis: [Jupyter](https://jupyter.org), [Seaborn](https://seaborn.pydata.org), [Matplotlib](https://matplotlib.org), [Altair](https://altair-viz.github.io)" + - + text: "Model and evaluate: [scikit-learn](https://scikit-learn.org), [statsmodels](https://www.statsmodels.org/stable/index.html), [PyMC3](https://docs.pymc.io), [spaCy](https://spacy.io)" + - + text: "Report in a dashboard: [Dash](https://plotly.com/dash), [Panel](https://panel.holoviz.org), [Voila](https://voila.readthedocs.io/)" + content: + - + text: For high data volumes, [Dask](https://dask.org) and [Ray](https://ray.io/) are designed to scale. Stable deployments rely on data versioning ([DVC](https://dvc.org)), experiment tracking ([MLFlow](https://mlflow.org)), and workflow automation ([Airflow](https://airflow.apache.org), [Dagster](https://dagster.io) and [Prefect](https://www.prefect.io)). + visualization: + images: + - + url: https://www.fusioncharts.com/blog/best-python-data-visualization-libraries + img: /images/content_images/v_matplotlib.png + alttext: A streamplot made in matplotlib + - + url: https://github.com/yhat/ggpy + img: /images/content_images/v_ggpy.png + alttext: A scatter-plot graph made in ggpy + - + url: https://www.journaldev.com/19692/python-plotly-tutorial + img: /images/content_images/v_plotly.png + alttext: A box-plot made in plotly + - + url: https://altair-viz.github.io/gallery/streamgraph.html + img: /images/content_images/v_altair.png + alttext: A streamgraph made in altair + - + url: https://seaborn.pydata.org + img: /images/content_images/v_seaborn.png + alttext: A pairplot of two types of graph, a plot-graph and a frequency graph made in seaborn" + - + url: https://docs.pyvista.org/ + img: /images/content_images/v_pyvista.png + alttext: A 3D volume rendering made in PyVista. + - + url: https://napari.org + img: /images/content_images/v_napari.png + alttext: A multi-dimensionan image made in napari. + - + url: https://vispy.org/gallery/index.html + img: /images/content_images/v_vispy.png + alttext: A Voronoi diagram made in vispy. + content: + - + text: NumPy is an essential component in the burgeoning [Python visualization landscape](https://pyviz.org/overviews/index.html), which includes [Matplotlib](https://matplotlib.org), [Seaborn](https://seaborn.pydata.org), [Plotly](https://plot.ly), [Altair](https://altair-viz.github.io), [Bokeh](https://docs.bokeh.org/en/latest/), [Holoviz](https://holoviz.org), [Vispy](http://vispy.org), [Napari](https://napari.org/), and [PyVista](https://docs.pyvista.org/), to name a few. + - + text: NumPy's accelerated processing of large arrays allows researchers to visualize datasets far larger than native Python could handle. diff --git a/content/ar/teams.md b/content/ar/teams.md new file mode 100644 index 0000000000..91cf5ca399 --- /dev/null +++ b/content/ar/teams.md @@ -0,0 +1,22 @@ +--- +title: NumPy Teams +sidebar: false +--- + +We are an international team on a mission to support scientific and research communities worldwide by building quality, open-source software. [Join us]({{< relref "/contribute" >}})! + +{{< include-html "static/gallery/maintainers.html" >}} + +{{< include-html "static/gallery/docs-team.html" >}} + +{{< include-html "static/gallery/web-team.html" >}} + +{{< include-html "static/gallery/triage-team.html" >}} + +{{< include-html "static/gallery/survey-team.html" >}} + +{{< include-html "static/gallery/emeritus-maintainers.html" >}} + +# Governance + +For the list of the Steering Council members, please see [here](https://numpy.org/about/). diff --git a/content/ar/user-survey-2020.md b/content/ar/user-survey-2020.md new file mode 100644 index 0000000000..b4349bcb7d --- /dev/null +++ b/content/ar/user-survey-2020.md @@ -0,0 +1,18 @@ +--- +title: 2020 NUMPY COMMUNITY SURVEY +sidebar: false +--- + +In 2020, the NumPy survey team in partnership with students and faculty from a Master’s course in Survey Methodology jointly hosted by the University of Michigan and the University of Maryland conducted the first official NumPy community survey. Over 1,200 users from 75 countries participated to help us map out a landscape of the NumPy community and voiced their thoughts about the future of the project. + +{{< figure >}} +src = '/surveys/NumPy_usersurvey_2020_report_cover.png' alt = 'Cover page of the 2020 NumPy user survey report, titled "NumPy Community Survey 2020 - results"' width = '250' +{{< /figure >}} + +**[Download the report](/surveys/NumPy_usersurvey_2020_report.pdf)** to take a closer look at the survey findings. + + +For the highlights, check out **[this infographic](https://github.com/numpy/numpy-surveys/blob/master/images/2020NumPysurveyresults_community_infographic.pdf)**. + +Ready for a deep dive? Visit **https://numpy.org/user-survey-2020-details/**. + diff --git a/content/ar/user-surveys.md b/content/ar/user-surveys.md new file mode 100644 index 0000000000..89a2aa0460 --- /dev/null +++ b/content/ar/user-surveys.md @@ -0,0 +1,10 @@ +--- +title: NUMPY USER SURVEYS +sidebar: false +--- + +**2020** The NumPy survey team in partnership with students and faculty from the University of Michigan and the University of Maryland conducted the first official NumPy community survey. Find the survey results [here](https://numpy.org/user-survey-2020/). + +**2021** The collected data is currently being analyzed. + +If you have any questions or suggestions for the past or future surveys, please open an issue [here](https://github.com/numpy/numpy-surveys/issues). diff --git a/content/es/404.md b/content/es/404.md new file mode 100644 index 0000000000..b38d73b758 --- /dev/null +++ b/content/es/404.md @@ -0,0 +1,8 @@ +--- +title: 404 +sidebar: false +--- + +¡Oh, oh! Has llegado a un callejón sin salida. + +Si crees que algo debería estar aquí, puedes [reportar este problema](https://github.com/numpy/numpy.org/issues) en GitHub. diff --git a/content/es/_index.md b/content/es/_index.md new file mode 100644 index 0000000000..be88f9e642 --- /dev/null +++ b/content/es/_index.md @@ -0,0 +1,49 @@ +--- +title: null +--- + +{{< grid columns="1 2 2 3" >}} + +[[item]] +type = 'card' +title = 'Powerful N-dimensional arrays' +body = ''' +Fast and versatile, the NumPy vectorization, indexing, and broadcasting concepts are the de-facto standards of array computing today. +''' + +[[item]] +type = 'card' +title = 'Numerical computing tools' +body = ''' +NumPy offers comprehensive mathematical functions, random number generators, linear algebra routines, Fourier transforms, and more. +''' + +[[item]] +type = 'card' +title = 'Open source' +body = ''' +Distributed under a liberal [BSD license](https://github.com/numpy/numpy/blob/main/LICENSE.txt), NumPy is developed and maintained [publicly on GitHub](https://github.com/numpy/numpy) by a vibrant, responsive, and diverse [community](/community). +''' + +[[item]] +type = 'card' +title = 'Interoperable' +body = ''' +NumPy supports a wide range of hardware and computing platforms, and plays well with distributed, GPU, and sparse array libraries. +''' + +[[item]] +type = 'card' +title = 'Performant' +body = ''' +The core of NumPy is well-optimized C code. Enjoy the flexibility of Python with the speed of compiled code. +''' + +[[item]] +type = 'card' +title = 'Easy to use' +body = ''' +NumPy's high level syntax makes it accessible and productive for programmers from any background or experience level. +''' + +{{< /grid>}} diff --git a/content/es/about.md b/content/es/about.md new file mode 100644 index 0000000000..a6e42626dc --- /dev/null +++ b/content/es/about.md @@ -0,0 +1,90 @@ +--- +title: Quiénes Somos +sidebar: false +--- + +NumPy es un proyecto de código abierto cuyo objetivo es permitir la computación numérica en Python. Se creó en el 2005, a partir de los primeros trabajos de las bibliotecas Numeric y Numarray. NumPy siempre será software 100% código abierto, de uso libre para todos. Fue liberado bajo los términos liberales de la [licencia BSD modificada](https://github.com/numpy/numpy/blob/main/LICENSE.txt). + +NumPy se desarrolla de forma abierta en GitHub, mediante el consenso de las comunidades NumPy y científica de Python en general. Para más información sobre nuestro enfoque de gobernanza, consulte nuestro [Documento de Gobernanza](https://www.numpy.org/devdocs/dev/governance/index.html). + + +## Consejo Directivo + +El Consejo de Dirección de NumPy es el órgano de gobernanza del proyecto. Su papel es el garantizar, a través del trabajo con la comunidad NumPy en general y al servicio de la misma, el bienestar a largo plazo del proyecto, tanto desde el punto de vista técnico como de la comunidad. El Consejo Directivo de NumPy está formado actualmente por los siguientes miembros (en orden alfabético): + +- Sebastian Berg +- Ralf Gommers +- Charles Harris +- Stephan Hoyer +- Inessa Pawson +- Matti Picus +- Stéfan van der Walt +- Melissa Weber Mendonça +- Eric Wieser + +Eméritos: + +- Alex Griffing (2015-2017) +- Allan Haldane (2015-2021) +- Marten van Kerkwijk (2017-2019) +- Travis Oliphant (project founder, 2005-2012) +- Nathaniel Smith (2012-2021) +- Julian Taylor (2013-2021) +- Jaime Fernández del Río (2014-2021) +- Pauli Virtanen (2008-2021) + +Para contactar con el Consejo Directivo de NumPy, por favor envía un correo electrónico a numpy-team@googlegroups.com. + +## Equipos + +La dirección del proyecto NumPy trabaja activamente para diversificar las vías de contribución al proyecto.
NumPy cuenta actualmente con los siguientes equipos: + +- desarrollo +- documentación +- clasificación +- página web +- encuesta +- traducción +- mentores de sprints +- optimization +- financiación y subvenciones + +Visita la página de [Equipos]({{< relref "/teams" >}}) para más información. + +## Subcomité NumFOCUS + +- Charles Harris +- Ralf Gommers +- Melissa Weber Mendonça +- Sebastian Berg +- Miembro externo: Thomas Caswell + +## Patrocinadores + +NumPy recibe financiación directa de las siguientes fuentes: +{{< sponsors >}} + + +## Socios institucionales + +Los socios institucionales son organizaciones que apoyan al proyecto empleando a personas que contribuyen a NumPy como parte de su trabajo. Entre los actuales socios institucionales se encuentran: + +- UC Berkeley (Stéfan van der Walt) +- Quansight (Nathan Goldbaum, Ralf Gommers, Matti Picus, Melissa Weber Mendonça) +- NVIDIA (Sebastian Berg) + +{{< partners >}} + + +## Donar + +Si has encontrado NumPy útil en tu trabajo, investigación o empresa, por favor considera una donación al proyecto proporcional a tus recursos. ¡Cualquier cantidad ayuda! Todas las donaciones se utilizarán estrictamente para financiar el desarrollo del software de código abierto, la documentación y la comunidad de NumPy. + +NumPy es un proyecto patrocinado por NumFOCUS, una organización benéfica sin fines de lucro 501(c)(3) de Estados Unidos. NumFOCUS proporciona a NumPy apoyo fiscal, legal y administrativo para ayudar a garantizar el bienestar y la sostenibilidad del proyecto. Visita [numfocus.org](https://numfocus.org) para más información. + +Las donaciones a NumPy son gestionadas por [NumFOCUS](https://numfocus.org). Para los donantes de Estados Unidos, su donación es deducible de impuestos en la medida prevista por la ley. Al igual que con cualquier donación, debes consultar a tu asesor de impuestos sobre tu situación fiscal particular. + +El Consejo Directivo de NumPy tomará las decisiones sobre el mejor uso de los fondos recibidos. Las prioridades técnicas y de infraestructura están documentadas en la [Hoja de Ruta de NumPy](https://www.numpy.org/neps/index.html#roadmap). + +{{}} + diff --git a/content/es/arraycomputing.md b/content/es/arraycomputing.md new file mode 100644 index 0000000000..b473f9330f --- /dev/null +++ b/content/es/arraycomputing.md @@ -0,0 +1,21 @@ +--- +title: Cómputo Matricial +sidebar: false +--- + +*El cómputo matricial es la base del cómputo estadístico, matemático y científico en varias aplicaciones contemporáneas de ciencia de datos y aplicaciones de analíticas, como la visualización de datos, el procesamiento digital de señales, el procesamiento de imágenes, la bioinformática, el aprendizaje automático, la IA y entre otras.* + +La manipulación y transformación de datos a gran escala depende de una computación matricial eficiente y de alto rendimiento. El lenguaje de elección para la analítica de datos, el aprendizaje automático y el cómputo numérico productivo es **Python.** + +**Num**erical **Py**thon o NumPy es la biblioteca estándar de-facto del lenguaje de programación Python que soporta conjuntos y matrices multidimensionales de gran tamaño, y viene con una amplia colección de funciones matemáticas de alto nivel para operar sobre estos conjuntos. + +Tras el lanzamiento de NumPy en 2006, Pandas apareció en el panorama en 2008, y no fue hasta hace un par de años que aparecieron sucesivamente varias bibliotecas de cómputo matricial, poblando este escenario. Muchas de estas nuevas bibliotecas imitan las características y capacidades de NumPy, y contienen nuevos algoritmos y características orientadas a las aplicaciones de aprendizaje automático e inteligencia artificial. + +arraycl + +El **cómputo matricial** está basado en los **conjuntos** como estructura de datos. *Los conjuntos* se utilizan para organizar grandes cantidades de datos de manera que un conjunto de valores relacionados pueda ordenarse, buscarse, manipularse matemáticamente y transformarse con facilidad y rapidez. + +La computación matricial es *única* ya que implica operar sobre todo el conjunto de datos *al mismo tiempo*. Esto significa que cualquier operación de conjuntos se aplica a un conjunto completo de valores de una sola vez. Este enfoque vectorial proporciona velocidad y simplicidad, al permitir a los programadores codificar y trabajar sobre los datos agregados, sin tener que utilizar bucles de instrucciones escalares individuales. diff --git a/content/es/case-studies/blackhole-image.md b/content/es/case-studies/blackhole-image.md new file mode 100644 index 0000000000..0087c7b632 --- /dev/null +++ b/content/es/case-studies/blackhole-image.md @@ -0,0 +1,80 @@ +--- +title: "Caso de estudio: La primera imagen de un Agujero Negro" +sidebar: false +--- + +{{< figure src="/images/content_images/cs/blackhole.jpg" caption="**Agujero Negro M87**" alt="imagen de un agujero negro" attr="*(Créditos de la imagen: Colaboración del Telescopio de Horizonte de Sucesos)*" attrlink="https://www.jpl.nasa.gov/images/universe/20190410/blackhole20190410.jpg" >}} +src = '/images/content_images/cs/blackhole.jpg' title = 'Black Hole M87' alt = 'black hole image' attribution = '(Image Credits: Event Horizon Telescope Collaboration)' attributionlink = 'https://www.jpl.nasa.gov/images/universe/20190410/blackhole20190410.jpg' +{{< /figure >}} + +{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="Katie Bouman, *Assistant Professor, Computing & Mathematical Sciences, Caltech*" +> }} Imaging the M87 Black Hole is like trying to see something that is by definition impossible to see. +> +> {{< /blockquote >}} + +## Un telescopio del tamaño de la Tierra + +El [ telescopio de Horizonte de Sucesos (EHT) ](https://eventhorizontelescope.org), es un conjunto de ocho radiotelescopios terrestres que forman un telescopio computacional del tamaño de la tierra, estudiando al universo con una sensibilidad y resolución sin precedente. El enorme telescopio virtual, que utiliza una técnica llamada Interferometría de línea de base muy larga (VLBI), tiene una resolución angular de [20 microsegundos de arco][resolution] — ¡suficiente para leer un periódico en Nueva York desde un café en la acera en París! + +### Objetivos clave y resultados + +* **Una nueva vista del universo:** El trabajo preliminar de la innovadora imagen de EHT se había establecido 100 años antes, cuando [Sir Arthur Eddington][eddington] dio el primer apoyo observacional a la teoría de la relatividad general de Einstein. + +* **El agujero negro:** EHT se entrenó en un enorme agujero negro aproximadamente a 55 millones de años luz de la tierra, situada en el centro de la galaxia Messier 87 (M87) en el cúmulo de Virgo. Su masa es 6.5 mil millones de veces la del sol. Se había estudiado por [más de 100 años](https://www.jpl.nasa.gov/news/news.php?feature=7385), pero nunca antes se había observado un agujero negro. + +* **Comparando las observaciones con la teoría:** A partir de la teoría de la relatividad general de Einstein, los científicos esperaban encontrar una región similar a las sombras causadas por la flexión gravitacional y la captura de la luz. Los científicos pudieron utilizarla para medir la enorme masa del agujero negro. + +### Los desafíos + +* **Escala computacional** + + EHT plantea desafíos de procesamiento de datos masivos, incluyendo rápidas fluctuaciones de fase atmosféricas, amplio ancho de banda de grabación, y telescopios que son ampliamente disímiles y geográficamente dispersos. + +* **Demasiada información** + + Cada día el EHT genera más de 350 terabytes de observaciones, almacenados en discos duros llenos de helio. Reducir el volumen y complejidad de estos datos es enormemente difícil. + +* **Hacia lo desconocido** + + Cuando el objetivo es ver algo nunca antes visto, ¿cómo pueden los científicos estar seguros de que la imagen es correcta? + +{{< figure >}} +src = '/images/content_images/cs/dataprocessbh.png' title = 'EHT Data Processing Pipeline' alt = 'data pipeline' align = 'center' attribution = '(Diagram Credits: The Astrophysical Journal, Event Horizon Telescope Collaboration)' attributionlink = 'https://iopscience.iop.org/article/10.3847/2041-8213/ab0c57' +{{< figure src="/images/content_images/cs/bh_numpy_role.png" class="fig-center" alt="el rol de numpy" caption="**El rol de NumPy en la fotografía del Agujero Negro**" >}} + +## El Rol de NumPy + +¿Qué pasa si hay un problema con los datos? O tal vez un algoritmo depende demasiado de una suposición en particular. ¿Cambiará drásticamente la imagen si se cambia un solo parámetro? + +La colaboración del EHT respondió a estos desafíos haciendo que los equipos independientes evaluaran los datos, utilizando técnicas de reconstrucción de imágenes ya establecidas y de vanguardia. Cuando los resultados se mostraron consistentes, se combinaron para producir la primera imagen de su tipo de un agujero negro. + +Su trabajo ilustra el rol que desempeña el ecosistema científico de Python en el avance de la ciencia a través del análisis de datos colaborativos. + +{{< figure >}} +{{< figure src="/images/content_images/cs/numpy_bh_benefits.png" class="fig-center" alt="ventajas de numpy" caption="**Capacidades Clave utilizadas de NumPy**" >}} +{{< /figure >}} + +Por ejemplo, el paquete de Python [`eht-imaging`][ehtim] proporciona herramientas para simular y realizar reconstrucción de imágenes en datos VLBI. NumPy está en el núcleo del procesamiento de datos de matrices utilizados en este paquete, como se muestra a continuación en el gráfico parcial de dependencias de software. + +{{< figure >}} +src = '/images/content_images/cs/ehtim_numpy.png' alt = 'ehtim dependency map highlighting numpy' title = 'Software dependency chart of ehtim package highlighting NumPy' +{{< /figure >}} + +Además de NumPy, muchos otros paquetes, como [SciPy](https://www.scipy.org) y [Pandas](https://pandas.io), son parte del flujo de procesamiento de datos para fotografiar el agujero negro. Los formatos estándar de archivos astronómicos y transformaciones de tiempo/coordenadas fueron manejados por [Astropy][astropy], mientras que [Matplotlib][mpl] fue utilizado en la visualización de datos a través del flujo de análisis, incluyendo la generación de la imagen final del agujero negro. + +## Resumen + +La matriz n-dimensional eficiente y adaptable que es la característica central de NumPy permitió a los investigadores manipular grandes conjuntos de datos numéricos, proporcionando una base para la primera imagen de un agujero negro. Un momento histórico en la ciencia ofrece una impresionante evidencia visual de la teoría de Einstein. Este logro abarca no solo los avances tecnológicos sino también la colaboración internacional de más de 200 científicos y algunos de los mejores radio observatorios del mundo. Algoritmos innovadores y técnicas de procesamiento de datos, mejorando los modelos astronómicos existentes, ayudaron a desvelar un misterio del universo. + +{{< figure >}} +src = '/images/content_images/cs/numpy_bh_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} + +[resolution]: https://eventhorizontelescope.org/press-release-april-10-2019-astronomers-capture-first-image-black-hole + +[eddington]: https://en.wikipedia.org/wiki/Eddington_experiment + +[ehtim]: https://github.com/achael/eht-imaging + +[astropy]: https://www.astropy.org/ +[mpl]: https://matplotlib.org/ diff --git a/content/es/case-studies/cricket-analytics.md b/content/es/case-studies/cricket-analytics.md new file mode 100644 index 0000000000..d0be271f50 --- /dev/null +++ b/content/es/case-studies/cricket-analytics.md @@ -0,0 +1,72 @@ +--- +title: "Case Study: Cricket Analytics, the game changer!" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/ipl-stadium.png' title = 'IPLT20, the biggest Cricket Festival in India' alt = 'Indian Premier League Cricket cup and stadium' attribution = '(Image credits: IPLT20 (cup and logo) & Akash Yadav (stadium))' attributionlink = 'https://unsplash.com/@aksh1802' +{{< /figure >}} + +{{< blockquote cite="https://www.scoopwhoop.com/sports/ms-dhoni/" by="M S Dhoni, *International Cricket Player, ex-captain, Indian Team, plays for Chennai Super Kings in IPL*" +> }} You don't play for the crowd, you play for the country. +> +> {{< /blockquote >}} + +## About Cricket + +It would be an understatement to state that Indians love cricket. The game is played in just about every nook and cranny of India, rural or urban, popular with the young and the old alike, connecting billions in India unlike any other sport. Cricket enjoys lots of media attention. There is a significant amount of [money](https://www.statista.com/topics/4543/indian-premier-league-ipl/) and fame at stake. Over the last several years, technology has literally been a game changer. Audiences are spoilt for choice with streaming media, tournaments, affordable access to mobile based live cricket watching, and more. + +The Indian Premier League (IPL) is a professional Twenty20 cricket league, founded in 2008. It is one of the most attended cricketing events in the world, valued at [$6.7 billion](https://en.wikipedia.org/wiki/Indian_Premier_League) in 2019. + +Cricket is a game of numbers - the runs scored by a batsman, the wickets taken by a bowler, the matches won by a cricket team, the number of times a batsman responds in a certain way to a kind of bowling attack, etc. The capability to dig into cricketing numbers for both improving performance and studying the business opportunities, overall market, and economics of cricket via powerful analytics tools, powered by numerical computing software such as NumPy, is a big deal. Cricket analytics provides interesting insights into the game and predictive intelligence regarding game outcomes. + +Today, there are rich and almost infinite troves of cricket game records and statistics available, e.g., [ESPN cricinfo](https://stats.espncricinfo.com/ci/engine/stats/index.html) and [cricsheet](https://cricsheet.org). These and several such cricket databases have been used for [cricket analysis](https://www.researchgate.net/publication/336886516_Data_visualization_and_toss_related_analysis_of_IPL_teams_and_batsmen_performances) using the latest machine learning and predictive modelling algorithms. Media and entertainment platforms along with professional sports bodies associated with the game use technology and analytics for determining key metrics for improving match winning chances: + +* batting performance moving average, +* score forecasting, +* gaining insights into fitness and performance of a player against different opposition, +* player contribution to wins and losses for making strategic decisions on team composition + +{{< figure >}} +src = '/images/content_images/cs/cricket-pitch.png' title = 'Cricket Pitch, the focal point in the field' alt = 'A cricket pitch with bowler and batsmen' align = 'center' attribution = '(Image credit: Debarghya Das)' attributionlink = 'http://debarghyadas.com/files/IPLpaper.pdf' +{{< /figure >}} + +### Key Data Analytics Objectives + +* Sports data analytics are used not only in cricket but many [other sports](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx) for improving the overall team performance and maximizing winning chances. +* Real-time data analytics can help in gaining insights even during the game for changing tactics by the team and by associated businesses for economic benefits and growth. +* Besides historical analysis, predictive models are harnessed to determine the possible match outcomes that require significant number crunching and data science know-how, visualization tools and capability to include newer observations in the analysis. + +{{< figure >}} +src = '/images/content_images/cs/player-pose-estimator.png' alt = 'pose estimator' title = 'Cricket Pose Estimator' attribution = '(Image credit: connect.vin)' attributionlink = 'https://connect.vin/2019/05/ai-for-cricket-batsman-pose-analysis/' +{{< /figure >}} + +### The Challenges + +* **Data Cleaning and preprocessing** + + IPL has expanded cricket beyond the classic test match format to a much larger scale. The number of matches played every season across various formats has increased and so has the data, the algorithms, newer sports data analysis technologies and simulation models. Cricket data analysis requires field mapping, player tracking, ball tracking, player shot analysis, and several other aspects involved in how the ball is delivered, its angle, spin, velocity, and trajectory. All these factors together have increased the complexity of data cleaning and preprocessing. + +* **Dynamic Modeling** + + In cricket, just like any other sport, there can be a large number of variables related to tracking various numbers of players on the field, their attributes, the ball, and several possibilities of potential actions. The complexity of data analytics and modeling is directly proportional to the kind of predictive questions that are put forth during analysis and are highly dependent on data representation and the model. Things get even more challenging in terms of computation, data comparisons when dynamic cricket play predictions are sought such as what would have happened if the batsman had hit the ball at a different angle or velocity. + +* **Predictive Analytics Complexity** + + Much of the decision making in cricket is based on questions such as "how often does a batsman play a certain kind of shot if the ball delivery is of a particular type", or "how does a bowler change his line and length if the batsman responds to his delivery in a certain way". This kind of predictive analytics query requires highly granular dataset availability and the capability to synthesize data and create generative models that are highly accurate. + +## NumPy’s Role in Cricket Analytics + +Sports Analytics is a thriving field. Many researchers and companies [use NumPy](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx) and other PyData packages like Scikit-learn, SciPy, Matplotlib, and Jupyter, besides using the latest machine learning and AI techniques. NumPy has been used for various kinds of cricket related sporting analytics such as: + +* **Statistical Analysis:** NumPy's numerical capabilities help estimate the statistical significance of observational data or match events in the context of various player and game tactics, estimating the game outcome by comparison with a generative or static model. [Causal analysis](https://amplitude.com/blog/2017/01/19/causation-correlation) and [big data approaches](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4996805/) are used for tactical analysis. + +* **Data Visualization:** Data graphing and [visualization](https://towardsdatascience.com/advanced-sports-visualization-with-pandas-matplotlib-and-seaborn-9c16df80a81b) provide useful insights into relationship between various datasets. + +## Summary + +Sports Analytics is a game changer when it comes to how professional games are played, especially how strategic decision making happens, which until recently was primarily done based on “gut feeling" or adherence to past traditions. NumPy forms a solid foundation for a large set of Python packages which provide higher level functions related to data analytics, machine learning, and AI algorithms. These packages are widely deployed to gain real-time insights that help in decision making for game-changing outcomes, both on field as well as to draw inferences and drive business around the game of cricket. Finding out the hidden parameters, patterns, and attributes that lead to the outcome of a cricket match helps the stakeholders to take notice of game insights that are otherwise hidden in numbers and statistics. + +{{< figure >}} +src = '/images/content_images/cs/numpy_ca_benefits.png' alt = 'Diagram showing benefits of using NumPy for cricket analytics' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} diff --git a/content/es/case-studies/deeplabcut-dnn.md b/content/es/case-studies/deeplabcut-dnn.md new file mode 100644 index 0000000000..257905de91 --- /dev/null +++ b/content/es/case-studies/deeplabcut-dnn.md @@ -0,0 +1,102 @@ +--- +title: "Case Study: DeepLabCut 3D Pose Estimation" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/mice-hand.gif' title = 'Analyzing mice hand-movement using DeepLapCut' alt = 'micehandanim' attribution = '(Source: www.deeplabcut.org )' attributionlink = 'http://www.mousemotorlab.org/deeplabcut' +{{< /figure >}} + +{{< blockquote cite="https://news.harvard.edu/gazette/story/newsplus/harvard-researchers-awarded-czi-open-source-award/" by="Alexander Mathis, *Assistant Professor, École polytechnique fédérale de Lausanne* ([EPFL](https://www.epfl.ch/en/))" +> }} Open Source Software is accelerating Biomedicine. DeepLabCut enables automated video analysis of animal behavior using Deep Learning. +> +> {{< /blockquote >}} + +## About DeepLabCut + +[DeepLabCut](https://github.com/DeepLabCut/DeepLabCut) is an open source toolbox that empowers researchers at hundreds of institutions worldwide to track behaviour of laboratory animals, with very little training data, at human-level accuracy. With DeepLabCut technology, scientists can delve deeper into the scientific understanding of motor control and behavior across animal species and timescales. + +Several areas of research, including neuroscience, medicine, and biomechanics, use data from tracking animal movement. DeepLabCut helps in understanding what humans and other animals are doing by parsing actions that have been recorded on film. Using automation for laborious tasks of tagging and monitoring, along with deep neural network based data analysis, DeepLabCut makes scientific studies involving observing animals, such as primates, mice, fish, flies etc., much faster and more accurate. + +{{< figure >}} +src = '/images/content_images/cs/race-horse.gif' title = 'Colored dots track the positions of a racehorse’s body part' alt = 'horserideranim' attribution = '(Source: Mackenzie Mathis)' +{{< /figure >}} + +DeepLabCut's non-invasive behavioral tracking of animals by extracting the poses of animals is crucial for scientific pursuits in domains such as biomechanics, genetics, ethology & neuroscience. Measuring animal poses non-invasively from video - without markers - in dynamically changing backgrounds is computationally challenging, both technically as well as in terms of resource needs and training data required. + +DeepLabCut allows researchers to estimate the pose of the subject, efficiently enabling them to quantify the behavior through a Python based software toolkit. With DeepLabCut, researchers can identify distinct frames from videos, digitally label specific body parts in a few dozen frames with a tailored GUI, and then the deep learning based pose estimation architectures in DeepLabCut learn how to pick out those same features in the rest of the video and in other similar videos of animals. It works across species of animals, from common laboratory animals such as flies and mice to more unusual animals like [cheetahs][cheetah-movement]. + +DeepLabCut uses a principle called [transfer learning](https://arxiv.org/pdf/1909.11229), which greatly reduces the amount of training data required and speeds up the convergence of the training period. Depending on the needs, users can pick different network architectures that provide faster inference (e.g. MobileNetV2), which can also be combined with real-time experimental feedback. DeepLabCut originally used the feature detectors from a top-performing human pose estimation architecture, called [DeeperCut](https://arxiv.org/abs/1605.03170), which inspired the name. The package now has been significantly changed to include additional architectures, augmentation methods, and a full front-end user experience. Furthermore, to support large-scale biological experiments DeepLabCut provides active learning capabilities so that users can increase the training set over time to cover edge cases and make their pose estimation algorithm robust within the specific context. + +Recently, the [DeepLabCut model zoo](http://www.mousemotorlab.org/dlc-modelzoo) was introduced, which provides pre-trained models for various species and experimental conditions from facial analysis in primates to dog posture. This can be run for instance in the cloud without any labeling of new data, or neural network training, and no programming experience is necessary. + +### Key Goals and Results + +* **Automation of animal pose analysis for scientific studies:** + + The primary objective of DeepLabCut technology is to measure and track posture of animals in a diverse settings. This data can be used, for example, in neuroscience studies to understand how the brain controls movement, or to elucidate how animals socially interact. Researchers have observed a [tenfold performance boost](https://www.biorxiv.org/content/10.1101/457242v1) with DeepLabCut. Poses can be inferred offline at up to 1200 frames per second (FPS). + +* **Creation of an easy-to-use Python toolkit for pose estimation:** + + DeepLabCut wanted to share their animal pose-estimation technology in the form of an easy to use tool that can be adopted by researchers easily. So they have created a complete, easy-to-use Python toolbox with project management features as well. These enable not only automation of pose-estimation but also managing the project end-to-end by helping the DeepLabCut Toolkit user right from the dataset collection stage to creating shareable and reusable analysis pipelines. + + Their [toolkit][DLCToolkit] is now available as open source. + + A typical DeepLabCut Workflow includes: + + - creation and refining of training sets via active learning + - creation of tailored neural networks for specific animals and scenarios + - code for large-scale inference on videos + - draw inferences using integrated visualization tools + +{{< figure >}} +src = '/images/content_images/cs/deeplabcut-toolkit-steps.png' title = 'Pose estimation steps with DeepLabCut' alt = 'dlcsteps' align = 'center' attribution = '(Source: DeepLabCut)' attributionlink = 'https://twitter.com/DeepLabCut/status/1198046918284210176/photo/1' +{{< /figure >}} + +### The Challenges + +* **Speed** + + Fast processing of animal behavior videos in order to measure their behavior and at the same time make scientific experiments more efficient, accurate. Extracting detailed animal poses for laboratory experiments, without markers, in dynamically changing backgrounds, can be challenging, both technically as well as in terms of resource needs and training data required. Coming up with a tool that is easy to use without the need for skills such as computer vision expertise that enables scientists to do research in more real-world contexts, is a non-trivial problem to solve. + +* **Combinatorics** + + Combinatorics involves assembly and integration of movement of multiple limbs into individual animal behavior. Assembling keypoints and their connections into individual animal movements and linking them across time is a complex process that requires heavy-duty numerical analysis, especially in case of multi-animal movement tracking in experiment videos. + +* **Data Processing** + + Last but not the least, array manipulation - processing large stacks of arrays corresponding to various images, target tensors and keypoints is fairly challenging. + +{{< figure >}} +src = '/images/content_images/cs/pose-estimation.png' title = 'Pose estimation variety and complexity' alt = 'challengesfig' align = 'center' attribution = '(Source: Mackenzie Mathis)' attributionlink = 'https://www.biorxiv.org/content/10.1101/476531v1.full.pdf' +{{< /figure >}} + +## NumPy's Role in meeting Pose Estimation Challenges + +NumPy addresses DeepLabCut technology's core need of numerical computations at high speed for behavioural analytics. Besides NumPy, DeepLabCut employs various Python software that utilize NumPy at their core, such as [SciPy](https://www.scipy.org), [Pandas](https://pandas.pydata.org), [matplotlib](https://matplotlib.org), [Tensorpack](https://github.com/tensorpack/tensorpack), [imgaug](https://github.com/aleju/imgaug), [scikit-learn](https://scikit-learn.org/stable/), [scikit-image](https://scikit-image.org) and [Tensorflow](https://www.tensorflow.org). + +The following features of NumPy played a key role in addressing the image processing, combinatorics requirements and need for fast computation in DeepLabCut pose estimation algorithms: + +* Vectorization +* Masked Array Operations +* Linear Algebra +* Random Sampling +* Reshaping of large arrays + +DeepLabCut utilizes NumPy’s array capabilities throughout the workflow offered by the toolkit. In particular, NumPy is used for sampling distinct frames for human annotation labeling, and for writing, editing and processing annotation data. Within TensorFlow the neural network is trained by DeepLabCut technology over thousands of iterations to predict the ground truth annotations from frames. For this purpose, target densities (scoremaps) are created to cast pose estimation as a image-to-image translation problem. To make the neural networks robust, data augmentation is employed, which requires the calculation of target scoremaps subject to various geometric and image processing steps. To make training fast, NumPy’s vectorization capabilities are leveraged. For inference, the most likely predictions from target scoremaps need to extracted and one needs to efficiently “link predictions to assemble individual animals”. + +{{< figure >}} +src = '/images/content_images/cs/deeplabcut-workflow.png' title = 'DeepLabCut Workflow' alt = 'workflow' attribution = '(Source: Mackenzie Mathis)' attributionlink = 'https://www.researchgate.net/figure/DeepLabCut-work-flow-The-diagram-delineates-the-work-flow-as-well-as-the-directory-and_fig1_329185962' +{{< /figure >}} + +## Summary + +Observing and efficiently describing behavior is a core tenant of modern ethology, neuroscience, medicine, and technology. [DeepLabCut](http://orga.cvss.cc/wp-content/uploads/2019/05/NathMathis2019.pdf) allows researchers to estimate the pose of the subject, efficiently enabling them to quantify the behavior. With only a small set of training images, the DeepLabCut Python toolbox allows training a neural network to within human level labeling accuracy, thus expanding its application to not only behavior analysis in the laboratory, but to potentially also in sports, gait analysis, medicine and rehabilitation studies. Complex combinatorics, data processing challenges faced by DeepLabCut algorithms are addressed through the use of NumPy's array manipulation capabilities. + +{{< figure >}} +src = '/images/content_images/cs/numpy_dlc_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} + +[cheetah-movement]: https://www.technologynetworks.com/neuroscience/articles/interview-a-deeper-cut-into-behavior-with-mackenzie-mathis-327618 + +[DLCToolkit]: https://github.com/DeepLabCut/DeepLabCut diff --git a/content/es/case-studies/gw-discov.md b/content/es/case-studies/gw-discov.md new file mode 100644 index 0000000000..08f54b7547 --- /dev/null +++ b/content/es/case-studies/gw-discov.md @@ -0,0 +1,77 @@ +--- +title: "Case Study: Discovery of Gravitational Waves" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/gw_sxs_image.png' title = 'Gravitational Waves' alt = 'binary coalesce black hole generating gravitational waves' attribution = '(Image Credits: The Simulating eXtreme Spacetimes (SXS) Project at LIGO)' attributionlink = 'https://youtu.be/Zt8Z_uzG71o' +{{< /figure >}} + +{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="David Shoemaker, *LIGO Scientific Collaboration*" >}} The scientific Python ecosystem is critical infrastructure for the research done at LIGO. +{{< /blockquote >}} + +## About [Gravitational Waves](https://www.nationalgeographic.com/news/2017/10/what-are-gravitational-waves-ligo-astronomy-science/) and [LIGO](https://www.ligo.caltech.edu) + +Gravitational waves are ripples in the fabric of space and time, generated by cataclysmic events in the universe such as collision and merging of two black holes or coalescing binary stars or supernovae. Observing GW can not only help in studying gravity but also in understanding some of the obscure phenomena in the distant universe and its impact. + +The [Laser Interferometer Gravitational-Wave Observatory (LIGO)](https://www.ligo.caltech.edu) was designed to open the field of gravitational-wave astrophysics through the direct detection of gravitational waves predicted by Einstein’s General Theory of Relativity. It comprises two widely separated interferometers within the United States — one in Hanford, Washington and the other in Livingston, Louisiana — operated in unison to detect gravitational waves. Each of them has multi-kilometer-scale gravitational wave detectors that use laser interferometry. The LIGO Scientific Collaboration (LSC), is a group of more than 1000 scientists from universities around the United States and in 14 other countries supported by more than 90 universities and research institutes; approximately 250 students actively contributing to the collaboration. The new LIGO discovery is the first observation of gravitational waves themselves, made by measuring the tiny disturbances the waves make to space and time as they pass through the earth. It has opened up new astrophysical frontiers that explore the warped side of the universe—objects and phenomena that are made from warped spacetime. + + +### Key Objectives + +* Though its [mission](https://www.ligo.caltech.edu/page/what-is-ligo) is to detect gravitational waves from some of the most violent and energetic processes in the Universe, the data LIGO collects may have far-reaching effects on many areas of physics including gravitation, relativity, astrophysics, cosmology, particle physics, and nuclear physics. +* Crunch observed data via numerical relativity computations that involves complex maths in order to discern signal from noise, filter out relevant signal and statistically estimate significance of observed data +* Data visualization so that the binary / numerical results can be comprehended. + + + +### The Challenges + +* **Computation** + + Gravitational Waves are hard to detect as they produce a very small effect and have tiny interaction with matter. Processing and analyzing all of LIGO's data requires a vast computing infrastructure.After taking care of noise, which is billions of times of the signal, there is still very complex relativity equations and huge amounts of data which present a computational challenge: [O(10^7) CPU hrs needed for binary merger analyses](https://youtu.be/7mcHknWWzNI) spread on 6 dedicated LIGO clusters + +* **Data Deluge** + + As observational devices become more sensitive and reliable, the challenges posed by data deluge and finding a needle in a haystack rise multi-fold. LIGO generates terabytes of data every day! Making sense of this data requires an enormous effort for each and every detection. For example, the signals being collected by LIGO must be matched by supercomputers against hundreds of thousands of templates of possible gravitational-wave signatures. + +* **Visualization** + + Once the obstacles related to understanding Einstein’s equations well enough to solve them using supercomputers are taken care of, the next big challenge was making data comprehensible to the human brain. Simulation modeling as well as signal detection requires effective visualization techniques. Visualization also plays a role in lending more credibility to numerical relativity in the eyes of pure science aficionados, who did not give enough importance to numerical relativity until imaging and simulations made it easier to comprehend results for a larger audience. Speed of complex computations and rendering, re-rendering images and simulations using latest experimental inputs and insights can be a time consuming activity that challenges researchers in this domain. + +{{< figure >}} +src = '/images/content_images/cs/gw_strain_amplitude.png' alt = 'gravitational waves strain amplitude' title = 'Estimated gravitational-wave strain amplitude from GW150914' attribution = '(Graph Credits: Observation of Gravitational Waves from a Binary Black Hole Merger, ResearchGate Publication)' attributionlink = 'https://www.researchgate.net/publication/293886905_Observation_of_Gravitational_Waves_from_a_Binary_Black_Hole_Merger' +{{< /figure >}} + +## NumPy’s Role in the Detection of Gravitational Waves + +Gravitational waves emitted from the merger cannot be computed using any technique except brute force numerical relativity using supercomputers. The amount of data LIGO collects is as incomprehensibly large as gravitational wave signals are small. + +NumPy, the standard numerical analysis package for Python, was utilized by the software used for various tasks performed during the GW detection project at LIGO. NumPy helped in solving complex maths and data manipulation at high speed. Here are some examples: + +* [Signal Processing](https://www.uv.es/virgogroup/Denoising_ROF.html): Glitch detection, [Noise identification and Data Characterization](https://ep2016.europython.eu/media/conference/slides/pyhton-in-gravitational-waves-research-communities.pdf) (NumPy, scikit-learn, scipy, matplotlib, pandas, pyCharm) +* Data retrieval: Deciding which data can be analyzed, figuring out whether it contains a signal - needle in a haystack +* Statistical analysis: estimate the statistical significance of observational data, estimating the signal parameters (e.g. masses of stars, spin velocity, and distance) by comparison with a model. +* Visualization of data + - Time series + - Spectrograms +* Compute Correlations +* Key [Software](https://github.com/lscsoft) developed in GW data analysis such as [GwPy](https://gwpy.github.io/docs/stable/overview.html) and [PyCBC](https://pycbc.org) uses NumPy and AstroPy under the hood for providing object based interfaces to utilities, tools, and methods for studying data from gravitational-wave detectors. + +{{< figure >}} +src = '/images/content_images/cs/gwpy-numpy-dep-graph.png' alt = 'gwpy-numpy depgraph' title = 'Dependency graph showing how GwPy package depends on NumPy' +{{< /figure >}} + +---- + +{{< figure >}} +src = '/images/content_images/cs/PyCBC-numpy-dep-graph.png' alt = 'PyCBC-numpy depgraph' title = 'Dependency graph showing how PyCBC package depends on NumPy' +{{< /figure >}} + +## Summary + +GW detection has enabled researchers to discover entirely unexpected phenomena while providing new insight into many of the most profound astrophysical phenomena known. Number crunching and data visualization is a crucial step that helps scientists gain insights into data gathered from the scientific observations and understand the results. The computations are complex and cannot be comprehended by humans unless it is visualized using computer simulations that are fed with the real observed data and analysis. NumPy along with other Python packages such as matplotlib, pandas, and scikit-learn is [enabling researchers](https://www.gw-openscience.org/events/GW150914/) to answer complex questions and discover new horizons in our understanding of the universe. + +{{< figure >}} +src = '/images/content_images/cs/numpy_gw_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} diff --git a/content/es/citing-numpy.md b/content/es/citing-numpy.md new file mode 100644 index 0000000000..50b9bb2b78 --- /dev/null +++ b/content/es/citing-numpy.md @@ -0,0 +1,35 @@ +--- +title: Citando a NumPy +sidebar: false +--- + +Si NumPy ha sido importante en tu investigación y deseas reconocer el proyecto en tu publicación académica, te sugerimos que cites el siguiente documento: + +* Harris, C.R., Millman, K.J., van der Walt, S.J. et al. _Array programming with NumPy_. Nature 585, 357–362 (2020). DOI: [10.1038/s41586-020-2649-2](https://doi.org/10.1038/s41586-020-2649-2). ([Publisher link](https://www.nature.com/articles/s41586-020-2649-2)). + +_En formato BibTeX:_ + + ``` +@Article{ harris2020array, + title = {Array programming with {NumPy}}, + author = {Charles R. Harris and K. Jarrod Millman and St{\'{e}}fan J. + van der Walt and Ralf Gommers and Pauli Virtanen and David + Cournapeau and Eric Wieser and Julian Taylor and Sebastian + Berg and Nathaniel J. Smith and Robert Kern and Matti Picus + and Stephan Hoyer and Marten H. van Kerkwijk and Matthew + Brett and Allan Haldane and Jaime Fern{\'{a}}ndez del + R{\'{i}}o and Mark Wiebe and Pearu Peterson and Pierre + G{\'{e}}rard-Marchant and Kevin Sheppard and Tyler Reddy and + Warren Weckesser and Hameer Abbasi and Christoph Gohlke and + Travis E. Oliphant}, + year = {2020}, + month = sep, + journal = {Nature}, + volume = {585}, + number = {7825}, + pages = {357--362}, + doi = {10.1038/s41586-020-2649-2}, + publisher = {Springer Science and Business Media {LLC}}, + url = {https://doi.org/10.1038/s41586-020-2649-2} +} +``` diff --git a/content/es/code-of-conduct.md b/content/es/code-of-conduct.md new file mode 100644 index 0000000000..5ee1f4bcbe --- /dev/null +++ b/content/es/code-of-conduct.md @@ -0,0 +1,83 @@ +--- +title: NumPy Code of Conduct +sidebar: false +aliases: + - /conduct.html +--- + +### Introduction + +This Code of Conduct applies to all spaces managed by the NumPy project, including all public and private mailing lists, issue trackers, wikis, blogs, Twitter, and any other communication channel used by our community. The NumPy project does not organise in-person events, however events related to our community should have a code of conduct similar in spirit to this one. + +This Code of Conduct should be honored by everyone who participates in the NumPy community formally or informally, or claims any affiliation with the project, in any project-related activities and especially when representing the project, in any role. + +This code is not exhaustive or complete. It serves to distill our common understanding of a collaborative, shared environment and goals. Please try to follow this code in spirit as much as in letter, to create a friendly and productive environment that enriches the surrounding community. + +### Specific Guidelines + +We strive to: + +1. Be open. We invite anyone to participate in our community. We prefer to use public methods of communication for project-related messages, unless discussing something sensitive. This applies to messages for help or project-related support, too; not only is a public support request much more likely to result in an answer to a question, it also ensures that any inadvertent mistakes in answering are more easily detected and corrected. +2. Be empathetic, welcoming, friendly, and patient. We work together to resolve conflict, and assume good intentions. We may all experience some frustration from time to time, but we do not allow frustration to turn into a personal attack. A community where people feel uncomfortable or threatened is not a productive one. +3. Be collaborative. Our work will be used by other people, and in turn we will depend on the work of others. When we make something for the benefit of the project, we are willing to explain to others how it works, so that they can build on the work to make it even better. Any decision we make will affect users and colleagues, and we take those consequences seriously when making decisions. +4. Be inquisitive. Nobody knows everything! Asking questions early avoids many problems later, so we encourage questions, although we may direct them to the appropriate forum. We will try hard to be responsive and helpful. +5. Be careful in the words that we choose. We are careful and respectful in our communication, and we take responsibility for our own speech. Be kind to others. Do not insult or put down other participants. We will not accept harassment or other exclusionary behaviour, such as: + * Violent threats or language directed against another person. + * Sexist, racist, or otherwise discriminatory jokes and language. + * Posting sexually explicit or violent material. + * Posting (or threatening to post) other people’s personally identifying information (“doxing”). + * Sharing private content, such as emails sent privately or non-publicly, or unlogged forums such as IRC channel history, without the sender’s consent. + * Personal insults, especially those using racist or sexist terms. + * Unwelcome sexual attention. + * Excessive profanity. Please avoid swearwords; people differ greatly in their sensitivity to swearing. + * Repeated harassment of others. In general, if someone asks you to stop, then stop. + * Advocating for, or encouraging, any of the above behaviour. + +### Diversity Statement + +The NumPy project welcomes and encourages participation by everyone. We are committed to being a community that everyone enjoys being part of. Although we may not always be able to accommodate each individual’s preferences, we try our best to treat everyone kindly. + +No matter how you identify yourself or how others perceive you: we welcome you. Though no list can hope to be comprehensive, we explicitly honour diversity in: age, culture, ethnicity, genotype, gender identity or expression, language, national origin, neurotype, phenotype, political beliefs, profession, race, religion, sexual orientation, socioeconomic status, subculture and technical ability, to the extent that these do not conflict with this code of conduct. + +Though we welcome people fluent in all languages, NumPy development is conducted in English. + +Standards for behaviour in the NumPy community are detailed in the Code of Conduct above. Participants in our community should uphold these standards in all their interactions and help others to do so as well (see next section). + +### Reporting Guidelines + +We know that it is painfully common for internet communication to start at or devolve into obvious and flagrant abuse. We also recognize that sometimes people may have a bad day, or be unaware of some of the guidelines in this Code of Conduct. Please keep this in mind when deciding on how to respond to a breach of this Code. + +For clearly intentional breaches, report those to the Code of Conduct Committee (see below). For possibly unintentional breaches, you may reply to the person and point out this code of conduct (either in public or in private, whatever is most appropriate). If you would prefer not to do that, please feel free to report to the Code of Conduct Committee directly, or ask the Committee for advice, in confidence. + +You can report issues to the NumPy Code of Conduct Committee at numpy-conduct@googlegroups.com. + +Currently, the Committee consists of: + +* Stefan van der Walt +* Melissa Weber Mendonça +* Rohit Goswami + +If your report involves any members of the Committee, or if they feel they have a conflict of interest in handling it, then they will recuse themselves from considering your report. Alternatively, if for any reason you feel uncomfortable making a report to the Committee, then you can also contact senior NumFOCUS staff at [conduct@numfocus.org](https://numfocus.org/code-of-conduct#persons-responsible). + +### Incident reporting resolution & Code of Conduct enforcement + +_This section summarizes the most important points, more details can be found in_ [NumPy Code of Conduct - How to follow up on a report](report-handling-manual). + +We will investigate and respond to all complaints. The NumPy Code of Conduct Committee and the NumPy Steering Committee (if involved) will protect the identity of the reporter, and treat the content of complaints as confidential (unless the reporter agrees otherwise). + +In case of severe and obvious breaches, e.g. personal threat or violent, sexist or racist language, we will immediately disconnect the originator from NumPy communication channels; please see the manual for details. + +In cases not involving clear severe and obvious breaches of this Code of Conduct the process for acting on any received Code of Conduct violation report will be: + +1. acknowledge report is received, +2. reasonable discussion/feedback, +3. mediation (if feedback didn’t help, and only if both reporter and reportee agree to this), +4. enforcement via transparent decision (see [Resolutions](report-handling-manual/#resolutions)) by the Code of Conduct Committee. + +The Committee will respond to any report as soon as possible, and at most within 72 hours. + +### Endnotes + +We are thankful to the groups behind the following documents, from which we drew content and inspiration: + +- [The SciPy Code of Conduct](https://docs.scipy.org/doc/scipy/dev/conduct/code_of_conduct.html) diff --git a/content/es/community.md b/content/es/community.md new file mode 100644 index 0000000000..0e1e53211b --- /dev/null +++ b/content/es/community.md @@ -0,0 +1,66 @@ +--- +title: Comunidad +sidebar: false +--- + +NumPy es un proyecto de código abierto impulsado por la comunidad y desarrollado por un grupo diverso de [colaboradores](/teams/). El liderazgo de NumPy se ha comprometido firmemente a crear una comunidad abierta, inclusiva y positiva. Por favor, lee el [Código de Conducta de NumPy](/code-of-conduct) para obtener orientación sobre cómo interactuar con los demás de una manera que haga que la comunidad prospere. + +Ofrecemos varios canales de comunicación para aprender, compartir conocimientos y conectarse con otros dentro de la comunidad de NumPy. + + +## Participa en línea + +Las siguientes son formas de relacionarse directamente con el proyecto y la comunidad de NumPy. _Ten en cuenta que animamos a los usuarios y a los miembros de la comunidad a apoyarse mutuamente por preguntas de uso - ver [Obtener ayuda](/gethelp)._ + + +### [Lista de correo de NumPy](https://mail.python.org/mailman/listinfo/numpy-discussion) + +Este es el foro principal para discusiones más extensas, como añadir nuevas características a NumPy, hacer cambios en el mapa de ruta de NumPy, y todo tipo de proceso de toma de decisiones sobre el proyecto. Aquí también se hacen los anuncios sobre NumPy, tales como lanzamientos, reuniones de desarrolladores, sprints o conferencias. + +En esta lista, por favor, utiliza el botón de envío inferior, responde a la lista (en lugar de a otro remitente) y no respondas a los resúmenes. El archivo de consulta de esta lista está disponible [aquí](https://mail.python.org/archives/list/numpy-discussion@python.org/). + +*** + +### [Seguimiento de incidencias en GitHub](https://github.com/numpy/numpy/issues) + +- Para informes de error (por ejemplo, "`np.arange(3).shape` devuelve `(5,)`, cuando debería devolver `(3,)`"); +- problemas en la documentación (por ejemplo, "Esta sección me pareció poco clara"); +- y solicitudes de funcionalidades (por ejemplo, "Me gustaría tener un nuevo método de interpolación en `np.percentile`"). + +_Ten en cuenta que GitHub no es el lugar adecuado para reportar una vulnerabilidad de seguridad. Si crees que has encontrado una vulnerabilidad de seguridad en NumPy, por favor repórtalo [aquí](https://tidelift.com/docs/security)._ + +*** + +### [Slack](https://numpy-team.slack.com) + +Una sala de chat en tiempo real para hacer preguntas sobre las _contribuciones_ a NumPy. Este es un espacio privado, destinado específicamente a las personas que no se atreven a plantear sus preguntas o ideas en la lista de correo pública o en GitHub. Por favor, visita [aquí](https://numpy.org/devdocs/dev/index.html#contributing-to-numpy) para más detalles, y sobre cómo obtener una invitación. + + +## Grupos de Estudio y Reuniones + +Si desea encontrar un grupo de estudio o reunión local para aprender más sobre NumPy y el ecosistema más amplio de paquetes de Python para ciencia de datos y computación científica, te recomendamos que explores los [PyData meetups](https://www.meetup.com/pro/pydata/) (más de 150 reuniones, más de 100,000 miembros). + +NumPy también organiza ocasionalmente sprints presenciales para su equipo y colaboradores interesados. Estos normalmente se planifican con varios meses de anticipación y se anunciarán en la [lista de correo](https://mail.python.org/mailman/listinfo/numpy-discussion) y en [Twitter](https://twitter.com/numpy_team). + + +## Conferencias + +El proyecto NumPy no organiza sus propias conferencias. Las conferencias que tradicionalmente han sido más populares entre los responsables, colaboradores y usuarios de NumPy son la serie de conferencias de SciPy y PyData: + +- [SciPy US](https://conference.scipy.org) +- [EuroSciPy](https://www.euroscipy.org) +- [SciPy Latinoamérica](https://www.scipyla.org) +- [SciPy India](https://scipy.in) +- [SciPy Japan](https://conference.scipy.org) +- [Conferencias PyData](https://pydata.org/event-schedule/) (de 15 a 20 eventos al año, repartidos entre muchos países) + +Muchas de estas conferencias incluyen tutoriales y/o sprints que cubren NumPy, en donde puedes aprender cómo contribuir a Numpy o proyectos de código abierto relacionados. + + +## Únete a la comunidad Numpy + +Para prosperar, el proyecto NumPy necesita tu experiencia y entusiasmo. ¿No sabes programar? ¡No es un problema! Hay muchas maneras de contribuir a NumPy. + +Si te interesa colaborar en NumPy (¡yupi!) te recomendamos que visites nuestra página [Contribuir](/contribute). + +No dudes en pasar a saludarnos en uno de nuestros encuentros de la comunidad. Para enterarte del próximo, consulta nuestro calendario de eventos [aquí](https://scientific-python.org/calendars/). diff --git a/content/es/config.yaml b/content/es/config.yaml new file mode 100644 index 0000000000..1205ea6f24 --- /dev/null +++ b/content/es/config.yaml @@ -0,0 +1,140 @@ +languageName: Inglés +params: + description: '¿Por qué NumPy? Potentes matrices n-dimensionales. Herramientas de cálculo numérico. Interoperable. Rendimiento. Código abierto.' + navbarlogo: + image: logo.svg + text: NumPy + link: / + hero: + #Main hero title + title: NumPy + #Hero subtitle (optional) + subtitle: El paquete fundamental para la computación científica con Python + #Button text + buttontext: "Última versión: NumPy 1.25. Ver todas las versiones" + #Where the main hero button links to + buttonlink: "/news/#releases" + #Hero image (from static/images/___) + image: logo.svg + shell: + title: marcador + intro: + - + title: Prueba NumPy + text: Utilice el terminal interactivo para probar NumPy en el navegador + docslink: No olvides echarle un ojo a la documentación. + casestudies: + title: CASOS DE ESTUDIO + features: + - + title: Primera imagen de un Agujero Negro + text: Cómo NumPy, junto con bibliotecas como SciPy y Matplotlib que dependen de NumPy, permitió al Event Horizon Telescope producir la primera imagen de un agujero negro + img: /images/content_images/case_studies/blackhole.png + alttext: Primera imagen de un agujero negro. Es un círculo anaranjado con fondo negro. + url: /case-studies/blackhole-image + - + title: Detección de Ondas Gravitacionales + text: En 1916 Albert Einstein predijo las ondas gravitacionales; 100 años después se confirmó su existencia por científicos del LIGO, utilizando NumPy. + img: /images/content_images/case_studies/gravitional.png + alttext: Dos cuerpos orbitándose mutuamente. Estos desplazan la gravedad a su alrededor. + url: /case-studies/gw-discov + - + title: Analíticas Deportivas + text: El uso de Analíticas está cambiando al Cricket, al mejorar el rendimiento de los jugadores y equipos, mediante modelos estadísticos y análisis predictivos. NumPy permite realizar muchos de estos análisis. + img: /images/content_images/case_studies/sports.jpg + alttext: Bola de Cricket sobre un campo verde. + url: /case-studies/cricket-analytics + - + title: Estimación de la pose mediante aprendizaje profundo + text: DeepLabCut utiliza NumPy para acelerar estudios científicos que implican la observación del comportamiento animal para una mejor comprensión del control motriz, a través de especies y escalas de tiempo. + img: /images/content_images/case_studies/deeplabcut.png + alttext: Análisis de la pose de un Guepardo + url: /case-studies/deeplabcut-dnn + tabs: + title: ECOSISTEMA + section5: false + navbar: + - + title: Instalar + url: /install + - + title: Documentación + url: https://numpy.org/doc/stable + - + title: Aprende + url: /learn + - + title: Comunidad + url: /community + - + title: Acerca de nosotros + url: /about + - + title: Noticias + url: /news + - + title: Contribuye + url: /contribute + footer: + logo: logo.svg + socialmediatitle: "" + socialmedia: + - + link: https://github.com/numpy/numpy + icon: github + - + link: https://www.youtube.com/channel/UCguIL9NZ7ybWK5WQ53qbHng + icon: youtube + - + link: https://twitter.com/numpy_team + icon: twitter + quicklinks: + column1: + title: "" + links: + - + text: Instalar + link: /install + - + text: Documentación + link: https://numpy.org/doc/stable + - + text: Aprende + link: /learn + - + text: Citando a NumPy + link: /citing-numpy + - + text: Mapa de ruta + link: https://numpy.org/neps/roadmap.html + column2: + links: + - + text: Acerca de nosotros + link: /about + - + text: Comunidad + link: /community + - + text: Encuestas a usuarios + link: /user-surveys + - + text: Contribuye + link: /contribute + - + text: Código de Conducta + link: /code-of-conduct + column3: + links: + - + text: Buscar ayuda + link: /gethelp + - + text: Términos de uso + link: /terms + - + text: Confidencialidad + link: /privacy + - + text: Kit de prensa + link: /press-kit diff --git a/content/es/contribute.md b/content/es/contribute.md new file mode 100644 index 0000000000..75e24296f6 --- /dev/null +++ b/content/es/contribute.md @@ -0,0 +1,66 @@ +--- +title: Contribuye a NumPy +sidebar: false +--- + +¡El proyecto NumPy agradece tu experiencia y entusiasmo! Tus opciones no se limitan a la programación. Como puedes ver más abajo, existen muchas áreas en las que necesitamos **tu** ayuda. + +Si no estás seguro por dónde empezar o cómo encajan tus habilidades, _¡acércate!_ Puedes preguntar en la [lista de correos](https://mail.python.org/mailman/listinfo/numpy-discussion) o [GitHub](http://github.com/numpy/numpy) (abre una [propuesta](https://github.com/numpy/numpy/issues) o comenta en una relevante). + +Estos son nuestros canales preferidos (el código abierto es abierto por naturaleza), pero si prefieres hablar de manera privada, contacta a nuestros coordinadores de la communidad en o en [Slack](https://numpy-team.slack.com) (escribe a para recibir una invitación). + +También hacemos _llamadas a la comunidad_ de manera quincenal, cuyos detalles se anuncian en la [lista de correo](https://mail.python.org/mailman/listinfo/numpy-discussion). Te invitamos a unirte. Si es la primera vez que contribuyes al código abierto, también te recomendamos encarecidamente que leas [esta guía](https://opensource.guide/how-to-contribute/). + +Nuestra comunidad aspira a tratar a todos por igual y a valorar todas las contribuciones. Tenemos un [Código de Conducta](/code-of-conduct) para fomentar un entorno abierto y acogedor. + +### Escribiendo código + +Programadores, esta [guía](https://numpy.org/devdocs/dev/index.html#development-process-summary) explica cómo contribuir al código base.
También revisa nuestro [canal de YouTube](https://www.youtube.com/playlist?list=PLCK6zCrcN3GXBUUzDr9L4__LnXZVtaIzS) por consejos adicionales. + + +### Revisando solicitudes de cambios +El proyecto tiene más de 250 solicitudes de cambios abiertos, lo que significa muchas mejoras potenciales y muchos colaboradores de código abierto esperando retroalimentación. Si eres un desarrollador que conoce NumPy, puedes ayudar aunque no estés familiarizado con el código base. Puedes: +* resumir un debate extenso +* categorizar PRs de documentación +* probar los cambios propuestos + + +### Creando material educativo + +La [Guía de usuario](https://numpy.org/devdocs) de NumPy está en proceso de rehabilitación. Necesitamos nuevos tutoriales, instrucciones y explicaciones detalladas, y la página necesita una reestructuración. Las oportunidades no se limitan a escritores. También ejemplos prácticos, notebooks y vídeos. La propuesta [NEP 44 - Reestructuración de la Documentación NumPy](https://numpy.org/neps/nep-0044-restructuring-numpy-docs.html) expone nuestras ideas -- y tal vez tú puedas tener otras. + + +### Clasificación de propuestas + +El [rastreador de propuestas de NumPy](https://github.com/numpy/numpy/issues) tiene _muchos_ temas abiertos. Algunos ya no son válidos, otros deberían priorizarse y otros serían buenos temas para nuevos colaboradores. Puedes: + +* revisar si errores antiguos siguen presentes +* encontrar propuestas duplicadas, y enlazar las relacionadas +* añadir buenos reproductores autónomos de las incidencias +* etiquetar correctamente las propuestas (para ello es necesario tener derechos de categorización, solo necesitas preguntar) + +Por favor, solo sumérgete. + + +### Desarrollo de la Página Web + +Acabamos de renovar nuestro sitio web, pero aún no hemos terminado. Si te gusta el desarrollo web, estas [incidencias](https://github.com/numpy/numpy.org/issues?q=is%3Aissue+is%3Aopen+label%3Adesign) enumeran algunas de nuestras necesidades insatisfechas -- y no dudes en compartir tus propias ideas. + + +### Diseño gráfico + +Apenas podemos empezar a enumerar las aportaciones que puede hacer un diseñador gráfico. Nuestra documentación está sedienta de ilustraciones; nuestro sitio web, en pleno crecimiento, ansía imágenes... las oportunidades abundan. + + +### Traduciendo el contenido de la página web + +Planeamos múltiples traducciones de [numpy.org](https://numpy.org) para hacer a NumPy accesible a los usuarios en su lengua materna. Los traductores voluntarios son el núcleo de este esfuerzo. Consulta [aquí](https://numpy.org/neps/nep-0028-website-redesign.html#translation-multilingual-i18n) para más información; comenta en [este tema de GitHub](https://github.com/numpy/numpy.org/issues/55) para inscribirte. + + +### Coordinación y divulgación de la comunidad + +A través del contacto con la comunidad compartimos nuestro trabajo más ampliamente, y aprendemos en dónde nos estamos quedando cortos. Estamos ansiosos por conseguir más gente involucrada en esfuerzos como nuestra cuenta de [Twitter](https://twitter.com/numpy_team), organizando [code sprints](https://scisprints.github.io/) de NumPy, un boletín y quizás un blog. + +### Recaudación de fondos + +NumPy fue durante muchos años un proyecto voluntario, pero a medida que crecía su importancia se hizo evidente que necesitaríamos apoyo financiero para garantizar su estabilidad y crecimiento. [Esta charla en SciPy'19](https://www.youtube.com/watch?v=dBTJD_FDVjU) explica cuánta diferencia ha supuesto este apoyo. Como todo en el mundo sin ánimo de lucro, estamos constantemente en busca de subvenciones, patrocinios y otros tipos de ayuda. Tenemos varias ideas y, por supuesto, aceptamos más. La recaudación de fondos es una habilidad escasa aquí -- apreciaríamos tu ayuda. diff --git a/content/es/gethelp.md b/content/es/gethelp.md new file mode 100644 index 0000000000..5da0e518f7 --- /dev/null +++ b/content/es/gethelp.md @@ -0,0 +1,20 @@ +--- +title: Buscar Ayuda +sidebar: false +--- + +**Development issues:** For NumPy development-related matters (e.g., bug reports), please see [Community](/community). + +**User questions:** The best way to get help is to post your question to a site like [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) or [Reddit](https://www.reddit.com/r/Numpy/). We wish we could keep an eye on these sites, or answer questions directly, but the volume is a little overwhelming! + +### [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) + +Un foro para hacer preguntas de uso, como por ejemplo: "¿Cómo hago X cosa en NumPy?". Por favor [usa la etiqueta `#numpy`](https://stackoverflow.com/help/tagging) + +*** + +### [Reddit](https://www.reddit.com/r/Numpy/) + +Otro foro para hacer preguntas de uso. + +*** diff --git a/content/es/history.md b/content/es/history.md new file mode 100644 index 0000000000..e4e988a3df --- /dev/null +++ b/content/es/history.md @@ -0,0 +1,21 @@ +--- +title: Historia de NumPy +sidebar: false +--- + +NumPy es una biblioteca fundamental de Python que proporciona estructuras de datos de matrices y rutinas numéricas rápidas relacionadas. Cuando se puso en marcha, la biblioteca contaba con escasos fondos y la escribían principalmente estudiantes de posgrado, muchos de ellos sin formación en informática y, a menudo, sin la bendición de sus asesores. Imaginar siquiera que un pequeño grupo de estudiantes programadores "rebeldes" pudiera derribar el ecosistema de software de investigación, ya establecido y respaldado por millones en financiación y cientos de ingenieros altamente cualificados, era absurdo. Sin embargo, las motivaciones filosóficas detrás de la pila de herramientas totalmente abierta, en combinación con una comunidad entusiasta y amistosa con un enfoque singular, han demostrado ser favorable a largo plazo. Hoy en día, científicos, ingenieros y muchos otros profesionales en todo el mundo confían en NumPy. Por ejemplo, los scripts publicados en el análisis de ondas gravitacionales utilizan NumPy, y el proyecto de imagen del agujero negro M87 cita directamente a NumPy. + +Para conocer en profundidad los hitos en el desarrollo de NumPy y las bibliotecas relacionadas, consulta [arxiv.org](arxiv.org/abs/1907.10121). + +Si deseas obtener una copia de las bibliotecas originales Numeric y Numarray, sigue los siguientes enlaces: + +[Página de Descarga de *Numeric*](https://sourceforge.net/projects/numpy/files/Old%20Numeric/)* + +[Página de Descarga de *Numarray*](https://sourceforge.net/projects/numpy/files/Old%20Numarray/)* + +*Ten en cuenta que estos paquetes antiguos ya no se mantienen, y se recomienda encarecidamente a los usuarios que utilicen NumPy para cualquier propósito relacionado con matrices o que refactoricen cualquier código preexistente para utilizar la biblioteca NumPy. + +### Documentación Histórica + +[Descarga el Manual de *`Numeric'*](static/numeric-manual.pdf) + diff --git a/content/es/install.md b/content/es/install.md new file mode 100644 index 0000000000..f64a17841d --- /dev/null +++ b/content/es/install.md @@ -0,0 +1,126 @@ +--- +title: Installing NumPy +sidebar: false +--- + +The only prerequisite for installing NumPy is Python itself. If you don't have Python yet and want the simplest way to get started, we recommend you use the [Anaconda Distribution](https://www.anaconda.com/download) - it includes Python, NumPy, and many other commonly used packages for scientific computing and data science. + +NumPy can be installed with `conda`, with `pip`, with a package manager on macOS and Linux, or [from source](https://numpy.org/devdocs/building). For more detailed instructions, consult our [Python and NumPy installation guide](#python-numpy-install-guide) below. + +**CONDA** + +If you use `conda`, you can install NumPy from the `defaults` or `conda-forge` channels: + +```bash +# Best practice, use an environment rather than install in the base env +conda create -n my-env +conda activate my-env +# If you want to install from conda-forge +conda config --env --add channels conda-forge +# The actual install command +conda install numpy +``` + +**PIP** + +If you use `pip`, you can install NumPy with: + +```bash +pip install numpy +``` +Also when using pip, it's good practice to use a virtual environment - see [Reproducible Installs](#reproducible-installs) below for why, and [this guide](https://dev.to/bowmanjd/python-tools-for-managing-virtual-environments-3bko#howto) for details on using virtual environments. + + + + +# Python and NumPy installation guide + +Installing and managing packages in Python is complicated, there are a number of alternative solutions for most tasks. This guide tries to give the reader a sense of the best (or most popular) solutions, and give clear recommendations. It focuses on users of Python, NumPy, and the PyData (or numerical computing) stack on common operating systems and hardware. + +## Recommendations + +We'll start with recommendations based on the user's experience level and operating system of interest. If you're in between "beginning" and "advanced", please go with "beginning" if you want to keep things simple, and with "advanced" if you want to work according to best practices that go a longer way in the future. + +### Beginning users + +On all of Windows, macOS, and Linux: + +- Install [Anaconda](https://www.anaconda.com/download) (it installs all packages you need and all other tools mentioned below). +- For writing and executing code, use notebooks in [JupyterLab](https://jupyterlab.readthedocs.io/en/stable/index.html) for exploratory and interactive computing, and [Spyder](https://www.spyder-ide.org/) or [Visual Studio Code](https://code.visualstudio.com/) for writing scripts and packages. +- Use [Anaconda Navigator](https://docs.anaconda.com/anaconda/navigator/) to manage your packages and start JupyterLab, Spyder, or Visual Studio Code. + + +### Advanced users + +#### Conda + +- Install [Miniforge](https://github.com/conda-forge/miniforge). +- Keep the `base` conda environment minimal, and use one or more [conda environments](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html) to install the package you need for the task or project you're working on. + +#### Alternative if you prefer pip/PyPI + +For users who know, from personal preference or reading about the main differences between conda and pip below, they prefer a pip/PyPI-based solution, we recommend: +- Install Python from [python.org](https://www.python.org/downloads/), [Homebrew](https://brew.sh/), or your Linux package manager. +- Use [Poetry](https://python-poetry.org/) as the most well-maintained tool that provides a dependency resolver and environment management capabilities in a similar fashion as conda does. + + +## Python package management + +Managing packages is a challenging problem, and, as a result, there are lots of tools. For web and general purpose Python development there's a whole [host of tools](https://packaging.python.org/guides/tool-recommendations/) complementary with pip. For high-performance computing (HPC), [Spack](https://github.com/spack/spack) is worth considering. For most NumPy users though, [conda](https://conda.io/en/latest/) and [pip](https://pip.pypa.io/en/stable/) are the two most popular tools. + + +### Pip & conda + +The two main tools that install Python packages are `pip` and `conda`. Their functionality partially overlaps (e.g. both can install `numpy`), however, they can also work together. We'll discuss the major differences between pip and conda here - this is important to understand if you want to manage packages effectively. + +The first difference is that conda is cross-language and it can install Python, while pip is installed for a particular Python on your system and installs other packages to that same Python install only. This also means conda can install non-Python libraries and tools you may need (e.g. compilers, CUDA, HDF5), while pip can't. + +The second difference is that pip installs from the Python Packaging Index (PyPI), while conda installs from its own channels (typically "defaults" or "conda-forge"). PyPI is the largest collection of packages by far, however, all popular packages are available for conda as well. + +The third difference is that conda is an integrated solution for managing packages, dependencies and environments, while with pip you may need another tool (there are many!) for dealing with environments or complex dependencies. + + + +### Reproducible installs + +As libraries get updated, results from running your code can change, or your code can break completely. It's important to be able to reconstruct the set of packages and versions you're using. Best practice is to: + +1. use a different environment per project you're working on, +2. record package names and versions using your package installer; each has its own metadata format for this: + - Conda: [conda environments and environment.yml](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html) + - Pip: [virtual environments](https://docs.python.org/3/tutorial/venv.html) and [requirements.txt](https://pip.readthedocs.io/en/latest/user_guide/#requirements-files) + - Poetry: [virtual environments and pyproject.toml](https://python-poetry.org/docs/basic-usage/) + + + +## NumPy packages & accelerated linear algebra libraries + +NumPy doesn't depend on any other Python packages, however, it does depend on an accelerated linear algebra library - typically [Intel MKL](https://software.intel.com/en-us/mkl) or [OpenBLAS](https://www.openblas.net/). Users don't have to worry about installing those (they're automatically included in all NumPy install methods). Power users may still want to know the details, because the used BLAS can affect performance, behavior and size on disk: + +- The NumPy wheels on PyPI, which is what pip installs, are built with OpenBLAS. The OpenBLAS libraries are included in the wheel. This makes the wheel larger, and if a user installs (for example) SciPy as well, they will now have two copies of OpenBLAS on disk. + +- In the conda defaults channel, NumPy is built against Intel MKL. MKL is a separate package that will be installed in the users' environment when they install NumPy. + +- In the conda-forge channel, NumPy is built against a dummy "BLAS" package. When a user installs NumPy from conda-forge, that BLAS package then gets installed together with the actual library - this defaults to OpenBLAS, but it can also be MKL (from the defaults channel), or even [BLIS](https://github.com/flame/blis) or reference BLAS. + +- The MKL package is a lot larger than OpenBLAS, it's about 700 MB on disk while OpenBLAS is about 30 MB. + +- MKL is typically a little faster and more robust than OpenBLAS. + +Besides install sizes, performance and robustness, there are two more things to consider: + +- Intel MKL is not open source. For normal use this is not a problem, but if a user needs to redistribute an application built with NumPy, this could be an issue. +- Both MKL and OpenBLAS will use multi-threading for function calls like `np.dot`, with the number of threads being determined by both a build-time option and an environment variable. Often all CPU cores will be used. This is sometimes unexpected for users; NumPy itself doesn't auto-parallelize any function calls. It typically yields better performance, but can also be harmful - for example when using another level of parallelization with Dask, scikit-learn or multiprocessing. + + +## Troubleshooting + +If your installation fails with the message below, see [Troubleshooting ImportError](https://numpy.org/doc/stable/user/troubleshooting-importerror.html). + +``` +IMPORTANT: PLEASE READ THIS FOR ADVICE ON HOW TO SOLVE THIS ISSUE! + +Importing the numpy c-extensions failed. This error can happen for +different reasons, often due to issues with your setup. +``` + diff --git a/content/es/learn.md b/content/es/learn.md new file mode 100644 index 0000000000..5140f5c664 --- /dev/null +++ b/content/es/learn.md @@ -0,0 +1,76 @@ +--- +title: Aprende +sidebar: false +--- + +Para la **documentación oficial de NumPy** visita [numpy.org/doc/stable](https://numpy.org/doc/stable). + +*** + +A continuación se muestra una colección de recursos educativos, tanto para el autoaprendizaje como para enseñar a otros, desarrollados por colaboradores de NumPy y aprobados por la comunidad. + +## Principiantes + +Hay un montón de información sobre NumPy allá afuera. Si eres nuevo, te recomendamos encarecidamente estos: + + **Tutoriales** + +* [NumPy Quickstart Tutorial](https://numpy.org/devdocs/user/quickstart.html) +* [Tutoriales de NumPy](https://numpy.org/numpy-tutorials) Una colección de tutoriales y materiales educativos en formato de cuadernos Jupyter desarrollados y mantenidos por el equipo de documentación de NumPy. Para enviar tu propio contenido, visita el repositorio [numpy-tutorials en GitHub](https://github.com/numpy/numpy-tutorials). +* [NumPy Illustrated: The Visual Guide to NumPy *by Lev Maximov*](https://betterprogramming.pub/3b1d4976de1d?sk=57b908a77aa44075a49293fa1631dd9b) +* [Lecturas de SciPy](https://scipy-lectures.org/) Además de cubrir NumPy, estas conferencias ofrecen una introducción más amplia al ecosistema científico de Python. +* [NumPy: the absolute basics for beginners](https://numpy.org/devdocs/user/absolute_beginners.html) +* [NumPy tutorial *by Nicolas Rougier*](https://github.com/rougier/numpy-tutorial) +* [Stanford CS231 *by Justin Johnson*](http://cs231n.github.io/python-numpy-tutorial/) +* [NumPy User Guide](https://numpy.org/devdocs) + + **Libros** + +* [Guide to NumPy *by Travis E. Oliphant*](http://web.mit.edu/dvp/Public/numpybook.pdf) Esta es la primera versión gratuita de 2006. Para conseguir la última versión (2015) mira [aquí](https://www.barnesandnoble.com/w/guide-to-numpy-travis-e-oliphant-phd/1122853007). +* [From Python to NumPy *by Nicolas P. Rougier*](https://www.labri.fr/perso/nrougier/from-python-to-numpy/) +* [Elegant SciPy](https://www.amazon.com/Elegant-SciPy-Art-Scientific-Python/dp/1491922877) *by Juan Nunez-Iglesias, Stefan van der Walt, and Harriet Dashnow* + +También puedes echar un vistazo a esta [lista de Goodreads](https://www.goodreads.com/shelf/show/python-scipy) sobre el tema "Python+SciPy". La mayoría de esos libros son sobre el "ecosistema SciPy", que tiene NumPy en su núcleo. + + **Vídeos** + +* [Introduction to Numerical Computing with NumPy](http://youtu.be/ZB7BZMhfPgk) *by Alex Chabot-Leclerc* + +*** + +## Avanzado + +Pruebe estos recursos avanzados para comprender mejor los conceptos de NumPy como indexación avanzada, división, apilamiento, álgebra lineal y mucho más. + + **Tutoriales** + +* [100 NumPy Exercises](http://www.labri.fr/perso/nrougier/teaching/numpy.100/index.html) *by Nicolas P. Rougier* +* [An Introduction to NumPy and Scipy](https://engineering.ucsb.edu/~shell/che210d/numpy.pdf) *by M. Scott Shell* +* [Numpy Medkits](http://mentat.za.net/numpy/numpy_advanced_slides/) *by Stéfan van der Walt* +* [Tutoriales de NumPy](https://numpy.org/numpy-tutorials) Una colección de tutoriales y materiales educativos en formato de cuadernos Jupyter desarrollados y mantenidos por el equipo de documentación de NumPy. Para enviar tu propio contenido, visita el repositorio [numpy-tutorials en GitHub](https://github.com/numpy/numpy-tutorials). + + **Libros** + +* [Python Data Science Handbook](https://www.amazon.com/Python-Data-Science-Handbook-Essential/dp/1491912057) *by Jake Vanderplas* +* [Python for Data Analysis](https://www.amazon.com/Python-Data-Analysis-Wrangling-IPython/dp/1491957662) *by Wes McKinney* +* [Numerical Python: Scientific Computing and Data Science Applications with Numpy, SciPy, and Matplotlib](https://www.amazon.com/Numerical-Python-Scientific-Applications-Matplotlib/dp/1484242459) *by Robert Johansson* + + **Vídeos** + +* [Advanced NumPy - broadcasting rules, strides, and advanced indexing](https://www.youtube.com/watch?v=cYugp9IN1-Q) *by Juan Nunez-Iglesias* + +*** + +## Charlas de NumPy + +* [The Future of NumPy Indexing](https://www.youtube.com/watch?v=o0EacbIbf58) *by Jaime Fernández* (2016) +* [Evolution of Array Computing in Python](https://www.youtube.com/watch?v=HVLPJnvInzM&t=10s) *by Ralf Gommers* (2019) +* [NumPy: what has changed and what is going to change?](https://www.youtube.com/watch?v=YFLVQFjRmPY) *by Matti Picus* (2019) +* [Inside NumPy](https://www.youtube.com/watch?v=dBTJD_FDVjU) *by Ralf Gommers, Sebastian Berg, Matti Picus, Tyler Reddy, Stefan van der Walt, Charles Harris* (2019) +* [Brief Review of Array Computing in Python](https://www.youtube.com/watch?v=f176j2g2eNc) *by Travis Oliphant* (2019) + +*** + +## Citando a NumPy + +Si NumPy ha sido importante en tu investigación y deseas reconocer el proyecto en tu publicación académica, consulta esta [información de citado](/citing-numpy). diff --git a/content/es/news.md b/content/es/news.md new file mode 100644 index 0000000000..7fc9ef2d82 --- /dev/null +++ b/content/es/news.md @@ -0,0 +1,279 @@ +--- +title: News +sidebar: false +newsHeader: "NumPy 2.0 release date: June 16" +date: 2024-05-23 +--- + +### NumPy 2.0 release date: June 16 + +_23 May, 2024_ -- We are excited to announce that NumPy 2.0 is planned to be released on June 16, 2024. This release has been over a year in the making, and is the first major release since 2006. Importantly, in addition to many new features and performance improvement, it contains **breaking changes** to the ABI as well as the Python and C APIs. It is likely that downstream packages and end user code needs to be adapted - if you can, please verify whether your code works with NumPy `2.0.0rc2`. **Please see the following for more details:** + +- The [NumPy 2.0 migration guide](https://numpy.org/devdocs/numpy_2_0_migration_guide.html) +- The [2.0.0 release notes](https://numpy.org/devdocs/release/2.0.0-notes.html) +- Announcement issue for status updates: [numpy#24300](https://github.com/numpy/numpy/issues/24300) + + +### NumFOCUS end of the year fundraiser +_Dec 19, 2023_ -- NumFOCUS has teamed up with PyCharm during their EOY campaign to offer a 30% discount on first-time PyCharm licenses. All year-one revenue from PyCharm purchases from now until December 23rd, 2023 will go directly to the NumFOCUS programs. + +Use unique URL that will allow to track purchases https://lp.jetbrains.com/support-data-science/ or a coupon code ISUPPORTDATASCIENCE  + +### NumPy 1.26.0 released + +_Sep 16, 2023_ -- [NumPy 1.26.0](https://numpy.org/doc/stable/release/1.26.0-notes.html) ahora está disponible. The highlights of the release are: + +* Soporte de Python 3.12.0. +* Compatibilidad con Cython 3.0.0. +* Utilización del sistema de compilación Meson +* Actualización del soporte de SIMD +* Correcciones de f2py, meson y soporte de bind(x) +* Soporte para la librería actualizada Accelerate BLAS/LAPACK + +La versión 1.26.0 de NumPy es la continuación de la serie 1.25.x que marca la transición al sistema de compilación Meson y que provee soporte para Cython 3.0.0. Un total de 20 personas contribuyeron a esta versión y 59 solicitudes de cambios fueron fusionadas. + +Las versiones de Python compatibles con esta versión son 3.9-3.12. + +### numpy.org is now available in Japanese and Portuguese + +_Aug 2, 2023_ -- numpy.org is now available in 2 additional languages: Japanese and Portuguese. This wouldn’t be possible without our dedicated volunteers: + +_Portuguese:_ +* Melissa Weber Mendonça (melissawm) +* Ricardo Prins (ricardoprins) +* Getúlio Silva (getuliosilva) +* Julio Batista Silva (jbsilva) +* Alexandre de Siqueira (alexdesiqueira) +* Alexandre B A Villares (villares) +* Vini Salazar (vinisalazar) + +_Japanese:_ +* Atsushi Sakai (AtsushiSakai) +* KKunai +* Tom Kelly (TomKellyGenetics) +* Yuji Kanagawa (kngwyu) +* Tetsuo Koyama (tkoyama010) + +The work on the translation infrastructure is supported with funding from CZI. + +Looking ahead, we’d love to translate the website into more languages. If you’d like to help, please connect with the NumPy Translations Team on Slack: https://join.slack.com/t/numpy-team/shared_invite/zt-1gokbq56s-bvEpo10Ef7aHbVtVFeZv2w. (Look for the #translations channel.) We are also building a Translations Team who will be working on localizing documentation and educational content across the Scientific Python ecosystem. If this piqued your interest, join us on the Scientific Python Discord: https://discord.gg/khWtqY6RKr. (Look for the #translation channel.) + +### NumPy 1.25.0 released + +_Jun 17, 2023_ -- [NumPy 1.25.0](https://numpy.org/doc/stable/release/1.25.0-notes.html) is now available. The highlights of the release are: + +* Support for MUSL, there are now MUSL wheels. +* Support for the Fujitsu C/C++ compiler. +* Object arrays are now supported in einsum. +* Support for the inplace matrix multiplication (`@=`). + +The NumPy 1.25.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase the execution speed, and clarify the documentation. There has also been preparatory work for the future NumPy 2.0.0, resulting in a large number of new and expired deprecations. + +A total of 148 people contributed to this release and 530 pull requests were merged. + +The Python versions supported by this release are 3.9-3.11. + +### Fostering an Inclusive Culture: Call for Participation + +_May 10, 2023_ -- Fostering an Inclusive Culture: Call for Participation + +How can we be better when it comes to diversity and inclusion? Read the report and find out how to get involved [here](https://contributor-experience.org/docs/posts/dei-report/). + +### NumPy documentation team leadership transition + +_Jan 6, 2023_ –- Mukulika Pahari and Ross Barnowski are appointed as the new NumPy documentation team leads replacing Melissa Mendonça. We thank Melissa for all her contributions to the NumPy official documentation and educational materials, and Mukulika and Ross for stepping up. + +### NumPy 1.24.0 released + +_Dec 18, 2022_ -- [NumPy 1.24.0](https://numpy.org/doc/stable/release/1.24.0-notes.html) is now available. The highlights of the release are: + +* New "dtype" and "casting" keywords for stacking functions. +* New F2PY features and fixes. +* Many new deprecations, check them out. +* Many expired deprecations, + +The NumPy 1.24.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase execution speed, and clarify the documentation. There are a large number of new and expired deprecations due to changes in dtype promotion and cleanups. It is the work of 177 contributors spread over 444 pull requests. The supported Python versions are 3.8-3.11. + +### Numpy 1.23.0 released + +_Jun 22, 2022_ -- [NumPy 1.23.0](https://numpy.org/doc/stable/release/1.23.0-notes.html) is now available. The highlights of the release are: + +* Implementation of `loadtxt` in C, greatly improving its performance. +* Exposure of DLPack at the Python level for easy data exchange. +* Changes to the promotion and comparisons of structured dtypes. +* Improvements to f2py. + +The NumPy 1.23.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase the execution speed, clarify the documentation, and expire old deprecations. It is the work of 151 contributors spread over 494 pull requests. The Python versions supported by this release 3.8-3.10. Python 3.11 will be supported when it reaches the rc stage. + +### NumFOCUS DEI research study: call for participation + +_Apr 13, 2022_ -- NumPy is working with [NumFOCUS](http://numfocus.org/) on a [research project](https://numfocus.org/diversity-inclusion-disc/a-pivotal-time-in-numfocuss-project-aimed-dei-efforts?eType=EmailBlastContent&eId=f41a86c3-60d4-4cf9-86cf-58eb49dc968c) funded by the [Gordon & Betty Moore Foundation](https://www.moore.org/) to understand the barriers to participation that contributors, particularly those from historically underrepresented groups, face in the open-source software community. The research team would like to talk to new contributors, project developers and maintainers, and those who have contributed in the past about their experiences joining and contributing to NumPy. + +**Interested in sharing your experiences?** + +Please complete this brief [“Participant Interest” form](https://numfocus.typeform.com/to/WBWVJSqe) which contains additional information on the research goals, privacy, and confidentiality considerations. Your participation will be valuable to the growth and sustainability of diverse and inclusive open-source software communities. Accepted participants will participate in a 30-minute interview with a research team member. + +### Numpy 1.22.0 release + +_Dec 31, 2021_ -- [NumPy 1.22.0](https://numpy.org/doc/stable/release/1.22.0-notes.html) is now available. The highlights of the release are: + +* Type annotations of the main namespace are essentially complete. Upstream is a moving target, so there will likely be further improvements, but the major work is done. This is probably the most user visible enhancement in this release. +* A preliminary version of the proposed [array API Standard](https://data-apis.org/array-api/latest/) is provided (see [NEP 47](https://numpy.org/neps/nep-0047-array-api-standard.html)). This is a step in creating a standard collection of functions that can be used across libraries such as CuPy and JAX. +* NumPy now has a DLPack backend. DLPack provides a common interchange format for array (tensor) data. +* New methods for `quantile`, `percentile`, and related functions. The new methods provide a complete set of the methods commonly found in the literature. +* The universal functions have been refactored to implement most of [NEP 43](https://numpy.org/neps/nep-0043-extensible-ufuncs.html). This also unlocks the ability to experiment with the future DType API. +* A new configurable memory allocator for use by downstream projects. + +NumPy 1.22.0 is a big release featuring the work of 153 contributors spread over 609 pull requests. The Python versions supported by this release are 3.8-3.10. + +### Advancing an inclusive culture in the scientific Python ecosystem + +_August 31, 2021_ -- We are happy to announce the Chan Zuckerberg Initiative has [awarded a grant](https://chanzuckerberg.com/newsroom/czi-awards-16-million-for-foundational-open-source-software-tools-essential-to-biomedicine/) to support the onboarding, inclusion, and retention of people from historically marginalized groups on scientific Python projects, and to structurally improve the community dynamics for NumPy, SciPy, Matplotlib, and Pandas. + +As a part of [CZI's Essential Open Source Software for Science program](https://chanzuckerberg.com/eoss/), this [Diversity & Inclusion supplemental grant](https://cziscience.medium.com/advancing-diversity-and-inclusion-in-scientific-open-source-eaabe6a5488b) will support the creation of dedicated Contributor Experience Lead positions to identify, document, and implement practices to foster inclusive open-source communities. This project will be led by Melissa Mendonça (NumPy), with additional mentorship and guidance provided by Ralf Gommers (NumPy, SciPy), Hannah Aizenman and Thomas Caswell (Matplotlib), Matt Haberland (SciPy), and Joris Van den Bossche (Pandas). + +This is an ambitious project aiming to discover and implement activities that should structurally improve the community dynamics of our projects. By establishing these new cross-project roles, we hope to introduce a new collaboration model to the Scientific Python communities, allowing community-building work within the ecosystem to be done more efficiently and with greater outcomes. We also expect to develop a clearer picture of what works and what doesn't in our projects to engage and retain new contributors, especially from historically underrepresented groups. Finally, we plan on producing detailed reports on the actions executed, explaining how they have impacted our projects in terms of representation and interaction with our communities. + +The two-year project is expected to start by November 2021, and we are excited to see the results from this work! [You can read the full proposal here](https://figshare.com/articles/online_resource/Advancing_an_inclusive_culture_in_the_scientific_Python_ecosystem/16548063). + +### 2021 NumPy survey + +_July 12, 2021_ -- At NumPy, we believe in the power of our community. 1,236 NumPy users from 75 countries participated in our inaugural survey last year. The survey findings gave us a very good understanding of what we should focus on for the next 12 months. + +It’s time for another survey, and we are counting on you once again. It will take about 15 minutes of your time. Besides English, the survey questionnaire is available in 8 additional languages: Bangla, French, Hindi, Japanese, Mandarin, Portuguese, Russian, and Spanish. + +Follow the link to get started: https://berkeley.qualtrics.com/jfe/form/SV_aaOONjgcBXDSl4q. + + +### Numpy 1.21.0 release + +_Jun 23, 2021_ -- [NumPy 1.21.0](https://numpy.org/doc/stable/release/1.21.0-notes.html) is now available. Los aspectos más destacados de esta versión son: + +- continued SIMD work covering more functions and platforms, +- initial work on the new dtype infrastructure and casting, +- universal2 wheels for Python 3.8 and Python 3.9 on Mac, +- improved documentation, +- improved annotations, +- new `PCG64DXSM` bitgenerator for random numbers. + +This NumPy release is the result of 581 merged pull requests contributed by 175 people. The Python versions supported for this release are 3.7-3.9, support for Python 3.10 will be added after Python 3.10 is released. + + +### 2020 NumPy survey results + +_Jun 22, 2021_ -- In 2020, the NumPy survey team in partnership with students and faculty from the University of Michigan and the University of Maryland conducted the first official NumPy community survey. Find the survey results here: https://numpy.org/user-survey-2020/. + + +### Numpy 1.20.0 release + +_Jan 30, 2021_ -- [NumPy 1.20.0](https://numpy.org/doc/stable/release/1.20.0-notes.html) is now available. This is the largest NumPy release to date, thanks to 180+ contributors. The two most exciting new features are: +- Type annotations for large parts of NumPy, and a new `numpy.typing` submodule containing `ArrayLike` and `DtypeLike` aliases that users and downstream libraries can use when adding type annotations in their own code. +- Multi-platform SIMD compiler optimizations, with support for x86 (SSE, AVX), ARM64 (Neon), and PowerPC (VSX) instructions. This yielded significant performance improvements for many functions (examples: [sin/cos](https://github.com/numpy/numpy/pull/17587), [einsum](https://github.com/numpy/numpy/pull/18194)). + +### Diversity in the NumPy project + +_Sep 20, 2020_ -- We wrote a [statement on the state of, and discussion on social media around, diversity and inclusion in the NumPy project](/diversity_sep2020). + + +### First official NumPy paper published in Nature! + +_Sep 16, 2020_ -- We are pleased to announce the publication of [the first official paper on NumPy](https://www.nature.com/articles/s41586-020-2649-2) as a review article in Nature. This comes 14 years after the release of NumPy 1.0. The paper covers applications and fundamental concepts of array programming, the rich scientific Python ecosystem built on top of NumPy, and the recently added array protocols to facilitate interoperability with external array and tensor libraries like CuPy, Dask, and JAX. + + +### Python 3.9 is coming, when will NumPy release binary wheels? + +_Sept 14, 2020_ -- Python 3.9 will be released in a few weeks. If you are an early adopter of Python versions, you may be dissapointed to find that NumPy (and other binary packages like SciPy) will not have binary wheels ready on the day of the release. It is a major effort to adapt the build infrastructure to a new Python version and it typically takes a few weeks for the packages to appear on PyPI and conda-forge. In preparation for this event, please make sure to +- update your `pip` to version 20.1 at least to support `manylinux2010` and `manylinux2014` +- use [`--only-binary=numpy`](https://pip.pypa.io/en/stable/reference/pip_install/#cmdoption-only-binary) or `--only-binary=:all:` to prevent `pip` from trying to build from source. + + +### Numpy 1.19.2 release + +_Sep 10, 2020_ -- [NumPy 1.19.2](https://numpy.org/devdocs/release/1.19.2-notes.html) is now available. This latest release in the 1.19 series fixes several bugs, prepares for the [upcoming Cython 3.x release](http://docs.cython.org/en/latest/src/changes.html) and pins setuptools to keep distutils working while upstream modifications are ongoing. The aarch64 wheels are built with the latest manylinux2014 release that fixes the problem of differing page sizes used by different linux distros. + +### The inaugural NumPy survey is live! + +_Jul 2, 2020_ -- This survey is meant to guide and set priorities for decision-making about the development of NumPy as software and as a community. The survey is available in 8 additional languages besides English: Bangla, Hindi, Japanese, Mandarin, Portuguese, Russian, Spanish and French. + +Please help us make NumPy better and take the survey [here](https://umdsurvey.umd.edu/jfe/form/SV_8bJrXjbhXf7saAl). + + +### NumPy has a new logo! + +_Jun 24, 2020_ -- NumPy now has a new logo: + +NumPy logo + +The logo is a modern take on the old one, with a cleaner design. Thanks to Isabela Presedo-Floyd for designing the new logo, as well as to Travis Vaught for the old logo that served us well for 15+ years. + + +### NumPy 1.19.0 release + +_Jun 20, 2020_ -- NumPy 1.19.0 is now available. This is the first release without Python 2 support, hence it was a "clean-up release". The minimum supported Python version is now Python 3.6. An important new feature is that the random number generation infrastructure that was introduced in NumPy 1.17.0 is now accessible from Cython. + + +### Season of Docs acceptance + +_May 11, 2020_ -- NumPy has been accepted as one of the mentor organizations for the Google Season of Docs program. We are excited about the opportunity to work with a technical writer to improve NumPy's documentation once again! For more details, please see [the official Season of Docs site](https://developers.google.com/season-of-docs/) and our [ideas page](https://github.com/numpy/numpy/wiki/Google-Season-of-Docs-2020-Project-Ideas). + + +### NumPy 1.18.0 release + +_Dec 22, 2019_ -- NumPy 1.18.0 is now available. After the major changes in 1.17.0, this is a consolidation release. It is the last minor release that will support Python 3.5. Highlights of the release includes the addition of basic infrastructure for linking with 64-bit BLAS and LAPACK libraries, and a new C-API for `numpy.random`. + +Please see the [release notes](https://github.com/numpy/numpy/releases/tag/v1.18.0) for more details. + + +### NumPy receives a grant from the Chan Zuckerberg Initiative + +_Nov 15, 2019_ -- We are pleased to announce that NumPy and OpenBLAS, one of NumPy's key dependencies, have received a joint grant for $195,000 from the Chan Zuckerberg Initiative through their [Essential Open Source Software for Science program](https://chanzuckerberg.com/eoss/) that supports software maintenance, growth, development, and community engagement for open source tools critical to science. + +This grant will be used to ramp up the efforts in improving NumPy documentation, website redesign, and community development to better serve our large and rapidly growing user base, and ensure the long-term sustainability of the project. While the OpenBLAS team will focus on addressing sets of key technical issues, in particular thread-safety, AVX-512, and thread-local storage (TLS) issues, as well as algorithmic improvements in ReLAPACK (Recursive LAPACK) on which OpenBLAS depends. + +More details on our proposed initiatives and deliverables can be found in the [full grant proposal](https://figshare.com/articles/Proposal_NumPy_OpenBLAS_for_Chan_Zuckerberg_Initiative_EOSS_2019_round_1/10302167). The work is scheduled to start on Dec 1st, 2019 and continue for the next 12 months. + + + + +## Releases + +Here is a list of NumPy releases, with links to release notes. Bugfix releases (only the `z` changes in the `x.y.z` version number) have no new features; minor releases (the `y` increases) do. + +- NumPy 1.26.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.4)) -- _5 Feb 2024_. +- NumPy 1.26.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.3)) -- _2 Jan 2024_. +- NumPy 1.26.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.2)) -- _12 Nov 2023_. +- NumPy 1.26.1 ([notas de publicación](https://github.com/numpy/numpy/releases/tag/v1.26.1)) -- _14 Oct 2023_. +- NumPy 1.26.0 ([notas de publicación](https://github.com/numpy/numpy/releases/tag/v1.26.0)) -- _16 Sep 2023_. +- NumPy 1.25.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.25.2)) -- _31 Jul 2023_. +- NumPy 1.25.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.25.1)) -- _8 Jul 2023_. +- NumPy 1.24.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.4)) -- _26 Jun 2023_. +- NumPy 1.25.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.25.0)) -- _17 Jun 2023_. +- NumPy 1.24.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.3)) -- _22 Apr 2023_. +- NumPy 1.24.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.2)) -- _5 Feb 2023_. +- NumPy 1.24.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.1)) -- _26 Dec 2022_. +- NumPy 1.24.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.0)) -- _18 Dec 2022_. +- NumPy 1.23.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.5)) -- _19 Nov 2022_. +- NumPy 1.23.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.4)) -- _12 Oct 2022_. +- NumPy 1.23.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.3)) -- _9 Sep 2022_. +- NumPy 1.23.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.2)) -- _14 Aug 2022_. +- NumPy 1.23.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.1)) -- _8 Jul 2022_. +- NumPy 1.23.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.0)) -- _22 Jun 2022_. +- NumPy 1.22.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.4)) -- _20 May 2022_. +- NumPy 1.21.6 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.21.6)) -- _12 Apr 2022_. +- NumPy 1.22.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.3)) -- _7 Mar 2022_. +- NumPy 1.22.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.2)) -- _3 Feb 2022_. +- NumPy 1.22.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.1)) -- _14 Jan 2022_. +- NumPy 1.22.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.0)) -- _31 Dec 2021_. +- NumPy 1.21.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.21.5)) -- _19 Dec 2021_. +- NumPy 1.21.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.21.0)) -- _22 Jun 2021_. +- NumPy 1.20.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.20.3)) -- _10 May 2021_. +- NumPy 1.20.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.20.0)) -- _30 Jan 2021_. +- NumPy 1.19.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.19.5)) -- _5 Jan 2021_. +- NumPy 1.19.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.19.0)) -- _20 Jun 2020_. +- NumPy 1.18.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.18.4)) -- _3 May 2020_. +- NumPy 1.17.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.17.5)) -- _1 Jan 2020_. +- NumPy 1.18.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.18.0)) -- _22 Dec 2019_. +- NumPy 1.17.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.17.0)) -- _26 Jul 2019_. +- NumPy 1.16.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.16.0)) -- _14 Jan 2019_. +- NumPy 1.15.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.15.0)) -- _23 Jul 2018_. +- NumPy 1.14.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.14.0)) -- _7 Jan 2018_. diff --git a/content/es/press-kit.md b/content/es/press-kit.md new file mode 100644 index 0000000000..9b66e3e683 --- /dev/null +++ b/content/es/press-kit.md @@ -0,0 +1,8 @@ +--- +title: Kit de prensa +sidebar: false +--- + +Nos gustaría facilitarte el trabajo para incluir la identidad del proyecto NumPy en tu próximo documento académico, material de curso o presentación. + +[Aquí](https://github.com/numpy/numpy/tree/main/branding/logo) encontrarás varias versiones en alta resolución del logo de NumPy. Ten en cuenta que al utilizar los recursos de numpy.org, aceptas el [Código de Conducta de NumPy](/code-of-conduct). diff --git a/content/es/privacy.md b/content/es/privacy.md new file mode 100644 index 0000000000..4b62b24be2 --- /dev/null +++ b/content/es/privacy.md @@ -0,0 +1,8 @@ +--- +title: Política de Privacidad +sidebar: false +--- + +**numpy.org** está operado por [NumFOCUS, Inc.](https://numfocus.org), el patrocinador fiscal del proyecto NumPy. Para ver la Política de Privacidad de este sitio web, por favor dirígete a https://numfocus.org/privacy-policy. + +Si tienes alguna pregunta sobre la política o la recolección, uso y divulgación de datos de NumFOCUS, por favor ponte en contacto con el personal de NumFOCUS en privacy@numfocus.org. diff --git a/content/es/report-handling-manual.md b/content/es/report-handling-manual.md new file mode 100644 index 0000000000..5586668cba --- /dev/null +++ b/content/es/report-handling-manual.md @@ -0,0 +1,95 @@ +--- +title: NumPy Code of Conduct - How to follow up on a report +sidebar: false +--- + +This is the manual followed by NumPy’s Code of Conduct Committee. It’s used when we respond to an issue to make sure we’re consistent and fair. + +Enforcing the [Code of Conduct](/code-of-conduct) impacts our community today and for the future. It’s an action that we do not take lightly. When reviewing enforcement measures, the Code of Conduct Committee will keep the following values and guidelines in mind: + +* Act in a personal manner rather than impersonal. The Committee can engage the parties to understand the situation while respecting the privacy and any necessary confidentiality of reporters. However, sometimes it is necessary to communicate with one or more individuals directly: the Committee’s goal is to improve the health of our community rather than only produce a formal decision. +* Emphasize empathy for individuals rather than judging behavior, avoiding binary labels of “good” and “bad/evil”. Overt, clear-cut aggression and harassment exist, and we will address them firmly. But many scenarios that can prove challenging to resolve are those where normal disagreements devolve into unhelpful or harmful behavior from multiple parties. Understanding the full context and finding a path that re-engages all is hard, but ultimately the most productive for our community. +* We understand that email is a difficult medium and can be isolating. Receiving criticism over email, without personal contact, can be particularly painful. This makes it especially important to keep an atmosphere of open-minded respect for the views of others. It also means that we must be transparent in our actions, and that we will do everything in our power to make sure that all our members are treated fairly and with sympathy. +* Discrimination can be subtle and it can be unconscious. It can show itself as unfairness and hostility in otherwise ordinary interactions. We know that this does occur, and we will take care to look out for it. We would very much like to hear from you if you feel you have been treated unfairly, and we will use these procedures to make sure that your complaint is heard and addressed. +* Help increase engagement in good discussion practice: try to identify where discussion may have broken down, and provide actionable information, pointers, and resources that can lead to positive change on these points. +* Be mindful of the needs of new members: provide them with explicit support and consideration, with the aim of increasing participation from underrepresented groups in particular. +* Individuals come from different cultural backgrounds and native languages. Try to identify any honest misunderstandings caused by a non-native speaker and help them understand the issue and what they can change to avoid causing offence. Complex discussion in a foreign language can be very intimidating, and we want to grow our diversity also across nationalities and cultures. + + +## Mediation + +Voluntary informal mediation is a tool at our disposal. In contexts such as when two or more parties have all escalated to the point of inappropriate behavior (something sadly common in human conflict), it may be useful to facilitate a mediation process. This is only an example: the Committee can consider mediation in any case, mindful that the process is meant to be strictly voluntary and no party can be pressured to participate. If the Committee suggests mediation, it should: + +* Find a candidate who can serve as a mediator. +* Obtain the agreement of the reporter(s). The reporter(s) have complete freedom to decline the mediation idea or to propose an alternate mediator. +* Obtain the agreement of the reported person(s). +* Settle on the mediator: while parties can propose a different mediator than the suggested candidate, only if a common agreement is reached on all terms can the process move forward. +* Establish a timeline for mediation to complete, ideally within two weeks. + +The mediator will engage with all the parties and seek a resolution that is satisfactory to all. Upon completion, the mediator will provide a report (vetted by all parties to the process) to the Committee, with recommendations on further steps. The Committee will then evaluate these results (whether a satisfactory resolution was achieved or not) and decide on any additional action deemed necessary. + + +## How the Committee will respond to reports + +When the Committee (or a Committee member) receives a report, they will first determine whether the report is about a clear and severe breach (as defined below). If so, immediate action needs to be taken in addition to the regular report handling process. + + +## Clear and severe breach actions + +We know that it is painfully common for internet communication to start at or devolve into obvious and flagrant abuse. We will deal quickly with clear and severe breaches like personal threats, violent, sexist or racist language. + +When a member of the Code of Conduct Committee becomes aware of a clear and severe breach, they will do the following: + +* Immediately disconnect the originator from all NumPy communication channels. +* Reply to the reporter that their report has been received and that the originator has been disconnected. +* In every case, the moderator should make a reasonable effort to contact the originator, and tell them specifically how their language or actions qualify as a “clear and severe breach”. The moderator should also say that, if the originator believes this is unfair or they want to be reconnected to NumPy, they have the right to ask for a review, as below, by the Code of Conduct Committee. The moderator should copy this explanation to the Code of Conduct Committee. +* The Code of Conduct Committee will formally review and sign off on all cases where this mechanism has been applied to make sure it is not being used to control ordinary heated disagreement. + + +## Report handling + +When a report is sent to the Committee they will immediately reply to the reporter to confirm receipt. This reply must be sent within 72 hours, and the group should strive to respond much quicker than that. + +If a report doesn’t contain enough information, the Committee will obtain all relevant data before acting. The Committee is empowered to act on the Steering Council’s behalf in contacting any individuals involved to get a more complete account of events. + +The Committee will then review the incident and determine, to the best of their ability: + +* What happened. +* Whether this event constitutes a Code of Conduct violation. +* Who are the responsible party(ies). +* Whether this is an ongoing situation, and there is a threat to anyone’s physical safety. + +This information will be collected in writing, and whenever possible the group’s deliberations will be recorded and retained (i.e. chat transcripts, email discussions, recorded conference calls, summaries of voice conversations, etc). + +It is important to retain an archive of all activities of this Committee to ensure consistency in behavior and provide institutional memory for the project. To assist in this, the default channel of discussion for this Committee will be a private mailing list accessible to current and future members of the Committee as well as members of the Steering Council upon justified request. If the Committee finds the need to use off-list communications (e.g. phone calls for early/rapid response), it should in all cases summarize these back to the list so there’s a good record of the process. + +The Code of Conduct Committee should aim to have a resolution agreed upon within two weeks. In the event that a resolution can’t be determined in that time, the Committee will respond to the reporter(s) with an update and projected timeline for resolution. + + +## Resolutions + +The Committee must agree on a resolution by consensus. If the group cannot reach consensus and deadlocks for over a week, the group will turn the matter over to the Steering Council for resolution. + +Possible responses may include: + +* Taking no further action: + - if we determine no violations have occurred; + - if the matter has been resolved publicly while the Committee was considering responses. +* Coordinating voluntary mediation: if all involved parties agree, the Committee may facilitate a mediation process as detailed above. +* Remind publicly, and point out that some behavior/actions/language have been judged inappropriate and why in the current context, or can but hurtful to some people, requesting the community to self-adjust. +* A private reprimand from the Committee to the individual(s) involved. In this case, the group chair will deliver that reprimand to the individual(s) over email, cc’ing the group. +* A public reprimand. In this case, the Committee chair will deliver that reprimand in the same venue that the violation occurred, within the limits of practicality. E.g., the original mailing list for an email violation, but for a chat room discussion where the person/context may be gone, they can be reached by other means. The group may choose to publish this message elsewhere for documentation purposes. +* A request for a public or private apology, assuming the reporter agrees to this idea: they may at their discretion refuse further contact with the violator. The chair will deliver this request. The Committee may, if it chooses, attach “strings” to this request: for example, the group may ask a violator to apologize in order to retain one’s membership on a mailing list. +* A “mutually agreed upon hiatus” where the Committee asks the individual to temporarily refrain from community participation. If the individual chooses not to take a temporary break voluntarily, the Committee may issue a “mandatory cooling off period”. +* A permanent or temporary ban from some or all NumPy spaces (mailing lists, gitter.im, etc.). The group will maintain records of all such bans so that they may be reviewed in the future or otherwise maintained. + +Once a resolution is agreed upon, but before it is enacted, the Committee will contact the original reporter and any other affected parties and explain the proposed resolution. The Committee will ask if this resolution is acceptable, and must note feedback for the record. + +Finally, the Committee will make a report to the NumPy Steering Council (as well as the NumPy core team in the event of an ongoing resolution, such as a ban). + +The Committee will never publicly discuss the issue; all public statements will be made by the chair of the Code of Conduct Committee or the NumPy Steering Council. + + +## Conflicts of Interest + +In the event of any conflict of interest, a Committee member must immediately notify the other members, and recuse themselves if necessary. diff --git a/content/es/tabcontents.yaml b/content/es/tabcontents.yaml new file mode 100644 index 0000000000..7eabca58e1 --- /dev/null +++ b/content/es/tabcontents.yaml @@ -0,0 +1,373 @@ +params: + machinelearning: + paras: + - + para1: NumPy constituye la base de potentes bibliotecas de aprendizaje automático como [scikit-learn](https://scikit-learn.org) y [SciPy](https://www.scipy.org). A medida que crece el aprendizaje automático, también lo hace la lista de bibliotecas basadas en NumPy. Las capacidades de aprendizaje profundo de [TensorFlow](https://www.tensorflow.org) tienen amplias aplicaciones— entre ellas el reconocimiento de voz e imágenes, las aplicaciones basadas en texto, el análisis de series temporales y la detección de vídeo. [PyTorch](https://pytorch.org), otra biblioteca de aprendizaje profundo, es popular entre los investigadores de visión por ordenador y procesamiento del lenguaje natural. [MXNet](https://github.com/apache/incubator-mxnet) es otro paquete de IA que proporciona modelos y plantillas para el aprendizaje profundo. + para2: Las técnicas estadísticas denominadas métodos [ensemble](https://towardsdatascience.com/ensemble-methods-bagging-boosting-and-stacking-c9214a10a205), como binning, bagging, stacking y boosting, se encuentran entre los algoritmos de ML implementados por herramientas como [XGBoost](https://github.com/dmlc/xgboost), [LightGBM](https://lightgbm.readthedocs.io/en/latest/) y [CatBoost](https://catboost.ai) — uno de los motores de inferencia más rápidos. [Yellowbrick](https://www.scikit-yb.org/en/latest/) y [Eli5](https://eli5.readthedocs.io/en/latest/) ofrecen visualizaciones de aprendizaje automático. + arraylibraries: + intro: + - + text: La API de NumPy es el punto de partida cuando se escriben bibliotecas para explotar hardware innovador, crear tipos de matrices especializadas o añadir capacidades más allá de lo que NumPy proporciona. + headers: + - + text: Biblioteca de matrices + - + text: Capacidades y áreas de aplicación + libraries: + - + title: Dask + text: Matrices distribuidas y paralelismo avanzado para análisis, que permiten un rendimiento a escala. + img: /images/content_images/arlib/dask.png + alttext: Dask + url: https://dask.org/ + - + title: CuPy + text: Biblioteca de matrices compatible con NumPy para cálculo acelerado en la GPU con Python. + img: /images/content_images/arlib/cupy.png + alttext: CuPy + url: https://cupy.chainer.org + - + title: JAX + text: "Transformaciones componibles de programas NumPy: diferenciar, vectorizar, compilación justo-a-tiempo a GPU/TPU." + img: /images/content_images/arlib/jax_logo_250px.png + alttext: JAX + url: https://github.com/google/jax + - + title: Xarray + text: Matrices multidimensionales indexadas y etiquetadas para análisis y visualización avanzados + img: /images/content_images/arlib/xarray.png + alttext: xarray + url: https://xarray.pydata.org/en/stable/index.html + - + title: Sparse + text: Biblioteca de matrices dispersas compatible con NumPy que se integra con el álgebra lineal dispersa de Dask y SciPy. + img: /images/content_images/arlib/sparse.png + alttext: sparse + url: https://sparse.pydata.org/en/latest/ + - + title: PyTorch + text: Marco de aprendizaje profundo que acelera el camino desde la creación de prototipos de investigación hasta la implantación en producción. + img: /images/content_images/arlib/pytorch-logo-dark.svg + alttext: PyTorch + url: https://pytorch.org/ + - + title: TensorFlow + text: Una plataforma integral de aprendizaje automático para crear y desplegar fácilmente aplicaciones basadas en ML. + img: /images/content_images/arlib/tensorflow-logo.svg + alttext: TensorFlow + url: https://www.tensorflow.org + - + title: MXNet + text: Marco de aprendizaje profundo adecuado para la creación flexible de prototipos de investigación y la producción. + img: /images/content_images/arlib/mxnet_logo.png + alttext: MXNet + url: https://mxnet.apache.org/ + - + title: Arrow + text: Plataforma de desarrollo multilenguaje para datos y análisis columnares en memoria. + img: /images/content_images/arlib/arrow.png + alttext: arrow + url: https://github.com/apache/arrow + - + title: xtensor + text: Matrices multidimensionales con emisión y computación de llamada-bajo-demanda para el análisis numérico. + img: /images/content_images/arlib/xtensor.png + alttext: xtensor + url: https://github.com/xtensor-stack/xtensor-python + - + title: Awkward Array + text: Manipula datos tipo JSON con modismos similares a los de NumPy. + img: /images/content_images/arlib/awkward.svg + alttext: awkward + url: https://awkward-array.org/ + - + title: uarray + text: Sistema de backend de Python que desvincula la API de la implementación; unumpy proporciona una API de NumPy. + img: /images/content_images/arlib/uarray.png + alttext: uarray + url: https://uarray.org/en/latest/ + scientificdomains: + intro: + - + text: Casi todos los científicos que trabajan en Python recurren a la potencia de NumPy. + - + text: "NumPy lleva la potencia de cálculo de lenguajes como C y Fortran a Python, un lenguaje mucho más fácil de aprender y utilizar. Con esta potencia viene la sencillez: una solución en NumPy suele ser clara y elegante." + libraries: + - + title: Quantum Computing + alttext: A computer chip. + img: /images/content_images/sc_dom_img/quantum_computing.svg + links: + - + url: http://qutip.org + label: QuTiP + - + url: https://pyquil-docs.rigetti.com/en/stable + label: PyQuil + - + url: https://qiskit.org + label: Qiskit + - + url: https://pennylane.ai + label: PennyLane + - + title: Statistical Computing + alttext: A line graph with the line moving up. + img: /images/content_images/sc_dom_img/statistical_computing.svg + links: + - + url: https://pandas.pydata.org/ + label: Pandas + - + url: https://www.statsmodels.org/ + label: statsmodels + - + url: https://xarray.pydata.org/en/stable/ + label: Xarray + - + url: https://seaborn.pydata.org/ + label: Seaborn + - + title: Signal Processing + alttext: A bar chart with positive and negative values. + img: /images/content_images/sc_dom_img/signal_processing.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://pywavelets.readthedocs.io/ + label: PyWavelets + - + url: https://python-control.org/ + label: python-control + - + url: https://hyperspy.org/ + label: HyperSpy + - + title: Image Processing + alttext: An photograph of the mountains. + img: /images/content_images/sc_dom_img/image_processing.svg + links: + - + url: https://scikit-image.org/ + label: Scikit-image + - + url: https://opencv.org/ + label: OpenCV + - + url: https://mahotas.rtfd.io/ + label: Mahotas + - + title: Graphs and Networks + alttext: A simple graph. + img: /images/content_images/sc_dom_img/sd6.svg + links: + - + url: https://networkx.org/ + label: NetworkX + - + url: https://graph-tool.skewed.de/ + label: graph-tool + - + url: https://igraph.org/python/ + label: igraph + - + url: https://pygsp.rtfd.io/ + label: PyGSP + - + title: Astronomy + alttext: A telescope. + img: /images/content_images/sc_dom_img/astronomy_processes.svg + links: + - + url: https://www.astropy.org/ + label: AstroPy + - + url: https://sunpy.org/ + label: SunPy + - + url: https://spacepy.github.io/ + label: SpacePy + - + title: Cognitive Psychology + alttext: A human head with gears. + img: /images/content_images/sc_dom_img/cognitive_psychology.svg + links: + - + url: https://www.psychopy.org/ + label: PsychoPy + - + title: Bioinformatics + alttext: A strand of DNA. + img: /images/content_images/sc_dom_img/bioinformatics.svg + links: + - + url: https://biopython.org/ + label: BioPython + - + url: http://scikit-bio.org/ + label: Scikit-Bio + - + url: https://github.com/openvax/pyensembl + label: PyEnsembl + - + url: http://etetoolkit.org/ + label: ETE + - + title: Bayesian Inference + alttext: A graph with a bell-shaped curve. + img: /images/content_images/sc_dom_img/bayesian_inference.svg + links: + - + url: https://pystan.readthedocs.io/en/latest/ + label: PyStan + - + url: https://docs.pymc.io/ + label: PyMC3 + - + url: https://arviz-devs.github.io/arviz/ + label: ArviZ + - + url: https://emcee.readthedocs.io/ + label: emcee + - + title: Mathematical Analysis + alttext: Four mathematical symbols. + img: /images/content_images/sc_dom_img/mathematical_analysis.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://www.sympy.org/ + label: SymPy + - + url: https://www.cvxpy.org/ + label: cvxpy + - + url: https://fenicsproject.org/ + label: FEniCS + - + title: Chemistry + alttext: A test tube. + img: /images/content_images/sc_dom_img/chemistry.svg + links: + - + url: https://cantera.org/ + label: Cantera + - + url: https://www.mdanalysis.org/ + label: MDAnalysis + - + url: https://github.com/rdkit/rdkit + label: RDKit + - + url: https://www.pybamm.org/ + label: PyBaMM + - + title: Geoscience + alttext: The Earth. + img: /images/content_images/sc_dom_img/geoscience.svg + links: + - + url: https://pangeo.io/ + label: Pangeo + - + url: https://simpeg.xyz/ + label: Simpeg + - + url: https://github.com/obspy/obspy/wiki + label: ObsPy + - + url: https://www.fatiando.org/ + label: Fatiando a Terra + - + title: Geographic Processing + alttext: A map. + img: /images/content_images/sc_dom_img/GIS.svg + links: + - + url: https://shapely.readthedocs.io/ + label: Shapely + - + url: https://geopandas.org/ + label: GeoPandas + - + url: https://python-visualization.github.io/folium + label: Folium + - + title: Architecture & Engineering + alttext: A microprocessor development board. + img: /images/content_images/sc_dom_img/robotics.svg + links: + - + url: https://compas.dev/ + label: COMPAS + - + url: https://cityenergyanalyst.com/ + label: City Energy Analyst + - + url: https://nortikin.github.io/sverchok/ + label: Sverchok + datascience: + intro: "NumPy es el núcleo de un rico ecosistema de librerías de ciencia de datos. Un flujo de trabajo exploratorio típico de ciencia de datos podría verse así:" + image1: + - + img: /images/content_images/ds-landscape.png + alttext: Diagrama de las librerías de Python. Las cinco categorías son "Extraer, Transformar, Cargar", "Exploración de Datos", "Modelado de Datos", "Evaluación de Datos" y "Presentación de Datos". + image2: + - + img: /images/content_images/data-science.png + alttext: Diagrama de tres círculos superpuestos. Los círculos se denominan "Matemáticas", "Informática" y "Conocimientos Especializados". En el centro del diagrama, con los tres círculos superpuestos, hay un área denominada "Ciencia de datos". + examples: + - + text: "Extraer, Transformar, Cargar: [Pandas](https://pandas.pydata.org), [Intake](https://intake.readthedocs.io), [PyJanitor](https://pyjanitor.readthedocs.io/)" + - + text: "Análisis Exploratorio: [Jupyter](https://jupyter.org), [Seaborn](https://seaborn.pydata.org), [Matplotlib](https://matplotlib.org), [Altair](https://altair-viz.github.io)" + - + text: "Modelado y evaluación: [scikit-learn](https://scikit-learn.org), [statsmodels](https://www.statsmodels.org/stable/index.html), [PyMC3](https://docs.pymc.io), [spaCy](https://spacy.io)" + - + text: "Informes en paneles: [Dash](https://plotly.com/dash), [Panel](https://panel.holoviz.org), [Voila](https://github.com/voila-dashboards/voila)" + content: + - + text: Para grandes volúmenes de datos, [Dask](https://dask.org) y [Ray](https://ray.io/) están diseñados para escalarse. Las implementaciones estables se basan en el versionado de datos ([DVC](https://dvc.org)), rastreo de experimentos ([MLFlow](https://mlflow.org)), y automatización del flujo de trabajo ([Airflow](https://airflow.apache.org), [Dagster](https://dagster.io) y [Prefect](https://www.prefect.io)). + visualization: + images: + - + url: https://www.fusioncharts.com/blog/best-python-data-visualization-libraries + img: /images/content_images/v_matplotlib.png + alttext: Un diagrama de flujo hecho en matplotlib + - + url: https://github.com/yhat/ggpy + img: /images/content_images/v_ggpy.png + alttext: Un diagrama de dispersión hecho en ggpy + - + url: https://www.journaldev.com/19692/python-plotly-tutorial + img: /images/content_images/v_plotly.png + alttext: Un diagrama de caja hecho en plotly + - + url: https://altair-viz.github.io/gallery/streamgraph.html + img: /images/content_images/v_altair.png + alttext: Un diagrama de flujo hecho en altair + - + url: https://seaborn.pydata.org + img: /images/content_images/v_seaborn.png + alttext: Un gráfico de pares de dos tipos de gráficos, un gráfico de trazado y un gráfico de frecuencias hecho en seaborn + - + url: https://docs.pyvista.org/examples/index.html + img: /images/content_images/v_pyvista.png + alttext: Un renderizado de volumen 3D realizado en PyVista. + - + url: https://napari.org + img: /images/content_images/v_napari.png + alttext: Una imagen multidimensional hecha en napari. + - + url: https://vispy.org/gallery/index.html + img: /images/content_images/v_vispy.png + alttext: Un diagrama de Voronoi hecho en vispy. + content: + - + text: NumPy es un componente esencial en el floreciente [panorama de visualización de Python](https://pyviz.org/overviews/index.html), que incluye [Matplotlib](https://matplotlib.org), [Seaborn](https://seaborn.pydata.org), [Plotly](https://plot.ly), [Altair](https://altair-viz.github.io), [Bokeh](https://docs.bokeh.org/en/latest/), [Holoviz](https://holoviz.org), [Vispy](http://vispy.org), [Napari](https://github.com/napari/napari), y [PyVista](https://github.com/pyvista/pyvista), por nombrar algunos. + - + text: El procesamiento acelerado de matrices de gran tamaño de NumPy permite a los investigadores visualizar conjuntos de datos mucho mayores a los que el Python nativo podría manejar. diff --git a/content/es/teams.md b/content/es/teams.md new file mode 100644 index 0000000000..ee292266a6 --- /dev/null +++ b/content/es/teams.md @@ -0,0 +1,22 @@ +--- +title: Equipos de NumPy +sidebar: false +--- + +Somos un equipo internacional en una misión para apoyar a las comunidades científicas y de investigaciones alrededor del mundo, mediante la construcción de software de código abierto de calidad. ¡[Únete a nosotros]({{< relref "/contribute" >}})! + +{{< include-html "static/gallery/maintainers.html" >}} + +{{< include-html "static/gallery/docs-team.html" >}} + +{{< include-html "static/gallery/web-team.html" >}} + +{{< include-html "static/gallery/triage-team.html" >}} + +{{< include-html "static/gallery/survey-team.html" >}} + +{{< include-html "static/gallery/emeritus-maintainers.html" >}} + +# Gobernanza + +Para la lista de personas del Consejo Directivo, por favor ve [aquí](https://numpy.org/about/). diff --git a/content/es/user-survey-2020.md b/content/es/user-survey-2020.md new file mode 100644 index 0000000000..d6d502b29a --- /dev/null +++ b/content/es/user-survey-2020.md @@ -0,0 +1,18 @@ +--- +title: ENCUESTA DE LA COMUNIDAD NUMPY 2020 +sidebar: false +--- + +En 2020, el equipo de encuestas de NumPy, en asociación con estudiantes y profesores de un curso de Maestría en Metodología de Encuestas organizado conjuntamente por la Universidad de Michigan y la Universidad de Maryland, llevaron a cabo la primera encuesta oficial de la comunidad NumPy. Más de 1,200 usuarios de 75 países participaron para ayudarnos a proyectar un panorama de la comunidad NumPy y expresaron sus pensamientos sobre el futuro del proyecto. + +{{< figure src="/surveys/NumPy_usersurvey_2020_report_cover.png" class="fig-left" alt="Cover page of the 2020 NumPy user survey report, titled 'NumPy Community Survey 2020 - results'" width="250">}} +src = '/surveys/NumPy_usersurvey_2020_report_cover.png' alt = 'Cover page of the 2020 NumPy user survey report, titled "NumPy Community Survey 2020 - results"' width = '250' +{{< /figure >}} + +**[Descarga el informe](/surveys/NumPy_usersurvey_2020_report.pdf)** para ver a detalle los resultados de la encuesta. + + +Para los puntos destacados, echa un vistazo a **[esta infografía](https://github.com/numpy/numpy-surveys/blob/master/images/2020NumPysurveyresults_community_infographic.pdf)**. + +¿Listo para una inmersión profunda? Visita **https://numpy.org/user-survey-2020-details/**. + diff --git a/content/es/user-surveys.md b/content/es/user-surveys.md new file mode 100644 index 0000000000..ef8467358d --- /dev/null +++ b/content/es/user-surveys.md @@ -0,0 +1,10 @@ +--- +title: ENCUESTAS DE USUARIOS DE NUMPY +sidebar: false +--- + +**2020** El equipo de encuestas de NumPy, en asociación con estudiantes y profesores de la Universidad de Michigan y de la Universidad de Maryland, realizó la primera encuesta oficial de la comunidad de NumPy. Encuentra los resultados de la encuesta [aquí](https://numpy.org/user-survey-2020/). + +**2021** Los datos recolectados están siendo analizados actualmente. + +Si tienes alguna pregunta o sugerencia sobre las encuestas pasadas o futuras, por favor abre una propuesta [aquí](https://github.com/numpy/numpy-surveys/issues). diff --git a/content/ja/404.md b/content/ja/404.md index 855d17f885..8e4db85255 100644 --- a/content/ja/404.md +++ b/content/ja/404.md @@ -5,4 +5,4 @@ sidebar: false おっとっと！ 間違った所にアクセスしているようです。 -何かがここにページがあるべきだと思ったら、GitHub で [issue](https://github.com/numpy/numpy.org/issues) を作成してください。 +何かがここにページがあるべきだと思ったら、GitHub で [issue](https://github.com/numpy/numpy.org/issues) を作成してください。 diff --git a/content/ja/_index.md b/content/ja/_index.md new file mode 100644 index 0000000000..be88f9e642 --- /dev/null +++ b/content/ja/_index.md @@ -0,0 +1,49 @@ +--- +title: null +--- + +{{< grid columns="1 2 2 3" >}} + +[[item]] +type = 'card' +title = 'Powerful N-dimensional arrays' +body = ''' +Fast and versatile, the NumPy vectorization, indexing, and broadcasting concepts are the de-facto standards of array computing today. +''' + +[[item]] +type = 'card' +title = 'Numerical computing tools' +body = ''' +NumPy offers comprehensive mathematical functions, random number generators, linear algebra routines, Fourier transforms, and more. +''' + +[[item]] +type = 'card' +title = 'Open source' +body = ''' +Distributed under a liberal [BSD license](https://github.com/numpy/numpy/blob/main/LICENSE.txt), NumPy is developed and maintained [publicly on GitHub](https://github.com/numpy/numpy) by a vibrant, responsive, and diverse [community](/community). +''' + +[[item]] +type = 'card' +title = 'Interoperable' +body = ''' +NumPy supports a wide range of hardware and computing platforms, and plays well with distributed, GPU, and sparse array libraries. +''' + +[[item]] +type = 'card' +title = 'Performant' +body = ''' +The core of NumPy is well-optimized C code. Enjoy the flexibility of Python with the speed of compiled code. +''' + +[[item]] +type = 'card' +title = 'Easy to use' +body = ''' +NumPy's high level syntax makes it accessible and productive for programmers from any background or experience level. +''' + +{{< /grid>}} diff --git a/content/ja/about.md b/content/ja/about.md index 58909a1333..6508ca265d 100644 --- a/content/ja/about.md +++ b/content/ja/about.md @@ -3,16 +3,14 @@ title: 私たちについて sidebar: false --- -_このページでは、NumPyのプロジェクトとそれを支えるコミュニティについて説明します。_ - -NumPy は、Python で数値計算を可能にするためのオープンソースプロジェクトです。 NumPyは、NumericやNumarrayといった初期のライブラリのコードをもとに、2005年から開発が開始されました。 NumPyは完全にオープンソースなソフトウェアであり、[修正BSD ライセンス](https://github.com/numpy/numpy/blob/main/LICENSE.txt) の条項の下で、すべての人が利用可能です。 +NumPy は、Python で数値計算を可能にするためのオープンソースプロジェクトです。 NumPyは、NumericやNumarrayといった初期のライブラリのコードをもとに、2005年から開発が開始されました。 NumPyは完全にオープンソースなソフトウェアです。 そして、NumPyは[修正BSD ライセンス](https://github.com/numpy/numpy/blob/main/LICENSE.txt) の条項の下で、すべての人が利用可能です。 NumPy は 、NumPyコミュニティやより広範な科学計算用Python コミュニティとの合意のもと、GitHub 上でオープンに開発されています。 NumPyのガバナンス方法の詳細については、 [Governance Document](https://www.numpy.org/devdocs/dev/governance/index.html) をご覧ください。 ## 運営委員会 -NumPy運営委員会の役割は、NumPyのコミュニティと協力しサポートすることを通じて、技術的にもコミュニティ的にも長期的にNumPyプロジェクトを良い状態に保つことです。 NumPy運営委員会は現在以下のメンバーで構成されています (アルファベット順、姓で): +Numpy運営委員会はこのプロジェクトの管理組織です。 その役割は、Numpy コミュニティと協力し、Numpyのソフトウェアサービスを確実にユーザに提供することです。 ソフトウェアパッケージとコミュニティの両方において、プロジェクトの長期的な持続可能性を保っていきます。 NumPy運営委員会は現在以下のメンバーで構成されています (姓のアルファベット順): - Sebastian Berg - Ralf Gommers @@ -20,33 +18,47 @@ NumPy運営委員会の役割は、NumPyのコミュニティと協力しサポ - Stephan Hoyer - Inessa Pawson - Matti Picus -- Stéfan van der Walt -- Melissa Weber Mendonça +- Stéfan van der Walt +- Melissa Weber Mendonça - Eric Wieser -終身名誉委員 +過去のメンバー - Alex Griffing (2015-2017) - Allan Haldane (2015-2021) - Marten van Kerkwijk (2017-2019) -- Travis Oliphant (project founder, 2005-2012) +- Travis Oliphant (プロジェクト創設者, 2005-2012) - Nathaniel Smith (2012-2021) - Julian Taylor (2013-2021) -- Jaime Fernández del Río (2014-2021) +- Jaime Fernández del Río (2014-2021) - Pauli Virtanen (2008-2021) +Numpy運営委員会に連絡するには、numpy-team@googlegroups.comまでメールしてください。 + ## チーム -NumPy プロジェクトは拡大しているため、いくつかのチームが設置されています。 +Numpy プロジェクトのコアメンバーは、プロジェクトへの貢献の方法の多様化に積極的に取り組んでいます。
Numpyには現在以下のチームがあります: -- コード +- 開発 - ドキュメント -- ウェブサイト - トリアージ +- ウェブサイト +- 調査 +- 翻訳 +- スプリントのメンター +- 最適化 - 資金と助成金 個々のチームメンバーについては、 [チーム](/teams/) のページを参照してください。 +## NumFOCUSサブ委員会 + +- Charles Harris +- Ralf Gommers +- Melissa Weber Mendonça +- Sebastian Berg +- 外部メンバー: Thomas Caswell + ## スポンサー情報 NumPyは以下の団体から直接資金援助を受けています。 @@ -56,6 +68,11 @@ NumPyは以下の団体から直接資金援助を受けています。 ## パートナー団体 パートナー団体は、NumPyへの開発を仕事の一つとして、社員を雇っている団体です。 現在のパートナー団体としては、下記の通りです。 + +- カルフォルニア大学 バークレー校 (Stéfan van der Walt) +- Quansight (Nathan Goldbaum, Ralf Gommers, Matti Picus, Melissa Weber Mendonça) +- NVIDIA (Sebastian Berg) + {{< partners >}} @@ -68,4 +85,6 @@ NumPy は NumFOCUS にスポンサーされたプロジェクトであり、米 NumPy への寄付は [NumFOCUS](https://numfocus.org) によって管理されています。 米国の寄付提供者の場合、その人の寄付は法律によって定められる範囲で免税されます。 但し、他の寄付と同様に、あなたはあなたの税務状況について、あなたの税務担当と相談する必要があることを忘れないで下さい。 NumPyの運営委員会は、受け取った資金をどのように使えば良いかを検討し、使用する方法について決定します. NumPyに関する技術とインフラの投資の優先順位に関しては、[NumPy Roadmap](https://www.numpy.org/neps/index.html#roadmap) に記載されています。 -{{< numfocus >}} + +{{}} + diff --git a/content/ja/arraycomputing.md b/content/ja/arraycomputing.md index fd7a74d380..7713e7e0f2 100644 --- a/content/ja/arraycomputing.md +++ b/content/ja/arraycomputing.md @@ -9,7 +9,7 @@ sidebar: false **Num**erical **Py**thon: NumPyは、Pythonにおけるデファクトスタンダードなライブラリであり、大規模な多次元配列や行列、そして、それらの配列を処理する様々な分野の数学ルーチンをサポートしています。 -2006年にNumPyが発表されてから、2008年にPandasが登場し、その後、数年間にいくつかの配列演算関連のライブラリが次々と現れるようになりました。そこから配列演算界隈は盛り上がり始めました。 これらの新しい配列演算ライブラリの多くは、NumPyの機能や能力を模倣しており、機械学習や人工知能向けの新しいアルゴリズムや機能を持っています。 +2006年にNumPyが発表されてから、2008年にPandasが登場し、その後、数年間にいくつかの配列演算関連のライブラリが次々と現れるようになりました。 これらの新しい配列演算ライブラリの多くは、NumPyの機能や能力を模倣しており、機械学習や人工知能向けの新しいアルゴリズムや機能を持っています。 black hole image}} +src = '/images/content_images/cs/blackhole.jpg' title = 'Black Hole M87' alt = 'black hole image' attribution = '(Image Credits: Event Horizon Telescope Collaboration)' attributionlink = 'https://www.jpl.nasa.gov/images/universe/20190410/blackhole20190410.jpg' +{{< /figure >}} -
-

M87ブラックホールを画像化することは、見ることのできないものを、あえて見ようとするようなものです。

-
カリフォルニア工科大学 計算・数理学部のKatie Bouman助教授
-
+{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="Katie Bouman, * カリフォルニア工科大学の計算数理科学の助教授 *" +> }} M87ブラックホールを画像化することは、見ることのできないものを、あえて見ようとするようなものです。 +> +> {{< /blockquote >}} ## 地球大の望遠鏡 -[Event Horizon telescope(EHT)](https:/eventhorizontelescope.org)は、地球サイズの解析望遠鏡を形成する8台の地上型電波望遠鏡から成るシステムで、これまでに前例のない感度と解像度で宇宙を研究することができます。超長基線干渉法(VLBI) と呼ばれる手法を用いた巨大な仮想望遠鏡の角度分解能は、[20マイクロ秒][resolution]で、ニューヨークにある新聞をパリの歩道のカフェから読むのに十分な解像度です! +[Event Horizon telescope(EHT)](https:/eventhorizontelescope.org)は、地球サイズの解析望遠鏡を形成する8台の地上型電波望遠鏡から成るシステムで、これまでに前例のない感度と解像度で宇宙を研究することができます。 超長基線干渉法(VLBI) と呼ばれる手法を用いた巨大な仮想望遠鏡の角度分解能は、[20マイクロ秒][resolution]で、ニューヨークにある新聞をパリの歩道のカフェから読むのに十分な解像度です! ### 主な目標と結果 * **宇宙の新しい見方:** EHTの画期的な考え方の基礎が築かれたのは、100年前に [Sir Arthur Eddington][eddington]がアインシュタインの一般相対性理論に沿った最初の観測を実施したことが始まりでした。 -* **ブラックホール:** EHTは、おとめ座銀河団のメシエ87銀河 (M87) の中心にある、地球から約5500万光年の距離にある超巨大ブラックホールを観測しました。 その質量は、太陽の65億倍です。[100年以上](https://www.jpl.nasa.gov/news/news.php?feature=7385)に渡る研究が行われてもなお、これまでに視覚的にブラックホールを観測できたことはありませんでした。 +* **ブラックホール:** EHTは、おとめ座銀河団のメシエ87銀河 (M87) の中心にある、地球から約5500万光年の距離にある超巨大ブラックホールを観測しました。 その質量は、太陽の65億倍です。 [100年以上](https://www.jpl.nasa.gov/news/news.php?feature=7385)に渡る研究が行われてもなお、これまでに視覚的にブラックホールを観測できたことはありませんでした。 -* **観測と理論の比較:** 科学者たちの間で、アインシュタインの一般相対性理論から、重力による光の曲げや光の捕獲による影のような領域が観測できるのではないかと期待されていました。これはブラックホールの巨大な質量を測定するために利用することができます。 +* **観測と理論の比較:** 科学者たちの間で、アインシュタインの一般相対性理論から、重力による光の曲げや光の捕獲による影のような領域が観測できるのではないかと期待されていました。 これはブラックホールの巨大な質量を測定するために利用することができます。 ### 課題 * **大規模な計算** - EHTは膨大なデータ処理の課題を抱えていました。大気の位相変動は急速で、記録帯域の幅は大きく、望遠鏡はそれぞれ異なっていて地理的にも分散しています。 + EHTは膨大なデータ処理の課題を抱えていました。 大気の位相変動は急速で、記録帯域の幅は大きく、望遠鏡はそれぞれ異なっていて地理的にも分散しています。 * **大量のデータ** - EHTは一日で350テラバイトを超える観測データを生成し、ヘリウムで満たされたハードドライブに保存しています。この大量のデータとデータの複雑さを軽減することは非常に難しいことです。 + EHTは一日で350テラバイトを超える観測データを生成し、ヘリウムで満たされたハードドライブに保存しています。 この大量のデータとデータの複雑さを軽減することは非常に難しいことです。 * **よくわからないものを観測する** 今までに見たことのないものを見るのが研究の目標なら、どうやって科学者はその画像が正しいと確信することができるのでしょうか? -{{< figure src="/images/content_images/cs/dataprocessbh.png" class="csfigcaption" caption="**EHTのデータ処理パイプライン**" alt="data pipeline" align="middle" attr="(Diagram Credits: The Astrophysical Journal, Event Horizon Telescope Collaboration)" attrlink="https://iopscience.iop.org/article/10.3847/2041-8213/ab0c57" >}} +{{< figure >}} +src = '/images/content_images/cs/dataprocessbh.png' title = 'EHT Data Processing Pipeline' alt = 'data pipeline' align = 'center' attribution = '(Diagram Credits: The Astrophysical Journal, Event Horizon Telescope Collaboration)' attributionlink = 'https://iopscience.iop.org/article/10.3847/2041-8213/ab0c57' +{{< figure src="/images/content_images/cs/bh_numpy_role.png" class="fig-center" alt="role of numpy" caption="**ブラックホール画像化でNumPyが果たした役割**" >}} ## NumPyが果たした役割 -データに問題がある場合はどうなるでしょう？ あるいは、アルゴリズムが特定の仮定に あまりにも大きく依存しているかもしれません。もしあるパラメータを変更した場合、画像は大きく変化するのでしょうか？ +データに問題がある場合はどうなるでしょう？ あるいは、アルゴリズムが特定の仮定に あまりにも大きく依存しているかもしれません。 もしあるパラメータを変更した場合、画像は大きく変化するのでしょうか？ -EHTの共同研究では、最先端の画像再構成技術を使用して、それぞれのチームがデータを評価することによって、これらの課題に対処しました。それぞれのチームの解析結果が同じであることが証明されると、それらの結果を組み合わせることで、ブラックホール画像を得ることができました。 +EHTの共同研究では、最先端の画像再構成技術を使用して、それぞれのチームがデータを評価することによって、これらの課題に対処しました。 それぞれのチームの解析結果が同じであることが証明されると、それらの結果を組み合わせることで、ブラックホール画像を得ることができました。 彼らの研究は、共同のデータ解析を通じて科学を進歩させる、科学的なPythonエコシステムが果たす役割を如実に表しています。 -{{< figure src="/images/content_images/cs/bh_numpy_role.png" class="fig-center" alt="role of numpy" caption="**ブラックホール画像化でNumPyが果たした役割**" >}} +{{< figure >}} +{{< figure src="/images/content_images/cs/numpy_bh_benefits.png" class="fig-center" alt="numpy benefits" caption="**利用されたNumPyの主要機能**" >}} +{{< /figure >}} -例えば、 [`eht-imaging`][ehtim] というPython パッケージは VLBI データで画像の再構築をシミュレートし、実行するためのツールです。NumPyは、以下のソフトウェア依存関係チャートで示されているように、このパッケージで使用される配列データ処理の中核を担っています。 +例えば、 [`eht-imaging`][ehtim] というPython パッケージは VLBI データで画像の再構築をシミュレートし、実行するためのツールです。 NumPyは、以下のソフトウェア依存関係チャートで示されているように、このパッケージで使用される配列データ処理の中核を担っています。 -{{< figure src="/images/content_images/cs/ehtim_numpy.png" class="fig-center" alt="ehtim dependency map highlighting numpy" caption="**NumPyの中心としたehtimのソフトウェア依存図**" >}} +{{< figure >}} +src = '/images/content_images/cs/ehtim_numpy.png' alt = 'ehtim dependency map highlighting numpy' title = 'Software dependency chart of ehtim package highlighting NumPy' +{{< /figure >}} NumPyだけでなく、[SciPy](https://www.scipy.org)や[Pandas](https://pandas.io)などのパッケージもブラックホール画像化におけるデータ処理パイプラインに利用されています。 天文学の標準的なファイル形式や時間/座標変換 は[Astropy][astropy]で実装され、ブラックホールの最終画像の生成を含め、解析パイプライン全体でのデータ可視化には [Matplotlib][mpl]が利用されました。 ## まとめ -NumPyの中心的な機能である、効率的で適用性の高いn次元配列は、研究者が大規模な数値データを操作することを可能にし、世界で初めてのブラックホールの画像化の基礎を築きました。 アインシュタインの理論に素晴らしい視覚的証拠を与えたのは、科学の画期的な瞬間だといえます。 この科学的に偉大な達成には、技術的の飛躍的な進歩だけでなく、200人以上の科学者と世界で 最高の電波観測所の間での国際協力も寄与しました。革新的なアルゴリズムとデータ処理技術は、既存の天文学モデルを改良し、宇宙の謎を解き明かす助けになったといえます。 +NumPyの中心的な機能である、効率的で適用性の高いn次元配列は、研究者が大規模な数値データを操作することを可能にし、世界で初めてのブラックホールの画像化の基礎を築きました。 アインシュタインの理論に素晴らしい視覚的証拠を与えたのは、科学の画期的な瞬間だといえます。 この科学的に偉大な達成には、技術的の飛躍的な進歩だけでなく、200人以上の科学者と世界で 最高の電波観測所の間での国際協力も寄与しました。 革新的なアルゴリズムとデータ処理技術は、既存の天文学モデルを改良し、宇宙の謎を解き明かす助けになったといえます。 -{{< figure src="/images/content_images/cs/numpy_bh_benefits.png" class="fig-center" alt="numpy benefits" caption="**利用されたNumPyの主要機能**" >}} +{{< figure >}} +src = '/images/content_images/cs/numpy_bh_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} [resolution]: https://eventhorizontelescope.org/press-release-april-10-2019-astronomers-capture-first-image-black-hole diff --git a/content/ja/case-studies/cricket-analytics.md b/content/ja/case-studies/cricket-analytics.md index ec68246720..2e366ad8d1 100644 --- a/content/ja/case-studies/cricket-analytics.md +++ b/content/ja/case-studies/cricket-analytics.md @@ -4,28 +4,32 @@ sidebar: false --- {{< figure src="/images/content_images/cs/ipl-stadium.png" caption="** IPLT20、インド最大のクリケットフェスティバル**" alt="Indian Premier League Cricket cup and stadium" attr="*(Image credits: IPLT20 (cup and logo) & Akash Yadav (stadium))*" attrlink="https://unsplash.com/@aksh1802" >}} +src = '/images/content_images/cs/ipl-stadium.png' title = 'IPLT20, the biggest Cricket Festival in India' alt = 'Indian Premier League Cricket cup and stadium' attribution = '(Image credits: IPLT20 (cup and logo) & Akash Yadav (stadium))' attributionlink = 'https://unsplash.com/@aksh1802' +{{< /figure >}} -
-

観客のために競技をするのではなく、国のために競技するのです。

-
—M S Dhoni、 インディアンチームの元キャプテン、インターナショナル・クリケットプレイヤー、チェンナイ・スーパー・キングスのためにIPLでプレイ
-
+{{< blockquote cite="https://www.scoopwhoop.com/sports/ms-dhoni/" by="M S Dhoni, *世界的なクリケットプレイヤーであり、前世界王者、インドチームで活躍し、IPLにおけるチェンナイスーパー王者となった*" +> }} 観客のために競技しているのではなく、国のために競技しているのです。 +> +> {{< /blockquote >}} ## クリケットについて -インド人はクリケットが大好きだと言っても過言ではないでしょう。この競技は、他のスポーツと異なり、インドの農村部や都市部を問わず、あらゆる場所でプレイされており、若者から年配の方まで広く人気があり、インドでは何十億人もの人々を結びつける役割を担っています。クリケットは多くのメディアの注目を集め、非常に[多額のお金](https://www.statista.com/topics/4543/indian-premier-league-ipl/)と名声がかかっています。過去数年間、テクノロジーは文字通りクリケットの試合を変えてきました。視聴者はストリーミングメディア、トーナメント、モバイルベースの手頃なアクセスによるライブクリケット視聴などを享受しています。 +インド人はクリケットが大好きだと言っても過言ではないでしょう。 この競技は、他のスポーツと異なり、インドの農村部や都市部を問わず、あらゆる場所でプレイされており、若者から年配の方まで広く人気があり、インドでは何十億人もの人々を結びつける役割を担っています。 クリケットは多くのメディアの注目を集めています。 クリケットは多くのメディアの注目を集め、非常に[多額のお金](https://www.statista.com/topics/4543/indian-premier-league-ipl/)と名声がかかっています。 過去数年間、テクノロジーは文字通りクリケットの試合を変えてきました。 視聴者はストリーミングメディア、トーナメント、モバイルベースの手頃なアクセスによるライブクリケット視聴などを享受しています。 インドプレミアリーグ (IPL) は、2008年に設立された20チームから成るプロクリケットリーグです。 これは世界で最も参加者が多いクリケットイベントの1つで、2019年の市場規模は[67億ドル](https://en.wikipedia.org/wiki/Indian_Premier_League)だと評価されています。 -クリケットは数のゲームです。バッツマンによってスコアされたランの数、ボウラーによって取られたウィケットの数、クリケットチームによって獲得した試合の数、バッツマンがボウリング攻撃に特定の方法で応答する回数。パフォーマンスを向上させたり、クリケットのビジネスチャンス・市場・経済性などを研究するため、NumPyなどの数値計算ソフトウェアを利用した強力な分析ツールによりクリケットの数字を掘り下げる能力は、大きな意味を持ちます。クリケット分析は、試合に関する興味深い洞察と、ゲームの結果に関する予測AIを提供します。 +クリケットは数のゲームです。 バッツマンによってスコアされたランの数、ボウラーによって取られたウィケットの数、クリケットチームによって獲得した試合の数、バッツマンがボウリング攻撃に特定の方法で応答する回数。 クリケットの数字を掘り下げてパフォーマンスを向上させるとともに、NumPyなどの数値計算ソフトウェアを利用した強力な分析ツールを介して、クリケットのビジネスチャンス、市場全体、経済性を研究することは、大きな意味を持ちます。 クリケット分析は、試合に関する興味深い洞察と、ゲームの結果に関する予測AIを提供します。 -現在では、クリケットゲームの記録と 利用可能な統計データは豊富で、ほぼ無限の宝の山だと言えます。: [ESPN cricinfo や](https://stats.espncricinfo.com/ci/engine/stats/index.html) [cricsheet](https://cricsheet.org). これらのクリケットデータベースは、最新の機械学習と予測モデリングアルゴリズムを使用して、 [クリケット 分析](https://www.researchgate.net/publication/336886516_Data_visualization_and_toss_related_analysis_of_IPL_teams_and_batsmen_performances) に使用されています。 メディアやプロスポーツ団体のエンターテインメントプラットフォームは、技術や分析を利用し、試合勝率を向上させるために、下記のような要素が主要なメトリックだと考え始めています。 +現在では、クリケットゲームの記録と 利用可能な統計データは豊富で、ほぼ無限の宝の山だと言えます。 : [ESPN cricinfo や](https://stats.espncricinfo.com/ci/engine/stats/index.html) [cricsheet](https://cricsheet.org). これらのクリケットデータベースは、最新の機械学習と予測モデリングアルゴリズムを使用して、 [クリケット 分析](https://www.researchgate.net/publication/336886516_Data_visualization_and_toss_related_analysis_of_IPL_teams_and_batsmen_performances) に使用されています。 メディアやプロスポーツ団体のエンターテインメントプラットフォームは、技術や分析を利用し、試合勝率を向上させるために、下記のような要素が主要なメトリックだと考え始めています。 * バッティング成績の移動平均 * スコア予測 * プレイヤーの体力や、異なる相手に対するパフォーマンスについての洞察 * チーム構成に戦略的な決定を下すための、各勝敗へのプレイヤーの貢献 -{{< figure src="/images/content_images/cs/cricket-pitch.png" class="csfigcaption" caption="** フィールドのフォーカルポイントとなるクリケットピッチ**" alt="A cricket pitch with bowler and batsmen" align="middle" attr="*(Image credit: Debarghya Das)*" attrlink="http://debarghyadas.com/files/IPLpaper.pdf" >}} +{{< figure >}} +src = '/images/content_images/cs/cricket-pitch.png' title = 'Cricket Pitch, the focal point in the field' alt = 'A cricket pitch with bowler and batsmen' align = 'center' attribution = '(Image credit: Debarghya Das)' attributionlink = 'http://debarghyadas.com/files/IPLpaper.pdf' +{{< figure src="/images/content_images/cs/numpy_ca_benefits.png" class="fig-center" alt="クリケット分析にNumPyを使用するメリットを示す図" caption="** 利用されている主なNumPy機能 **" >}} ### データ分析の主要な目標 @@ -33,32 +37,36 @@ sidebar: false * リアルタイムデータ分析は、ゲーム中の洞察を得ることができ、チームや関連ビジネスが経済的利益と成長のために戦術を変更するためも役立ちます。 * 履歴分析に加えて、予測モデルは可能性のある結果を求めることができますが、かなりの数のナンバークランチングとデータサイエンスのノウハウ、可視化ツール、および分析に新しい観測データを含める機能などが必要になります。 -{{< figure src="/images/content_images/cs/player-pose-estimator.png" class="fig-center" alt="pose estimator" caption="**クリケットの姿勢推定**" attr="*(Image credit: connect.vin)*" attrlink="https://connect.vin/2019/05/ai-for-cricket-batsman-pose-analysis/" >}} +{{< figure >}} +src = '/images/content_images/cs/player-pose-estimator.png' alt = 'pose estimator' title = 'Cricket Pose Estimator' attribution = '(Image credit: connect.vin)' attributionlink = 'https://connect.vin/2019/05/ai-for-cricket-batsman-pose-analysis/' +{{< /figure >}} ### 課題 * **データのクリーニングと前処理** - IPLは、クリケットを古典的なテストマッチ形式から、はるかに大規模に拡大させました。 毎シーズン、様々なフォーマットで行われる試合の数は増加しており、データ、アルゴリズム、最新のスポーツデータ分析技術、シミュレーションモデルも増加しています。クリケットのデータ分析には、フィールドマッピング、プレイヤートラッキング、ボールトラッキング、プレイヤーショット分析、およびボールがどのように動くのか、その角度、スピン、速度、軌道など、他の沢山の種類のデータを必要とします。 これらの要因により、データクリーニングと前処理の複雑さが増してしまいました。 + IPLは、クリケットを古典的なテストマッチ形式から、はるかに大規模に拡大させました。 毎シーズン、様々なフォーマットで行われる試合の数は増加しており、データ、アルゴリズム、最新のスポーツデータ分析技術、シミュレーションモデルも増加しています。 クリケットのデータ分析には、フィールドマッピング、プレイヤートラッキング、ボールトラッキング、プレイヤーショット分析、およびボールがどのように動くのか、その角度、スピン、速度、軌道など、他の沢山の種類のデータを必要とします。 これらの要因により、データクリーニングと前処理の複雑さが増してしまいました。 * **動的モデリング** - クリケットでは、他のスポーツと同様、フィールド上の選手の様々な数字を追跡するために、関連する変数の数が多くなってしまいがちです。たとえば、ボールやその属性情報、およびいくつかの行動をとるアクションのいくつかの可能性などの変数です。データ分析とモデリングの複雑さは、分析中に必要となる予測のための質問の種類に正比例しており、データ表現とモデルにも大きく依存しています。バッツマンが異なる角度や速度でボールを打った場合に何が起こるのかのような、動的なクリケットのプレーの予測が必要な場合、計算量やデータ比較が更に困難になります。 + クリケットでは、他のスポーツと同様、フィールド上の選手の様々な数字を追跡するために、関連する変数の数が多くなってしまいがちです。 たとえば、ボールやその属性情報、およびいくつかの行動をとるアクションのいくつかの可能性などの変数です。 データ分析とモデリングの複雑さは、分析中に必要となる予測のための質問の種類に正比例しており、データ表現とモデルにも大きく依存しています。 バッツマンが異なる角度や速度でボールを打った場合に何が起こるのかのような、動的なクリケットのプレーの予測が必要な場合、計算量やデータ比較が更に困難になります。 * **予測分析の複雑さ** - クリケットにおいて、意思決定の多くは「ボウラーがある特定のタイプの場合、打者はどのくらいの頻度で特定の種類のショットを打つのか」「バッツマンが特定の方法であるボウラーに反応した場合、ボウラーはどのようにラインと長さを変更するのか 」などの質問に基づいています。この種の予測分析クエリでは、精度の良いデータセットが利用できることと、データを合成して高精度な生成モデルを作成できることが必要とされます。 + クリケットにおいて、意思決定の多くは「ボウラーがある特定のタイプの場合、打者はどのくらいの頻度で特定の種類のショットを打つのか」「バッツマンが特定の方法であるボウラーに反応した場合、ボウラーはどのようにラインと長さを変更するのか 」などの質問に基づいています。 この種の予測分析クエリでは、精度の良いデータセットが利用できることと、データを合成して高精度な生成モデルを作成できることが必要とされます。 ## クリケット解析におけるNumPyの役割 -スポーツ分析は現在、非常に盛んな分野です。 多くの研究者や企業は、最新の機械学習やAI技術以外にも、NumPyや、Scikit-learn, SciPy, Matplotlib, Jupyterなどの他のPyDataパッケージを[使っています](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx)。NumPyは以下のように、クリケット関連の様々なスポーツ分析に使用されています。 +スポーツ分析は現在、非常に盛んな分野です。 多くの研究者や企業は、最新の機械学習やAI技術以外にも、NumPyや、Scikit-learn, SciPy, Matplotlib, Jupyterなどの他のPyDataパッケージを[使っています](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx)。 NumPyは以下のように、クリケット関連の様々なスポーツ分析に使用されています。 -* **統計分析:** NumPyの数値計算機能は、様々なプレイヤーやゲーム戦術のコンテキストでの観測データで、試合中のイベントの統計的有意性を推定し、生成モデルや静的モデルと比較して試合結果を推定するのに役立ちます。[因果分析](https://amplitude.com/blog/2017/01/19/causation-correlation) と [ビッグデータアプローチ](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4996805/)が戦術的分析に使用されています。 +* **統計分析:** NumPyの数値計算機能は、様々なプレイヤーやゲーム戦術のコンテキストでの観測データで、試合中のイベントの統計的有意性を推定し、生成モデルや静的モデルと比較して試合結果を推定するのに役立ちます。 [因果分析](https://amplitude.com/blog/2017/01/19/causation-correlation) と [ビッグデータアプローチ](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4996805/)が戦術的分析に使用されています。 * **データ可視化:** データのグラフ化・[可視化](https://towardsdatascience.com/advanced-sports-visualization-with-pandas-matplotlib-and-seaborn-9c16df80a81b) は、さまざまなデータセット間の関係について、有益な洞察を与えてくれます。 ## まとめ -スポーツアナリティクスは、プロの試合についてはまさにゲームチェンジャーです。特に戦略的な意思決定については、最近まで主に「直感」や過去の伝統的な考え方に基づいて行われていたため、大きな影響があります。NumPyは、データ分析・機械学習・人工知能のアルゴリズムに関連する高レベル関数を提供する沢山のPythonパッケージ群の、堅固な基盤となっています。これらのパッケージは、ゲームの結果を変えるような意思決定を支援するリアルタイムのインサイトを得るため、クリケットの試合だけでなく関連する推論やビジネスの推進にも広く使用されています。クリケットの試合結果につながる隠れたパラメータや、パターン、属性を見つけることは、ステークホルダーが数字や統計に隠されているゲームの洞察方法を見つけるのにも役に立つのです。 +スポーツアナリティクスは、プロの試合についてはまさにゲームチェンジャーです。 特に戦略的な意思決定については、最近まで主に「直感」や過去の伝統的な考え方に基づいて行われていたため、大きな影響があります。 NumPyは、データ分析・機械学習・人工知能のアルゴリズムに関連する高レベル関数を提供する沢山のPythonパッケージ群の、堅固な基盤となっています。 これらのパッケージは、ゲームの結果を変えるような意思決定を支援するリアルタイムのインサイトを得るため、クリケットの試合だけでなく関連する推論やビジネスの推進にも広く使用されています。 クリケットの試合結果につながる隠れたパラメータや、パターン、属性を見つけることは、ステークホルダーが数字や統計に隠されているゲームの洞察方法を見つけるのにも役に立つのです。 -{{< figure src="/images/content_images/cs/numpy_ca_benefits.png" class="fig-center" alt="クリケット分析にNumPyを使用するメリットを示す図" caption="** 利用されている主なNumPy機能 **" >}} +{{< figure >}} +src = '/images/content_images/cs/numpy_ca_benefits.png' alt = 'Diagram showing benefits of using NumPy for cricket analytics' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} diff --git a/content/ja/case-studies/deeplabcut-dnn.md b/content/ja/case-studies/deeplabcut-dnn.md index c597261ef4..d5db28843a 100644 --- a/content/ja/case-studies/deeplabcut-dnn.md +++ b/content/ja/case-studies/deeplabcut-dnn.md @@ -4,37 +4,41 @@ sidebar: false --- {{< figure src="/images/content_images/cs/mice-hand.gif" class="fig-center" caption="**DeepLapCutを用いたマウスの手の動きの解析 **" alt="micehandanim" attr="*(Source: www.deeplabcut.org )*" attrlink="http://www.mousemotorlab.org/deeplabcut">}} +src = '/images/content_images/cs/mice-hand.gif' title = 'Analyzing mice hand-movement using DeepLapCut' alt = 'micehandanim' attribution = '(Source: www.deeplabcut.org )' attributionlink = 'http://www.mousemotorlab.org/deeplabcut' +{{< /figure >}} -
-

オープンソースソフトウェアは生体臨床医学を加速させています。 DeepLabCut を使用すると、深層学習を使用して動物の行動を自動的にビデオ解析することができます。

-
—Alexander Mathis、 准教授、École polytechnology fe’rale de Lausanne (EPFL)
-
+{{< blockquote cite="https://news.harvard.edu/gazette/story/newsplus/harvard-researchers-awarded-czi-open-source-award/" by="Alexander Mathis, * ローザンヌ連邦工科大学([EPFL](https://www.epfl.ch/en/))の助教授 *" +> }} オープンソースソフトウェアは生体臨床医学を加速させています。 DeepLabCut を使用すると、深層学習を使用して動物の行動を自動的にビデオ解析することができます。 +> +> {{< /blockquote >}} ## DeepLabCut について -[DeepLabCut](https://github.com/DeepLabCut/DeepLabCut)は、ごくわずかなトレーニングデータで人間レベルの精度で実験動物の行動を追跡可能にするオープンソースのツールボックスです。世界中の何百もの研究機関の研究者により使用されています。DeepLabCutの技術を使うことで、科学者は動物の種類と時系列のデータをもとに、運動制御と行動に関する科学的な理解を深めることができるようになりました。 +[DeepLabCut](https://github.com/DeepLabCut/DeepLabCut)は、ごくわずかなトレーニングデータで人間レベルの精度で実験動物の行動を追跡可能にするオープンソースのツールボックスです。 DeepLabCutの技術を使うことで、科学者は動物の種類と時系列のデータをもとに、運動制御と行動に関する科学的な理解を深めることができるようになりました。 -神経科学、医学、生体力学などのいくつかの研究分野では、動物の動きを追跡したデータを使用しています。DeepLabCutは、動画に記録された動きを解析することで、人間やその他の動物が何をしているのかを理解することができます。タグ付けや監視などの、手間のかかる作業を自動化し、深層学習ベースのデータ解析を実施します。DeepLabCutは、霊長類、マウス、魚、ハエなどの動物を観察する科学研究をより速く正確にしています。 +神経科学、医学、生体力学などのいくつかの研究分野では、動物の動きを追跡したデータを使用しています。 DeepLabCutは、動画に記録された動きを解析することで、人間やその他の動物が何をしているのかを理解することができます。 タグ付けや監視などの、手間のかかる作業を自動化し、深層学習ベースのデータ解析を実施します。 DeepLabCutは、霊長類、マウス、魚、ハエなどの動物を観察する科学研究をより速く正確にしています。 -{{< figure src="/images/content_images/cs/race-horse.gif" class="fig-center" caption="**色のついた点は競走馬の体の位置を追跡**" alt="horserideranim" attr="*(Source: Mackenzie Mathis)*">}} +{{< figure >}} +src = '/images/content_images/cs/race-horse.gif' title = 'Colored dots track the positions of a racehorse’s body part' alt = 'horserideranim' attribution = '(Source: Mackenzie Mathis)' +{{< /figure >}} -DeepLabCutは、動物の姿勢を抽出することで非侵襲的な行動追跡を行います。これは、生体力学、遺伝学、倫理学、神経科学などの分野での研究に必要不可欠です。動的に変化する背景の中で、動物の姿勢をビデオデータから非侵襲的に測定することは、技術的にも、必要な計算リソースやトレーニングデータの点でも、非常に困難な計算処理です。 +DeepLabCutは、動物の姿勢を抽出することで非侵襲的な行動追跡を行います。 これは、生体力学、遺伝学、倫理学、神経科学などの分野での研究に必要不可欠です。 動的に変化する背景の中で、動物の姿勢をビデオデータから非侵襲的に測定することは、技術的にも、必要な計算リソースやトレーニングデータの点でも、非常に困難な計算処理です。 -DeepLabCutは、研究者が対象の姿勢を推定し、Pythonベースのソフトウェアを使って効率的に対象の行動を定量化することを可能にします。DeepLabCutを使用すると、研究者は動画から異なるフレームを識別し、数十個のフレームの特定の身体部位を、よくできたGUIによってラベルづけできます。すると、DeepLabCutの深層学習ベースのポーズ推定アーキテクチャにより、動画の残りの部分や動物の他の類似した動画から同じ特徴を抽出する方法を学習できます。ハエやマウスなどの一般的な実験動物から [チーター][cheetah-movement]のようなより珍しい動物まで、動物の種類を問わず利用できます。 +DeepLabCutは、研究者が対象の姿勢を推定し、Pythonベースのソフトウェアを使って効率的に対象の行動を定量化することを可能にします。 DeepLabCutを使用すると、研究者は動画から異なるフレームを識別し、数十個のフレームの特定の身体部位を、よくできたGUIによってラベルづけできます。 すると、DeepLabCutの深層学習ベースのポーズ推定アーキテクチャにより、動画の残りの部分や動物の他の類似した動画から同じ特徴を抽出する方法を学習できます。 ハエやマウスなどの一般的な実験動物から [チーター][cheetah-movement]のようなより珍しい動物まで、動物の種類を問わず利用できます。 -DeepLabCutでは[転移学習](https://arxiv.org/pdf/1909.11229)という技術を使用しています。これにより必要な学習データの量を大幅に削減し、学習の収束を加速させることができます。必要に応じて、より高速な推論を提供するさまざまなネットワークアーキテクチャ(MobileNetV2など)を選択することができ、リアルタイムの実験データフィードバックと組み合わせることもできます。DeepLabCutはもともと[DeeperCut](https://arxiv.org/abs/1605.03170)と呼ばれるパフォーマンスのよい人用のポーズ推定アーキテクチャの特徴検出器を使用しており、これが名前の由来になりました。今ではこのパッケージは大幅に変更され、追加のアーキテクチャ・データの水増し・一通りのユーザー用フロントエンドを含んでいます。さらに、 大規模な生物学的実験をサポートするため、DeepLabCutはオンライン学習の機能を提供しています。これにより、動画の時間をこえて学習データを増やすことができ、エッジケースをカバーしたり、特定のコンテキスト内でポーズ推定アルゴリズムを堅牢にしたりできます。 +DeepLabCutでは[転移学習](https://arxiv.org/pdf/1909.11229)という技術を使用しています。 これにより必要な学習データの量を大幅に削減し、学習の収束を加速させることができます。 必要に応じて、より高速な推論を提供するさまざまなネットワークアーキテクチャ(MobileNetV2など)を選択することができ、リアルタイムの実験データフィードバックと組み合わせることもできます。 DeepLabCutはもともと[DeeperCut](https://arxiv.org/abs/1605.03170)と呼ばれるパフォーマンスのよい人用のポーズ推定アーキテクチャの特徴検出器を使用しており、これが名前の由来になりました。 今ではこのパッケージは大幅に変更され、追加のアーキテクチャ・データの水増し・一通りのユーザー用フロントエンドを含んでいます。 さらに、 大規模な生物学的実験をサポートするため、DeepLabCutはオンライン学習の機能を提供しています。 これにより、動画の時間をこえて学習データを増やすことができ、エッジケースをカバーしたり、特定のコンテキスト内でポーズ推定アルゴリズムを堅牢にしたりできます。 -最近、[DeepLabCut model zoo](http://www.mousemotorlab.org/dlc-modelzoo)が発表されました。これは、霊長類の顔分析から犬の姿勢まで、様々な種や実験条件に対応した事前訓練済みモデルを提供しています。これにより、例えば、新しいデータのラベルを付けることなくクラウドで予測を実行することができたり、ニューラルネットワークの学習を実行することができます。プログラミング経験は必要ありません。 +最近、[DeepLabCut model zoo](http://www.mousemotorlab.org/dlc-modelzoo)が発表されました。 これは、霊長類の顔分析から犬の姿勢まで、様々な種や実験条件に対応した事前訓練済みモデルを提供しています。 これにより、例えば、新しいデータのラベルを付けることなくクラウドで予測を実行することができたり、ニューラルネットワークの学習を実行することができます。 プログラミング経験は必要ありません。 ### 主な目標と結果 * **科学研究のための動物姿勢解析の自動化:** - DeepLabCutという技術の主な目的は、多様な環境で動物の姿勢を測定し追跡することです。このデータは例えば神経科学の研究において、脳がどのように運動を制御しているかを理解するためのや、動物がどのように社会的に交流しているかを明らかにするために利用することができます。研究者はDeepLabCutで [10倍のパフォーマンス向上](https://www.biorxiv.org/content/10.1101/457242v1) が可能であると発表しています。オフラインでは最大1200フレーム/秒(FPS) で姿勢を推定することができます。 + DeepLabCutという技術の主な目的は、多様な環境で動物の姿勢を測定し追跡することです。 このデータは例えば神経科学の研究において、脳がどのように運動を制御しているかを理解するためのや、動物がどのように社会的に交流しているかを明らかにするために利用することができます。 研究者はDeepLabCutで [10倍のパフォーマンス向上](https://www.biorxiv.org/content/10.1101/457242v1) が可能であると発表しています。 オフラインでは最大1200フレーム/秒(FPS) で姿勢を推定することができます。 * **姿勢推定のための使いやすいPythonツールキットの作成:** - DeepLabCutは、動物の姿勢推定技術を研究者が簡単に利用できるツールとして共有したいという考えから開発されています。そこで開発者らはプロジェクト管理機能を備えた、単独で機能し、使いやすいPythonツールボックスとしてこのツールを作成しました。 これにより、姿勢推定を自動化するだけでなく、DeepLabCutツールキットユーザーをデータセット収集段階から共有可能・再利用可能な分析パイプラインを作成する段階まで補助し、プロジェクトをエンドツーエンドで管理することも可能になりました。 + DeepLabCutは、動物の姿勢推定技術を研究者が簡単に利用できるツールとして共有したいという考えから開発されています。 そこで開発者らはプロジェクト管理機能を備えた、単独で機能し、使いやすいPythonツールボックスとしてこのツールを作成しました。 これにより、姿勢推定を自動化するだけでなく、DeepLabCutツールキットユーザーをデータセット収集段階から共有可能・再利用可能な分析パイプラインを作成する段階まで補助し、プロジェクトをエンドツーエンドで管理することも可能になりました。 この[ツールキット][DLCToolkit] はオープンソースとして利用できます。 @@ -45,27 +49,31 @@ DeepLabCutでは[転移学習](https://arxiv.org/pdf/1909.11229)という技術 - 動画における大規模推論のためのコード作成 - 統合された可視化ツールを使用した推論の描画 -{{< figure src="/images/content_images/cs/deeplabcut-toolkit-steps.png" class="csfigcaption" caption="**DeepLabCutによる姿勢推定のステップ**" alt="dlcsteps" align="middle" attr="(Source: DeepLabCut)" attrlink="https://twitter.com/DeepLabCut/status/1198046918284210176/photo/1" >}} +{{< figure >}} +{{< figure src="/images/content_images/cs/deeplabcut-workflow.png" class="fig-center" caption="**DeepLabCutのワークフロー**" alt="workflow" attr="*(Source: Mackenzie Mathis)*" attrlink="https://www.researchgate.net/figure/DeepLabCut-work-flow-The-diagram-delineates-the-work-flow-as-well-as-the-directory-and_fig1_329185962">}} +{{< /figure >}} ### 課題 * **速度** - 動物行動動画の高速な処理は、動物の行動を測定し、科学実験をより効率的で正確にするために重要です。動的に変化する背景の中で、マーカーを使用せずに、実験室での実験のために動物の詳細な姿勢を抽出することは、技術的にも、必要なリソース的にも、必要なトレーニングデータの面でも、困難な場合があります。科学者が、より現実的な状況で研究を行うために、コンピュータビジョンなどの専門知識のスキルを必要とせずに使うことができるツールを開発することは、解決すべき重要な問題です。 + 動物行動動画の高速な処理は、動物の行動を測定し、科学実験をより効率的で正確にするために重要です。 動的に変化する背景の中で、マーカーを使用せずに、実験室での実験のために動物の詳細な姿勢を抽出することは、技術的にも、必要なリソース的にも、必要なトレーニングデータの面でも、困難な場合があります。 科学者が、より現実的な状況で研究を行うために、コンピュータビジョンなどの専門知識のスキルを必要とせずに使うことができるツールを開発することは、解決すべき重要な問題です。 * **組み合わせ問題** - 組合せ問題とは、複数の四肢の動きを個々の動物行動に統合することを指します。 キーポイントと、その個々の動物行動との関連性を組み合わせ、時間的に結びつけることは、複雑なプロセスであり、非常に膨大な数値解析が必要となります。特に、実験映像の中で複数の動物の動きを追跡する場合は大変です。 + 組合せ問題とは、複数の四肢の動きを個々の動物行動に統合することを指します。 キーポイントと、その個々の動物行動との関連性を組み合わせ、時間的に結びつけることは、複雑なプロセスであり、非常に膨大な数値解析が必要となります。 特に、実験映像の中で複数の動物の動きを追跡する場合は大変です。 * **データ処理** - 最後に、配列の操作もかなり難しい問題です。様々な画像や、目標のテンソル、キーポイントに対応する大きな配列のスタックを処理しなければならないからです。 + 最後に、配列の操作もかなり難しい問題です。 様々な画像や、目標のテンソル、キーポイントに対応する大きな配列のスタックを処理しなければならないからです。 -{{< figure src="/images/content_images/cs/pose-estimation.png" class="csfigcaption" caption="**姿勢推定の多様性と難しさ**" alt="challengesfig" align="middle" attr="(Source: Mackenzie Mathis)" attrlink="https://www.biorxiv.org/content/10.1101/476531v1.full.pdf" >}} +{{< figure src="/images/content_images/cs/numpy_dlc_benefits.png" class="fig-center" alt="numpy benefits" caption="**NumPyの主要機能**" >}} +src = '/images/content_images/cs/pose-estimation.png' title = 'Pose estimation variety and complexity' alt = 'challengesfig' align = 'center' attribution = '(Source: Mackenzie Mathis)' attributionlink = 'https://www.biorxiv.org/content/10.1101/476531v1.full.pdf' +{{< /figure >}} ## 姿勢推定の課題に対応するためのNumPyの役割 -NumPy は DeepLabCutにおける、行動分析の高速化のための数値計算の核となっています。NumPyだけでなく、DeepLabCutは様々なNumPyをベースとしているPythonライブラリを利用しています。[SciPy](https://www.scipy.org)、[Pandas](https://pandas.pydata.org)、[matplotlib](https://matplotlib.org)、[Tensorpack](https://github.com/tensorpack/tensorpack), [imgaug](https://github.com/aleju/imgaug)、[scikit-learn](https://scikit-learn.org/stable/)、[scikit-image](https://scikit-image.org)、[Tensorflow](https://www.tensorflow.org)などです。 +NumPy は DeepLabCutにおける、行動分析の高速化のための数値計算の核となっています。 NumPyだけでなく、DeepLabCutは様々なNumPyをベースとしているPythonライブラリを利用しています。 [SciPy](https://www.scipy.org)、[Pandas](https://pandas.pydata.org)、[matplotlib](https://matplotlib.org)、[Tensorpack](https://github.com/tensorpack/tensorpack), [imgaug](https://github.com/aleju/imgaug)、[scikit-learn](https://scikit-learn.org/stable/)、[scikit-image](https://scikit-image.org)、[Tensorflow](https://www.tensorflow.org)などです。 以下に挙げるNumPyの特徴が、DeepLabCutの姿勢推定アルゴリズムでの画像処理・組み合わせ処理・高速計算において、重要な役割を果たしました。 @@ -75,15 +83,19 @@ NumPy は DeepLabCutにおける、行動分析の高速化のための数値計 * ランダムサンプリング * 大きな配列の再構成 -DeepLabCutは、ツールキットが提供するワークフローを通じてNumPyの配列機能を利用しています。 特に、NumPyはヒューマンアノテーションのラベル付けや、アノテーションの書き込み、編集、処理のために、特定のフレームをサンプリングするために使用されています。TensorFlowを使ったニューラルネットワークは、DeepLabCutの技術によって何千回も訓練され、 フレームから真のアノテーション情報を予測します。この目的のため、姿勢推定問題を画像-画像変換問題として変換する目標密度(スコアマップ) を作成します。ニューラルネットワークのロバスト化のため、データの水増しを使用していますが、このためには幾何学・画像的処理を施したスコアマップの計算を行うことが必要になります。また学習を高速化するため、NumPyのベクトル化機能が利用されています。 推論には、目標のスコアマップから最も可能性の高い予測値を抽出し、効率的に「予測値をリンクさせて個々の動物を組み立てる」ことが必要になります。 +DeepLabCutは、ツールキットが提供するワークフローを通じてNumPyの配列機能を利用しています。 特に、NumPyはヒューマンアノテーションのラベル付けや、アノテーションの書き込み、編集、処理のために、特定のフレームをサンプリングするために使用されています。 TensorFlowを使ったニューラルネットワークは、DeepLabCutの技術によって何千回も訓練され、 フレームから真のアノテーション情報を予測します。 この目的のため、姿勢推定問題を画像-画像変換問題として変換する目標密度(スコアマップ) を作成します。 ニューラルネットワークのロバスト化のため、データの水増しを使用していますが、このためには幾何学・画像的処理を施したスコアマップの計算を行うことが必要になります。 また学習を高速化するため、NumPyのベクトル化機能が利用されています。 推論には、目標のスコアマップから最も可能性の高い予測値を抽出し、効率的に「予測値をリンクさせて個々の動物を組み立てる」ことが必要になります。 -{{< figure src="/images/content_images/cs/deeplabcut-workflow.png" class="fig-center" caption="**DeepLabCutのワークフロー**" alt="workflow" attr="*(Source: Mackenzie Mathis)*" attrlink="https://www.researchgate.net/figure/DeepLabCut-work-flow-The-diagram-delineates-the-work-flow-as-well-as-the-directory-and_fig1_329185962">}} +{{< figure >}} +src = '/images/content_images/cs/deeplabcut-workflow.png' title = 'DeepLabCut Workflow' alt = 'workflow' attribution = '(Source: Mackenzie Mathis)' attributionlink = 'https://www.researchgate.net/figure/DeepLabCut-work-flow-The-diagram-delineates-the-work-flow-as-well-as-the-directory-and_fig1_329185962' +{{< /figure >}} ## まとめ -行動を観察し、効率的に表現することは、現代倫理学、神経科学、医学、工学の根幹です。[DeepLabCut](http://orga.cvss.cc/wp-content/uploads/2019/05/NathMathis2019.pdf) により、研究者は対象の姿勢を推定し、行動を効率的に定量化できるようになりました。DeepLabCutというPythonツールボックスを使えば、わずかな学習画像のセットでニューラルネットワークを人間レベルのラベリング精度で学習することができ、実験室での行動分析だけでなく、スポーツ、歩行分析、医学、リハビリテーション研究などへの応用が可能になります。DeepLabCutアルゴリズムに必要な複雑な組み合わせ処理やデータ処理の問題を、NumPyの配列操作機能が解決しています。 +行動を観察し、効率的に表現することは、現代倫理学、神経科学、医学、工学の根幹です。 [DeepLabCut](http://orga.cvss.cc/wp-content/uploads/2019/05/NathMathis2019.pdf) により、研究者は対象の姿勢を推定し、行動を効率的に定量化できるようになりました。 DeepLabCutというPythonツールボックスを使えば、わずかな学習画像のセットでニューラルネットワークを人間レベルのラベリング精度で学習することができ、実験室での行動分析だけでなく、スポーツ、歩行分析、医学、リハビリテーション研究などへの応用が可能になります。 DeepLabCutアルゴリズムに必要な複雑な組み合わせ処理やデータ処理の問題を、NumPyの配列操作機能が解決しています。 -{{< figure src="/images/content_images/cs/numpy_dlc_benefits.png" class="fig-center" alt="numpy benefits" caption="**NumPyの主要機能**" >}} +{{< figure >}} +src = '/images/content_images/cs/numpy_dlc_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} [cheetah-movement]: https://www.technologynetworks.com/neuroscience/articles/interview-a-deeper-cut-into-behavior-with-mackenzie-mathis-327618 diff --git a/content/ja/case-studies/gw-discov.md b/content/ja/case-studies/gw-discov.md index 49d88c2845..e7faa49de9 100644 --- a/content/ja/case-studies/gw-discov.md +++ b/content/ja/case-studies/gw-discov.md @@ -4,17 +4,17 @@ sidebar: false --- {{< figure src="/images/content_images/cs/gw_sxs_image.png" class="fig-center" caption="**重力波**" alt="binary coalesce black hole generating gravitational waves" attr="*(Image Credits: The Simulating eXtreme Spacetimes (SXS) Project at LIGO)*" attrlink="https://youtu.be/Zt8Z_uzG71o" >}} +src = '/images/content_images/cs/gw_sxs_image.png' title = 'Gravitational Waves' alt = 'binary coalesce black hole generating gravitational waves' attribution = '(Image Credits: The Simulating eXtreme Spacetimes (SXS) Project at LIGO)' attributionlink = 'https://youtu.be/Zt8Z_uzG71o' +{{< /figure >}} -
-

科学計算のためのPythonエコシステムはLIGOで行われている研究のための重要なインフラです。

-
David Shoemaker, LIGOサイエンティフィック・コラボレーション
-
+{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="David Shoemaker, *LIGOの科学調査協力により*" >}} LIGOで行われる研究には、科学的なPythonエコシステムが重要なインフラとなっています。 +{{< /blockquote >}} ## [重力波](https://www.nationalgeographic.com/news/2017/10/what-are-gravitational-waves-ligo-astronomy-science/) と [LIGO](https://www.ligo.caltech.edu) について -重力波は、空間と時間の基本構造の波紋です。 2つのブラックホールの衝突や合体、2連星や超新星の合体など、大きな変動現象によって生成されます。重力波の観測は、重力を研究する上で重要なだけでなく、遠い宇宙におけるいくつかの不明瞭な現象と、その影響を理解するためにも役立ちます。 +重力波は、空間と時間の基本構造の波紋です。 2つのブラックホールの衝突や合体、2連星や超新星の合体など、大きな変動現象によって生成されます。 重力波の観測は、重力を研究する上で重要なだけでなく、遠い宇宙におけるいくつかの不明瞭な現象と、その影響を理解するためにも役立ちます。 -[レーザー干渉計重力波天文台(LIGO)](https://www. ligo. caltech. edu)は、アインシュタインの一般相対性理論によって予測された重力波の直接検出を通して、重力波天体物理学の分野を切り開くために設計されました。このシステムは、アメリカのワシントン州ハンフォードとルイジアナ州リビングストンにある2つの干渉計が一体となって構成され、重力波を検出します。それぞれのシステムには、レーザー干渉法を用いた数キロ規模の重力波検出器が設置されています。LIGO Scientific Collaboration（LSC）は、米国をはじめとする14カ国の大学から1000人以上の科学者が集まり、90以上の大学・研究機関によって支援されています。また、約250人の学生も参加しています。今回のLIGOの発見は、重力波が地球を通過する際に生じる空間と時間の微小な乱れの測定により、重力波そのものを初めて観測しました。これにより、新しい天体物理学のフロンティアが開かれました。これは、宇宙の歪んだ側面、つまり歪んだ時空から作られた物体とそれに現象を切り拓くものです。 +[レーザー干渉計重力波天文台(LIGO)](https://www.ligo.caltech.edu)は、アインシュタインの一般相対性理論によって予測された重力波の直接検出を通して、重力波天体物理学の分野を切り開くために設計されました。 重力波は非常に小さい効果を生み、物質と微小な相互作用を持つため、検出が困難です。 LIGOのすべてのデータを処理・分析するには、膨大な計算インフラが必要です。 信号の数十億倍のノイズを除去した後も、非常に複雑な相対性理論の方程式と膨大な量のデータがあり、計算上の課題となっています。 それぞれのシステムには、レーザー干渉法を用いた数キロ規模の重力波検出器が設置されています。 LIGO Scientific Collaboration（LSC）は、米国をはじめとする14カ国の大学から1000人以上の科学者が集まり、90以上の大学・研究機関によって支援されています。 また、約250人の学生も参加しています。 今回のLIGOの発見は、重力波が地球を通過する際に生じる空間と時間の微小な乱れの測定により、重力波そのものを初めて観測しました。 これにより、新しい天体物理学のフロンティアが開かれました。 これは、宇宙の歪んだ側面、つまり歪んだ時空から作られた物体とそれに伴う現象を切り拓くものです。 ### 主な目的 @@ -29,23 +29,25 @@ sidebar: false * **計算** - 重力波は非常に小さい効果を生み、物質と微小な相互作用を持つため、検出が困難です。 LIGOのすべてのデータを処理・分析するには、膨大な計算インフラが必要です。信号の数十億倍のノイズを除去した後も、非常に複雑な相対性理論の方程式と膨大な量のデータがあり、計算上の課題となっています。例えば6つのLIGO専用クラスターに分散されたバイナリ結合解析には[10の7乗オーダーのCPU時間](https:/youtu.be7mcHknWWzNI)が必要です。 + 合成により放出される重力波は、スーパーコンピュータを用いて数値相対性を手あたり次第に試すような方法では計算できません。 LIGOが収集するデータ量は、重力波の信号が少ないのと同じくらい不可解です。 * **データの氾濫** - 観測装置の感度と信頼性が高まると、様々な場所でデータの氾濫による困難が待ち受けています。それは、干し草の中から針を探すようなものです。LIGOは毎日テラバイトのデータを生成しているのです！この大量のデータを解釈するには、各検出ごとに多大な労力が必要です。例えば、LIGOによって収集される信号は、数十万個の重力波シグネチャのテンプレートで構成されており、スーパーコンピュータでしか解析できません。 + 観測装置がより高感度で信頼性を持つようになると、データの大洪水によって、干し草の中から針を探すような問題が、多重に発生することがわかります。 LIGOは毎日テラバイトのデータを生成しているのです！ この大量のデータを解釈するには、各検出ごとに多大な労力が必要です。 例えば、LIGOによって収集される信号は、数十万個の重力波シグネチャのテンプレートで構成されており、スーパーコンピュータでしか解析できません。 * **可視化** - アインシュタイン方程式を元にスーパーコンピュータでデータを解析できるようになったら、次はデータを人間の脳で理解できるようにしなければなりません。 シミュレーションのモデリングや信号の検出には、わかりやすい可視化技術が必要です。 画像処理やシミュレーションによって、解析結果をより多くの人に理解してもらえる状態になる前の段階において、可視化は、純粋な理論家に対し、数値相対性が、より信頼性の高いものとして映るようにするという役割も果たしています。理論家は、可視化とシミュレーションが結果の把握を容易にするまで、数値相対性を十分に重要視していませんでした。複雑な計算と描画を行い、また最新の実験結果と洞察に基づいてシミュレーションと再描画を行う作業は時間のかかるもので、この分野の研究者にとっての課題です。 + アインシュタイン方程式を元にスーパーコンピュータでデータを解析できるようになったら、次はデータを人間の脳で理解できるようにしなければなりません。 シミュレーションのモデリングや信号の検出には、わかりやすい可視化技術が必要です。 画像処理やシミュレーションによって、解析結果をより多くの人に理解してもらえる状態になる前の段階において、可視化は、数値相対性を十分に重要視していなかった純粋な科学愛好家の目に、数値相対性が、より信頼性の高いものとして映るようにするという役割も果たしています。 複雑な計算と描画を行い、また最新の実験結果と洞察に基づいてシミュレーションと再描画を行う作業は時間のかかるもので、この分野の研究者にとっての課題です。 -{{< figure src="/images/content_images/cs/gw_strain_amplitude.png" class="fig-center" alt="gravitational waves strain amplitude" caption="**GW150914から推定される重力波の歪みの振幅**" attr="(**Graph Credits:** Observation of Gravitational Waves from a Binary Black Hole Merger, ResearchGate Publication)" attrlink="https://www.researchgate.net/publication/293886905_Observation_of_Gravitational_Waves_from_a_Binary_Black_Hole_Merger" >}} +{{< figure >}} +{{< figure src="/images/content_images/cs/gwpy-numpy-dep-graph.png" class="fig-center" alt="gwpy-numpy depgraph" caption="**GwPyのNumPy依存グラフ**" >}} +{{< figure src="/images/content_images/cs/numpy_bh_benefits.png" class="fig-center" alt="numpy benefits" caption="**利用されたNumPyの主要機能**" >}} ## 重力波の検出におけるNumPyの役割 -合成により放出される重力波は、スーパーコンピュータを用いて数値相対性を手あたり次第に試すような方法では計算できません。LIGOが収集するデータ量は、重力波の信号が少ないのと同じくらい不可解です。 +ブラックホール合成により放出される重力波は、スーパーコンピュータを用いたブルートフォースの数値相対性処理以外の手法では計算できません。 重力波は非常に小さい効果を生み、物質と微小な相互作用を持つため、検出が困難です。 LIGOのすべてのデータを処理・分析するには、膨大な計算インフラが必要です。 信号の数十億倍のノイズを除去した後も、非常に複雑な相対性理論の方程式と膨大な量のデータがあり、計算上の課題となっています。 -Python用の標準的な数値解析パッケージNumPyは、LIGOの重力波検出プロジェクトで実行される様々なタスクに使用されるソフトウェアで利用されています。NumPyは、複雑な数学処理や高速なデータ操作に役立ちました。次にいくつかの例を示します。 +Python用の標準的な数値解析パッケージNumPyは、LIGOの重力波検出プロジェクトで実行される様々なタスクに使用されるソフトウェアで利用されています。 NumPyは、複雑な数学処理や高速なデータ操作に役立ちました。 次にいくつかの例を示します。 * [信号処理](https://www.uv.es/virgogroup/Denoising_ROF.html): グリッジ検出、[ノイズ同定とデータ判定](https://ep2016.europython.eu/media/conference/slides/pyhton-in-gravitational-waves-research-communities.pdf) (NumPy, scikit-learn, scipy, matplotlib, pandas, pyCharm)。 * データ取得: どのデータが解析できるかを決定し、干し草の中の針のような信号が入っているかどうかを突き止める。 @@ -56,14 +58,20 @@ Python用の標準的な数値解析パッケージNumPyは、LIGOの重力波 * 相関計算 * 重力波データ解析のために開発された[ソフトウェア群](https://github.com/lscsoft): [GwPy](https://gwpy.github.io/docs/stable/overview.html)や [PyCBC](https://pycbc.org)は、NumPyやAstroPyを用いて、重力波検出器データを研究するためのユーティリティー・ツール・関数へのオブジェクト指向インターフェースを提供しています。 -{{< figure src="/images/content_images/cs/gwpy-numpy-dep-graph.png" class="fig-center" alt="gwpy-numpy depgraph" caption="**GwPyのNumPy依存グラフ**" >}} +{{< figure >}} +src = '/images/content_images/cs/gwpy-numpy-dep-graph.png' alt = 'gwpy-numpy depgraph' title = 'Dependency graph showing how GwPy package depends on NumPy' +{{< /figure >}} ---- -{{< figure src="/images/content_images/cs/PyCBC-numpy-dep-graph.png" class="fig-center" alt="PyCBC-numpy depgraph" caption="**PyCBCのNumPy依存グラフ**" >}} +{{< figure >}} +src = '/images/content_images/cs/PyCBC-numpy-dep-graph.png' alt = 'PyCBC-numpy depgraph' title = 'Dependency graph showing how PyCBC package depends on NumPy' +{{< /figure >}} ## まとめ -重力波の検出により、研究者はこれまでに予期しなかった現象を発見することができました。 一方で、これまで知られてきた深遠な天体物理学の現象に、多くに新たな洞察を提供しました。数値処理とデータの可視化は、科学者が科学的な観測から収集したデータについての洞察を得て、その結果を理解するのに役立つ重要なステップです。 しかし、その計算は複雑であり、実際の観測データと分析を用いたコンピュータシミュレーションを用いて可視化されない限り、人間が理解することはできませんでした。 NumPyは、matplotlib・pandas・scikit-learnなどのPythonパッケージとともに、研究者が複雑な質問に答え、私たちの宇宙に対するの理解において、新しい地平を発見することを[可能にしています](https://www.gw-openscience.org/events/GW150914/)。 +GW検出により、研究者は完全に予期せぬ現象 を発見することができ、同時に知られている最も深遠な天体物理学 現象の多くに新たな洞察を提供しています。 数値処理とデータの可視化は、科学者が科学的な観測から収集したデータについての洞察を得て、その結果を理解するのに役立つ重要なステップです。 しかし、その計算は複雑であり、実際の観測データと分析を用いたコンピュータシミュレーションを用いて可視化されない限り、人間が理解することはできませんでした。 NumPyは、matplotlib・pandas・scikit-learnなどのPythonパッケージとともに、研究者が複雑な現象の疑問に答え、私たちの宇宙に対するの理解において、新しい地平を発見することを[可能にしています](https://www.gw-openscience.org/events/GW150914/)。 -{{< figure src="/images/content_images/cs/numpy_bh_benefits.png" class="fig-center" alt="numpy benefits" caption="**利用されたNumPyの主要機能**" >}} +{{< figure >}} +src = '/images/content_images/cs/numpy_gw_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} diff --git a/content/ja/citing-numpy.md b/content/ja/citing-numpy.md index 9696c6e4d1..397ca192ab 100644 --- a/content/ja/citing-numpy.md +++ b/content/ja/citing-numpy.md @@ -1,5 +1,5 @@ --- -title: NumPy を引用する場合 +title: NumPyを引用する sidebar: false --- @@ -12,8 +12,7 @@ _BibTeX形式:_ ``` @Article{ harris2020array, title = {Array programming with {NumPy}}, - author = {Charles R. Harris and K. Jarrod Millman and St{'{e}}fan J. - van der Walt and Ralf Gommers and Pauli Virtanen and David + author = {Charles R. Harris and K. Jarrod Millman and St{'{e}}fan J. van der Walt and Ralf Gommers and Pauli Virtanen and David Cournapeau and Eric Wieser and Julian Taylor and Sebastian Berg and Nathaniel J. Smith and Robert Kern and Matti Picus and Stephan Hoyer and Marten H. van Kerkwijk and Matthew diff --git a/content/ja/code-of-conduct.md b/content/ja/code-of-conduct.md index 9ae59b74b2..044123f3d1 100644 --- a/content/ja/code-of-conduct.md +++ b/content/ja/code-of-conduct.md @@ -7,9 +7,9 @@ aliases: ### はじめに -この行動規範は、NumPy プロジェクトによって管理されるすべての場所で適用されます。この場所とは、すべてのパブリックおよびプライベートのメーリングリスト、イシュートラッカー、Wiki、ブログ、Twitter、コミュニティで使用されているその他の通信チャンネルなどを含みます。 NumPy プロジェクトでは対面でのイベントは開催していません。しかし、我々のコミュニティに関連するものであれば、対面のイベントでも同様の行動規範を持つ必要があります。 +この行動規範は、NumPy プロジェクトによって管理されるすべての場所で適用されます。 この場所とは、すべてのパブリックおよびプライベートのメーリングリスト、イシュートラッカー、Wiki、ブログ、Twitter、コミュニティで使用されているその他の通信チャンネルなどを含みます。 NumPy プロジェクトでは対面でのイベントは開催していません。 しかし、我々のコミュニティに関連するものであれば、対面のイベントでも同様の行動規範を持つ必要があります。 -この行動規範は、NumPy コミュニティに正式または非公式に参加するすべての人が順守する必要があります。その他にも、NumPyとの提携・関連するプロジェクト活動においては、特にそれらのプロジェクトを代表する場合、同様の行動規範に従う必要があります。 +この行動規範は、NumPy コミュニティに正式または非公式に参加するすべての人が順守する必要があります。 その他にも、NumPyとの提携・関連するプロジェクト活動においては、特にそれらのプロジェクトを代表する場合、同様の行動規範に従う必要があります。 この行動規範は完全ではありません。 しかし、行動規範は我々が理解すべき、互いの協力の仕方や、共通の場所のあるべき姿、我々のゴールなどをまとめるのに重要な役目を果たします。 フレンドリーで生産的な環境を生み出し、周囲のコミュニティにより良い影響を与えるため、ぜひこの行動規範に従ってください。 @@ -18,14 +18,14 @@ aliases: 私たちは下記の内容に真摯に取り組みます。 1. 開けたコミュニティにしましょう。 私たちは、誰でもコミュニティに参加できるようにします。 私たちは、公にすべきではない内容を議論する場合以外、プロジェクトに関連するメッセージを公の場で告知することを選びます。 これは、NumPyに関するヘルプやプロジェクトサポートにも適用されます。公式なサポートだけでなく、NumPyに関する質問に答える場合もです。 これにより、質問に答えた際の意図しない間違いを、より簡単に検出し、訂正できるようになります。 -2. 共感し、歓迎し、友好的で、そして我慢強くありましょう。 私たちは互いに争いを解決し合い、互いの善意を信じ合います。 私たちは時折り不満を感じるかもしれません。しかしそのような場合も、不満を個人的な攻撃に変えることは許容されません。 人々が不快や脅威を感じるコミュニティは、生産的ではないからです。 -3. 互いに協力し合おう。 私たちの開発成果は他の人々によって利用され、一方で、たちは他の人々の開発成果に依存しているのです。 私たちがプロジェクトために何かを作るとき、私たちはそれがどのように動作するかを他の人に説明する必要があります。しかし、この作業により、より良いものを作り上げることができるのです。 私たちが下す全ての決断は、ユーザと開発コミュニティに影響を与えうるし、その決断がもたらす結果を私たちは真摯に受け止めます。 -4. 好奇心を大事にしよう。 全てを知っている人はいないのです！ 早め早めに質問をすることで、後に生じうる多くの問題を回避できます。そのため私たちは質問を奨励しています。もっとも、その質問に対して、適切なフォーラムを紹介する場合もありますが。 私たちは、出来るだけ質問に良く対応し、手助けできるよう努力します。 -5. 使う言葉に注意しましょう。 私たちは、コミュニティにおけるコミュニケーションに注意と敬意を払います。そして、私たちは自分の言葉に責任を持ちます。 他人に優しくしましょう。 他のコミュニティの参加者を侮辱しないでください。 私たちは、以下のようなハラスメントやその他の排斥行為を許しません。: +2. 共感し、歓迎し、友好的で、そして我慢強くありましょう。 私たちは互いに争いを解決し合い、互いの善意を信じ合います。 私たちは時折り不満を感じるかもしれません。 しかしそのような場合も、不満を個人的な攻撃に変えることは許容されません。 人々が不快や脅威を感じるコミュニティは、生産的ではないからです。 +3. 互いに協力し合おう。 私たちの開発成果は他の人々によって利用され、一方で、たちは他の人々の開発成果に依存しているのです。 私たちがプロジェクトために何かを作るとき、私たちはそれがどのように動作するかを他の人に説明する必要があります。 しかし、この作業により、より良いものを作り上げることができるのです。 私たちが下す全ての決断は、ユーザと開発コミュニティに影響を与えうるし、その決断がもたらす結果を私たちは真摯に受け止めます。 +4. 好奇心を大事にしよう。 全てを知っている人はいないのです！ 早め早めに質問をすることで、後に生じうる多くの問題を回避できます。 そのため私たちは質問を奨励しています。 私たちは、出来るだけ質問に良く対応し、手助けできるよう努力します。 +5. 使う言葉に注意しましょう。 私たちは、コミュニティにおけるコミュニケーションに注意と敬意を払います。 そして、私たちは自分の言葉に責任を持ちます。 他人に優しくしましょう。 他のコミュニティの参加者を侮辱しないでください。 私たちは、以下のようなハラスメントやその他の排斥行為を許しません。 : * 他の人に向けられた暴力的な行為や言葉。 * 性差別や人種差別、その他の差別的なジョークや言動。 * 性的または暴力的な内容の投稿。 - * 他のユーザーの個人情報を投稿すること。(または投稿すると脅すこと)。 + * 他のユーザーの個人情報を投稿すること。 (または投稿すると脅すこと)。 * 公開目的のない電子メールや、ICRチャットのようなログの残らないフォーラムの履歴など、プライベートなコンテンツを送信者の同意なしに共有すること。 * 個人的な侮辱, 特に人種差別や性差別的な用語を使用して侮辱すること。 * 不快な思いをさせる性的な言動。 @@ -37,7 +37,7 @@ aliases: NumPyプロジェクトは、全ての人々の参加を歓迎しています。 私たちは、誰もがコミュニティの一員であることを楽しめるように尽力します。 全ての人の好みを満足はさせられないかもしれませんが、全員に対し出来るだけ親切な対応ができるよう最善を尽くします。 -あなたの自己認識や、他者のあなたへの認識は関係ありません。私たちはあなたを歓迎します。 完璧なリストは望むべくもありませんが、私たちは行動規範に反しない限り、下記の多様性を尊重すると明言します: 年齢、文化。 民族、遺伝、性同一性あるいは関連する表現、言語、国籍、神経学的な差異、生物学的な差異、 政治的信条、職業、人種、宗教、性的指向、社会経済的地位、文化的な差異、技術的な能力。 +あなたの自己認識や、他者のあなたへの認識は関係ありません。 私たちはあなたを歓迎します。 民族、遺伝、性同一性あるいは関連する表現、言語、国籍、神経学的な差異、生物学的な差異、 政治的信条、職業、人種、宗教、性的指向、社会経済的地位、文化的な差異、技術的な能力。 私たちはすべての種類の言語言語話者の参加を歓迎しますが、NumPy 開発は英語で行われます。 @@ -61,7 +61,7 @@ NumPy行動規範委員会に問題を報告する場合は、こちらにご連 ### インシデント報告の解決 & 行動規範の実施 -本節では、_最も重要な点のみをまとめます。_詳細については、[NumPy Code of Conduct - How to follow up on a report](/report-handling-manual) をご覧ください。 +本節では、_最も重要な点のみをまとめます。 _詳細については、[NumPy Code of Conduct - How to follow up on a report](/report-handling-manual) をご覧ください。 私たちはすべての訴えを調査し、対応するようにします。 NumPy行動規範委員会およびNumPy運営委員会(もし関係する場合) は、報告者の身元を保護します。 また(報告者が同意しない限り) 苦情の内容を機密として扱うこととします。 @@ -78,6 +78,6 @@ NumPy行動規範委員会に問題を報告する場合は、こちらにご連 ### 文末脚注: -私たちは下記のドキュメントを作成したグループに感謝します。内容・発想ともに大いに影響されています。 +私たちは下記のドキュメントを作成したグループに感謝します。 内容・発想ともに大いに影響されています。 -- [SciPy行動規範](https://docs.scipy.org/doc/scipy/reference/dev/conduct/code_of_conduct.html) +- [SciPy行動規範](https://docs.scipy.org/doc/scipy/dev/conduct/code_of_conduct.html) diff --git a/content/ja/community.md b/content/ja/community.md index 35a58f4a7f..2629f72358 100644 --- a/content/ja/community.md +++ b/content/ja/community.md @@ -17,7 +17,7 @@ NumPy プロジェクトやコミュニティと直接交流する方法は次 このメーリングリストは、NumPy に新しい機能を追加するなど、より長い期間の議論のための主なコミュニケーションの場です。 NumPyのRoadmapに変更を加えたり、プロジェクト全体での意思決定を行います。 このメーリングリストでは、リリース、開発者会議、スプリント、カンファレンストークなど、NumPy についてのアナウンスなどにも利用されます。 -このメーリングリストでは、一番下のメールを使用し、メーリングリストに返信して下さい( 他の送信者ではなく)。 また、自動送信のメールには返信しないでください。 このメーリングリストの検索可能なアーカイブは [こちら](https://mail.python.org/archives/list/numpy-discussion@python.org/) にあります。 +このメーリングリストでは、一番下のメールを使用し、メーリングリストに返信して下さい( 他の送信者ではなく)。 このメーリングリストの検索可能なアーカイブは [こちら](https://mail.python.org/archives/list/numpy-discussion@python.org/) にあります。 *** @@ -33,7 +33,7 @@ _ちなみに、セキュリティの脆弱性を報告するには、GitHubの ### [Slack](https://numpy-team.slack.com) -SlackはNumPyに_ 貢献するための質問をする_、リアルタイムのチャットルームです。 Slackはプライベートな空間です。具体的には、 公開のメーリングリストやGitHubで質問やアイデアを持ち出すことを躊躇している人々のためのものです。 Slackに招待してもらいたい場合は[こちら](https://numpy.org/devdocs/dev/index.html#contributing-to-numpy)を確認下さい。 +SlackはNumpyに_ 貢献するための質問をするための_、リアルタイムのチャットルームです。 具体的には、 公開のメーリングリストやGitHubで質問やアイデアを持ち出すことを躊躇している人々のためのものです。 Slackに招待してもらいたい場合は[こちら](https://numpy.org/devdocs/dev/index.html#contributing-to-numpy)を確認下さい。 ## 勉強会とミートアップ @@ -52,7 +52,7 @@ NumPy プロジェクトは独自のカンファレンスは開催していま - [SciPy Latin America](https://www.scipyla.org) - [SciPy India](https://scipy.in) - [SciPy Japan](https://conference.scipy.org) -- [PyData conference](https://pydata.org/event-schedule/) (年に15~20のイベントが様々な国で開催されています。) +- [PyData conference](https://pydata.org/event-schedule/) (年に15~20のイベントが様々な国で開催されています。 ) これらのカンファレンスの多くは、NumPyの使い方や関連するオープンソースプロジェクトに貢献する方法を学ぶことができるチュートリアルを開催しています。 @@ -63,4 +63,4 @@ NumPyプロジェクトを成功させるには、あなたの専門知識とプ もし、NumPyに貢献したい場合は、 [コントリビュート](/ja/contribute) ページをご覧いただくことをお勧めします。 - +また、私たちのコミュニティミーティングにもぜひ参加してみてください。 コミュニティミーティングの活動を確認するには、[こちら](https://scientific-python.org/calendars/)のイベントカレンダーを確認ください。 diff --git a/content/ja/config.yaml b/content/ja/config.yaml index f031c22750..ec90ad7023 100644 --- a/content/ja/config.yaml +++ b/content/ja/config.yaml @@ -1,139 +1,140 @@ -# FOR TRANSLATORS: this is a YAML file, with lines being of the form: -# -# key: value -# -# Please translate the `value`s, not the `key`s! -# Comments (starting with `#`) and `url:` or `link:` lines also do not -# need to be translated. `title:` and `text:` lines do need translation. -# languageName: 日本語 (Japanese) params: description: NumPyが広く利用される理由 強力な多次元配列、数値計算ツール群、相互運用性、高いパフォーマンス、オープンソース navbarlogo: image: logo.svg + text: NumPy link: /ja/ - hero: + #Main hero title title: NumPy + #Hero subtitle (optional) subtitle: Pythonによる科学技術計算の基礎パッケージ - buttontext: 使い始める - buttonlink: "/ja/install" + #Button text + buttontext: "最新リリース: Numpy 1.25. すべてのリリースを表示する" + #Where the main hero button links to + buttonlink: "/ja/news/#releases" + #Hero image (from static/images/___) image: logo.svg - - news: - title: NumPy v1.20.0 - content: 型アノテーションサポート - 複数のプラットフォームにおけるSIMDを利用したパフォーマンス改善 - url: /ja/news - shell: - title: placeholder # do not translate - + title: placeholder intro: - - title: Try NumPy - text: Use the interactive shell to try NumPy in the browser - - docslink: Don't forget to check out the docs. - + - + title: NumPy を試す + text: インタラクティブシェルを使用して、ブラウザ上で Numpy を試してみてください。 + docslink: ドキュメント を確認することを忘れないでください。 casestudies: title: ケーススタディ features: - - title: 世界初のブラックホール画像 - text: NumPyはどのように、SciPyやMatplotlibなどのNumPyに依存するライブラリとともに、イベントホライズンテレスコープによる世界初のブラックホール画像の作成を可能にしたのでしょうか。 - img: /images/content_images/case_studies/blackhole.png - alttext: 世界初のブラックホール画像。黒い背景にオレンジ色の円で描かれています。 - url: /ja/case-studies/blackhole-image - - title: 重力波の検知 - text: 1916年、アルバート・アインシュタインは重力波を予言しました。100年後、LIGOの研究者たちはNumPyを使ってその存在を確認しました。 - img: /images/content_images/case_studies/gravitional.png - alttext: 2つのオーブがお互いに周回し、周りの重力を変位させています。 - url: /ja/case-studies/gw-discov - - title: スポーツ分析 - text: クリケット分析は、統計的モデリングと予測分析によって選手やチームのパフォーマンスを向上させることで、クリケットの試合を変えようとしています。多くの分析が、NumPyにより可能になりました。 - img: /images/content_images/case_studies/sports.jpg - alttext: 緑のフィールド上にあるクリケットボール。 - url: /ja/case-studies/cricket-analytics - - title: 深層学習による姿勢推定 - text: DeepLabCutはNumPyを利用し、種族・時間スケールによらない運動制御の理解へ向け、動物の行動観察を含む科学技術研究を加速しています。 - img: /images/content_images/case_studies/deeplabcut.png - alttext: チータの姿勢推定 - url: /ja/case-studies/deeplabcut-dnn - - keyfeatures: - features: - - title: 強力な多次元配列 - text: NumPyの高速で多機能なベクトル化計算、インデックス処理、ブロードキャスティングのコンセプトは、今日の配列計算のデファクト・スタンダードです。 - - title: 数値計算ツール群 - text: NumPyは、様々な数学関数、乱数生成器、線形代数ルーチン、フーリエ変換などを提供しています。 - - title: 相互運用性 - text: NumPyは、幅広いハードウェアとコンピューティング・プラットフォームをサポートしており、分散処理、GPU、疎行列ライブラリにも対応しています。 - - title: 高パフォーマンス - text: NumPyの中核は最適化されたC言語のコードです。Pythonの柔軟性を、コンパイルされたコードの高速さとともに享受できます。 - - title: 使いやすさ - text: NumPyの高水準なシンタックスは、どんなバックグラウンドや経験値のプログラマーでも利用でき、生産性を高めることができます。 - - title: オープンソース - text: 寛容な[BSDライセンス](https://github.com/numpy/numpy/blob/main/LICENSE.txt)で公開されています。NumPyは活発で、互いを尊重し、多様性を認め合う[コミュニティ](/ja/community)によって、 [GitHub](https://github.com/numpy/numpy)上でオープンに開発されています. - + - + title: 世界初のブラックホール画像 + text: NumPyはどのように、SciPyやMatplotlibなどのNumPyに依存するライブラリとともに、イベントホライズンテレスコープによる世界初のブラックホール画像の作成を可能にしたのでしょうか。 + img: /images/content_images/case_studies/blackhole.png + alttext: 世界初のブラックホール画像。黒い背景にオレンジ色の円で描かれています。 + url: /ja/case-studies/blackhole-image + - + title: 重力波の検知 + text: 1916年、アルバート・アインシュタインは重力波を予言しました。100年後、LIGOの研究者たちはNumPyを使ってその存在を確認しました。 + img: /images/content_images/case_studies/gravitional.png + alttext: 2つのオーブがお互いに周回し、周りの重力を変位させています。 + url: /ja/case-studies/gw-discov + - + title: スポーツ分析 + text: クリケット分析は、統計的モデリングと予測分析によって選手やチームのパフォーマンスを向上させることで、クリケットの試合を変えようとしています。多くの分析が、NumPyにより可能になりました。 + img: /images/content_images/case_studies/sports.jpg + alttext: 緑のフィールド上にあるクリケットボール。 + url: /ja/case-studies/cricket-analytics + - + title: 深層学習による姿勢推定 + text: DeepLabCutはNumPyを利用し、動物の種類や時間スケールによらない運動制御の理解へ向け、動物の行動観察を含む科学技術研究を加速させています。 + img: /images/content_images/case_studies/deeplabcut.png + alttext: チータの姿勢推定 + url: /ja/case-studies/deeplabcut-dnn tabs: - title: エコシステム + title: NumPyのエコシステム section5: false - -navbar: - - title: インストール - url: /ja/install - - title: ドキュメント - url: https://numpy.org/doc/stable - - title: 学び方 - url: /ja/learn - - title: コミュニティ - url: /ja/community - - title: 私達について - url: /ja/about - - title: NumPyに貢献する - url: /ja/contribute - -footer: - logo: logo.svg - socialmediatitle: "" - socialmedia: - - link: https://github.com/numpy/numpy - icon: github - - link: https://www.youtube.com/channel/UCguIL9NZ7ybWK5WQ53qbHng - icon: youtube - - link: https://twitter.com/numpy_team - icon: twitter - quicklinks: - column1: - title: "" - links: - - text: インストール - link: /ja/install - - text: ドキュメント - link: https://numpy.org/doc/stable - - text: 学び方 - link: /ja/learn - - text: 引用する - link: /ja/citing-numpy - - text: ロードマップ - link: https://numpy.org/neps/roadmap.html - column2: - links: - - text: 私達について - link: /ja/about - - text: コミュニティ - link: /ja/community - - text: User surveys - link: /ja/user-surveys - - text: NumPyに貢献する - link: /ja/contribute - - text: 行動規範 - link: /ja/code-of-conduct - column3: - links: - - text: サポートを得る方法 - link: /ja/gethelp - - text: 利用規約 - link: /ja/terms - - text: プライバシーポリシー - link: /ja/privacy - - text: プレス用資料 - link: /ja/press-kit + navbar: + - + title: インストール + url: /install + - + title: ドキュメント + url: https://numpy.org/doc/stable + - + title: 学習方法 + url: /learn + - + title: コミュニティ + url: /community + - + title: 私たちについて + url: /about + - + title: ニュース + url: /news + - + title: 貢献 + url: /contribute + footer: + logo: logo.svg + socialmediatitle: "" + socialmedia: + - + link: https://github.com/numpy/numpy + icon: github + - + link: https://www.youtube.com/channel/UCguIL9NZ7ybWK5WQ53qbHng + icon: youtube + - + link: https://twitter.com/numpy_team + icon: twitter + quicklinks: + column1: + title: "" + links: + - + text: インストール + link: /install + - + text: ドキュメント + link: https://numpy.org/doc/stable + - + text: 学習方法 + link: /learn + - + text: NumPy を引用する + link: /citing-numpy + - + text: ロードマップ + link: https://numpy.org/neps/roadmap.html + column2: + links: + - + text: 私達について + link: /about + - + text: コミュニティ + link: /community + - + text: ユーザ調査 + link: /user-surveys + - + text: 貢献 + link: /contribute + - + text: 行動規範 + link: /code-of-conduct + column3: + links: + - + text: サポートを得る方法 + link: /gethelp + - + text: 利用規約 + link: /terms + - + text: プライバシーポリシー + link: /privacy + - + text: プレス用資料 + link: /press-kit diff --git a/content/ja/contribute.md b/content/ja/contribute.md index 1963693a98..90db608852 100644 --- a/content/ja/contribute.md +++ b/content/ja/contribute.md @@ -3,32 +3,20 @@ title: NumPy に貢献する sidebar: false --- -NumPyプロジェクトを成功させるには、あなたの専門知識とプロジェクトに関する熱意が必要です。 NumPyに貢献する方法はコーディングだけではありません。 +NumPyプロジェクトを成功させるには、あなたの専門知識とプロジェクトに関する熱意が必要です。 貢献方法はプログラミングに限定されません。 このページには**あなたができる** 様々な種類の貢献方法が示されています。 -- [コードを書く]({{< relref "contribute.md#writing-code" >}}) +もしどこから始めればいいか、あなたのスキルをどう生かせばいいかがわからない場合は、 _是非ご連絡下さい。 _ 連絡の方法としては、 [メーリングリスト](https://mail.python.org/mailman/listinfo/numpy-discussion) 、 [GitHub](http://github.com/numpy/numpy)、 [イシューの作成](https://github.com/numpy/numpy/issues) 、関連するイシューへのコメントがあります。 -以外にも、下記の貢献の方法があります: +連絡先としては、 または、[Slack](https://numpy-team.slack.com) (グループに招待するためにこちらに連絡お願いします: )があります。 -- [プルリクエストをレビューする]({{< relref "contribute.md#reviewing-pull-requests" >}}) -- [チュートリアル・プレゼンテーションなど教育的資料を作成する]({{< relref "contribute.md#developing-educational-materials" >}}) -- [イシューをトリアージする]({{< relref "contribute.md#issue-triaging" >}}) -- [ウェブサイトをメンテナンスをする]({{< relref "contribute.md#website-development" >}}) -- [グラフィックデザインに貢献する]({{< relref "contribute.md#graphic-design" >}}) -- [ウェブサイトを翻訳する]({{< relref "contribute.md#translating-website-content" >}}) -- [コミュニティのコーディネーターをつとめる]({{< relref "contribute.md#community-coordination-and-outreach" >}}) -- [助成金のプロポーザルを書くなど、資金調達をサポートする]({{< relref "contribute.md#fundraising" >}}) - -もしどこから始めればいいか、あなたのスキルをどう生かせばいいかがわからない場合は、 _是非ご連絡下さい。_ 連絡の方法としては、 [メーリングリスト](https://mail.python.org/mailman/listinfo/numpy-discussion) 、 [GitHub](http://github.com/numpy/numpy)、 [イシューの作成](https://github.com/numpy/numpy/issues) 、関連するイシューへのコメントがあります。 - -これらが私たちにとって好ましい連絡手段ですが(元来、オープンソースプロジェクトはオープンな方法を好みます)、もしどうしても非公開の方法で連絡を取りたい場合は、コミュニティコーディネーターに連絡して下さい。連絡先としては、 または、[Slack](https://numpy-team.slack.com) (グループに招待するためにこちらに連絡お願いします: )があります。 - -また、隔週の _コミュニティミーティング_もあり、詳細は [メーリングリスト](https://mail.python.org/mailman/listinfo/numpy-discussion) で発表されています。あなたの参加を大いに歓迎します。オープンソースプロジェクトに貢献するのが初めての方は、是非、 [このガイド](https://opensource.guide/how-to-contribute/) を読んでみて下さい。 +また、隔週の _コミュニティミーティング_もあり、詳細は [メーリングリスト](https://mail.python.org/mailman/listinfo/numpy-discussion) で発表されています。 あなたの参加を大いに歓迎します。 オープンソースプロジェクトに貢献するのが初めての方は、是非、 [このガイド](https://opensource.guide/how-to-contribute/) を読んでみて下さい。 私たちのコミュニティは、誰もが平等に扱われ、すべての貢献を平等に評価することを目指しています。 私たちはオープンで居心地の良いコミュニティを作るために [行動基準](/ja/code-of-conduct) を制定しています。 ### コードを書く -プログラマーの方には、こちらの [ガイド](https://numpy.org/devdocs/dev/index.html#development-process-summary)でNumPyのコードに貢献する方法を説明しています。 +プログラマーの方には、こちらの [ガイド](https://numpy.org/devdocs/dev/index.html#development-process-summary)でNumPyのコードに貢献する方法を説明しています。
追加情報に関しては、 こちらの[YouTube チャンネル](https://www.youtube.com/playlist?list=PLCK6zCrcN3GXBUUzDr9L4__LnXZVtaIzS) もご覧ください。 + ### プルリクエストのレビュー NumPyプロジェクトには現時点で250以上のオープンなプルリクエストがあり、多くの 改善要求と多くのレビュワーからのフィードバックを待っています。 もしあなたがNumPy を使ったことがある場合、 たとえNumPyコードベースに慣れていない場合でも貢献する方法はあります。 例えば、 @@ -44,7 +32,7 @@ NumPy の [ユーザガイド](https://numpy.org/devdocs) は現在、大規模 ### イシューのトリアージ -[NumPyのイシュートラッカー](https://github.com/numpy/numpy/issues) には、 _沢山の_Open状態のイシューがあります。すでに解決されたもの、優先順位付けされるべきもの、 初心者が取り組むのに適したものがあります。あなたができることは、いくつもあります: +[NumPyのイシュートラッカー](https://github.com/numpy/numpy/issues) には、 _沢山の_Open状態のイシューがあります。 すでに解決されたもの、優先順位付けされるべきもの、 初心者が取り組むのに適したものがあります。 あなたができることは、いくつもあります: * 古いバグがまだ残っているか確認する * 重複したイシューを見つけ、お互いに関連づける @@ -56,12 +44,12 @@ NumPy の [ユーザガイド](https://numpy.org/devdocs) は現在、大規模 ### ウェブサイトの開発 -私たちはちょうどウェブサイトを作り直し始めたところですが、それらはまだ完了していません。Web開発が好きなら、この[イシュー](https://github.com/numpy/numpy.org/issues?q=is%3Aissue+is%3Aopen+label%3Adesign) に未完成な要求が列挙されています。ぜひ、あなたのアイデアを共有してください。 +私たちはちょうどウェブサイトを作り直し始めたところですが、それらはまだ完了していません。 Web開発が好きなら、この[イシュー](https://github.com/numpy/numpy.org/issues?q=is%3Aissue+is%3Aopen+label%3Adesign) に未完成な要求が列挙されています。 ぜひ、あなたのアイデアを共有してください。 ### グラフィックデザイン -グラフィックデザイナーの方が可能な貢献は、枚挙にいとまがありません。私たちのドキュメントには可視化が必要で、私たちの拡大しているウェブサイトには良い画像が必要です。貢献する機会は沢山あります。 +グラフィックデザイナーの方が可能な貢献は、枚挙にいとまがありません。 しかし、私たちのドキュメントは説明のために可視化が重要であり、私たちの拡大しているウェブサイトは良い画像を求めていることから、 貢献する機会が沢山あると言えます。 ### ウェブサイトの翻訳 @@ -75,5 +63,4 @@ NumPy の [ユーザガイド](https://numpy.org/devdocs) は現在、大規模 ### 資金調達 -NumPyは何年にも渡ってボランティアだけ活動していましたが、その重要性が高まるにつれ、安定性と成長のためには資金面での支援が必要であることがわかってきました。 こちらの[SciPy'19のプレゼン](https://www.youtube.com/watch?v=dBTJD_FDVjU) では、資金的なサポートを受けたことで、どれだけ違いが出たかを説明しています。 他の非営利団体のように、私たちは助成金や、スポンサーシップ、その他の資金支援を常に探しています。 私たちはすでにいくつかの資金調達のアイデアを持っていますが、他にもより多くを資金調達を受けたいと思っています。 資金調達に関する知識は、我々には不足しているスキルです。是非、あなたのサポートをお待ちしています。 - +NumPyは何年にも渡ってボランティアだけ活動していましたが、その重要性が高まるにつれ、安定性と成長のためには資金面での支援が必要であることがわかってきました。 こちらの[SciPy'19のプレゼン](https://www.youtube.com/watch?v=dBTJD_FDVjU) では、資金的なサポートを受けたことで、どれだけ違いが出たかを説明しています。 他の非営利団体のように、私たちは助成金や、スポンサーシップ、その他の資金支援を常に探しています。 私たちはすでにいくつかの資金調達のアイデアを持っていますが、他にもより多くを資金調達を受けたいと思っています。 資金調達に関する知識は、我々には不足しているスキルです。 是非、あなたのサポートをお待ちしています。 diff --git a/content/ja/gethelp.md b/content/ja/gethelp.md index 8ac9f7b9b4..6105ed9eb4 100644 --- a/content/ja/gethelp.md +++ b/content/ja/gethelp.md @@ -3,15 +3,13 @@ title: サポートを得る方法 sidebar: false --- -**ユーザーからの質問:** ユーザーからの質問に対して回答を得る最も良い方法は、[StackOverflow](http://stackoverflow.com/questions/tagged/numpy)に質問を投稿することです。すでに数千ものユーザーからの回答を見ることができます。 規模は小さいですが、下記のような質問をする場所もあります: [IRC](https://webchat.freenode.net/?channels=%23numpy)、 [Gitter](https://gitter.im/numpy/numpy)、 [Reddit](https://www.reddit.com/r/Numpy/)。 私たちはこれらのサイトを定期的に確認して、直接質問に答えるようにしていますが、質問の数は膨大です。 - -**開発関連の問題:** NumPyの開発関連の問題 (例: バグレポート) については、[コミュニティ](/community) のページを参照してください。 - +**Development issues:** For NumPy development-related matters (e.g., bug reports), please see [Community](/community). +**User questions:** The best way to get help is to post your question to a site like [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) or [Reddit](https://www.reddit.com/r/Numpy/). We wish we could keep an eye on these sites, or answer questions directly, but the volume is a little overwhelming! ### [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) -NumPyの使用方法に関する質問をするためのフォーラムです。例えば、「NumPyでXをするにはどうすればいいですか？」というような質問です。 質問をする時は、[ `#numpy` タグ](https://stackoverflow.com/help/tagging) を使用してください。 +NumPyの使用方法に関する質問をするためのフォーラムです。 例えば、「NumPyでXをするにはどうすればいいですか？ 質問をする時は、[ `#numpy` タグ](https://stackoverflow.com/help/tagging) を使用してください。 *** @@ -20,15 +18,3 @@ NumPyの使用方法に関する質問をするためのフォーラムです。 もう一つの使い方に関する質問の場です。 *** - -### [Gitter](https://gitter.im/numpy/numpy) - -ユーザーとコミュニティメンバーがお互いに助け合うリアルタイムのチャットルームです。 - -*** - -### [IRC](https://webchat.freenode.net/?channels=%23numpy) - -ユーザーとコミュニティメンバーがお互いを助け合うもう一つのリアルタイムチャットルームです。 - -*** diff --git a/content/ja/history.md b/content/ja/history.md index 75ae93fbed..04a5eb6432 100644 --- a/content/ja/history.md +++ b/content/ja/history.md @@ -3,20 +3,23 @@ title: NumPyの歴史 sidebar: false --- -NumPy は配列データ構造と配列に関連する高速な数値ルーチンを提供する Python 基礎的なライブラリです。 開始当初は資金も少なく、主に大学院生により開発されていました。その多くはコンピュータサイエンスの教育を受けておらず、指導教官のサポートも受けていませんでした。少数の "野良"学生プログラマーのグループが、すでに確立されていた商用研究ソフトウェアのエコシステムをひっくり返すなんて、想像することすら馬鹿げていました。 -商用ソフトは、何百万もの資金と何百人もの優秀なエンジニアに支えられていましたから。それでも、独特の視点を持つ熱狂的でフレンドリーなコミュニティに助けられ、完全にオープンなツールスタックの背後にある哲学的な動機は、長い目では日の目を見てきました。現在では、NumPyは科学者、技術者、および世界中の多くの専門家によって信頼され、使われています。 例えば、重力波の解析に用いられた公開スクリプトはNumPyを利用していますし、「M87ブラックホール画像化プロジェクト」では、直接NumPyを引用しています。 +NumPy は配列データ構造と配列に関連する高速な数値ルーチンを提供する Python 基礎的なライブラリです。 開始当初は資金も少なく、主に大学院生により開発されていました。その多くはコンピュータサイエンスの教育を受けておらず、指導教官のサポートも受けていませんでした。少数の "野良"学生プログラマーのグループが、すでに確立されていた商用研究ソフトウェアのエコシステムをひっくり返すなんて、想像することすら馬鹿げていました。 商用ソフトは、何百万もの資金と何百人もの優秀なエンジニアに支えられていましたから。それでも、独特の視点を持つ熱狂的でフレンドリーなコミュニティに助けられ、完全にオープンなツールスタックの背後にある哲学的な動機は、長い目では日の目を見てきました。現在では、NumPyは科学者、技術者、および世界中の多くの専門家によって信頼され、使われています。 例えば、重力波の解析に用いられた公開スクリプトはNumPyを利用していますし、「M87ブラックホール画像化プロジェクト」では、直接NumPyを引用しています。 このライブラリの開発開始当初は資金も少なく、主に大学院生が開発していましたが、その多くはコンピュータサイエンスの教育を受けておらず、指導教官のサポートも受けていませんでした。 何百万もの資金調達と何百人もの優秀なエンジニアに支えられている当時の商用研究ソフトウェアのエコシステムを、少数の "野良"学生プログラマーのグループがひっくり返すことができると想像することさえ、当時は馬鹿げていると考えられていました。 それでも、独特の視点を持つ熱狂的でフレンドリーなコミュニティに助けられ、完全にオープンなツールスタックの背後にある哲学的な動機は、長い目では日の目を見てきました。 現在では、Numpy は科学者、技術者、および世界中の多くの専門家によって信頼され、使われています。 例えば、重力波の解析に用いられた公開スクリプトはNumPyを利用していますし、「M87ブラックホール画像化プロジェクト」では、直接NumPyを引用しています。 NumPy および関連ライブラリの開発におけるマイルストーンの詳細については、 [arxiv.org](arxiv.org/abs/1907.10121) を参照してください。 NumPyのベースとなったNumericとNumarrayライブラリのコピーを入手したい場合は、以下のリンクを参照してください。 -[ *Numeric*](https://sourceforge.net/projects/numpy/files/Old%20Numeric/) のダウンロード* +[ *Numeric*](https://sourceforge.net/projects/numpy/files/Old%20Numeric/) のダウンロード** -[*Numarray *](https://sourceforge.net/projects/numpy/files/Old%20Numarray/) のダウンロード* +[*Numarray *](https://sourceforge.net/projects/numpy/files/Old%20Numarray/) のダウンロード** -*これらの古いパッケージはもはや保守されていないことに注意してください。配列関連の処理をしたい場合は、NumPyを使用するか、NumPyライブラリを利用するために既存のコードをリファクタリングすることを強くお勧めします。 +*これらの古いパッケージはもはや保守されていないことに注意してください。 配列関連の処理をしたい場合は、NumPyを使用するか、NumPyライブラリを利用するために既存のコードをリファクタリングすることを強くお勧めします。

-### 過去の資料 +

+ 過去の資料 +

-[*Numeric*マニュアルのダウンロード](static/numeric-manual.pdf) +

+ Numericマニュアルのダウンロード +

diff --git a/content/ja/install.md b/content/ja/install.md index 8a4e049d09..fad90210ea 100644 --- a/content/ja/install.md +++ b/content/ja/install.md @@ -3,9 +3,9 @@ title: NumPyのインストール sidebar: false --- -NumPy をインストールするために必ず必要なものはPython本体です。もしまだPythonをインストールしていないのであれば、最もシルプルな始め方として、[Anaconda Distribution](https://www.anaconda.com/distribution)を推奨します。AnacondaはPython・NumPyの他に、科学技術計算やデータサイエンスのために一般的に使用される沢山のパッケージが含まれています。 +NumPyをインストールするための唯一必要なものは、Pythonそのものだけです。 もしまだPythonをイントールしておらず、最もシンプルなインストール方法をお探しなら、[Anaconda Distribution](https://www.anaconda.com/distribution)の使用をおすすめします。これにはPython、NumPy、および科学計算やデータサイエンスでよく使われる様々な多くのパッケージが含まれています。 -NumPyは`conda`、`pip` 、macOSやLinuxのパッケージマネージャー、または [ソースコード](https://numpy.org/devdocs/user/building.html)からインストールすることが出来ます。 詳細な手順について、以下の [Python と NumPyの インストールガイド](#python-numpy-install-guide) を参照してください。 +NumPyは`conda`、`pip` 、macOSやLinuxのパッケージマネージャー、または [ソースコード](https://numpy.org/devdocs/user/building.html)からインストールすることが出来ます。 詳細な手順については、以下の [Python と Numpyの インストールガイド](#python-numpy-install-guide) を参照してください。 **CONDA** @@ -28,7 +28,8 @@ conda install numpy ```bash pip install numpy ``` -またpipを使う場合、仮想環境を使うことをおすすめします。[再現可能なインストール](#reproducible-installs)を参照ください。[こちらの記事](https://dev.to/bowmanjd/python-tools-for-managing-virtual-environments-3bko#howto)では仮想環境を使う詳細について説明されています。 +またpipを使う場合、仮想環境を使うことをおすすめします。 [再現可能なインストール](#reproducible-installs)を参照ください。 [こちらの記事](https://dev.to/bowmanjd/python-tools-for-managing-virtual-environments-3bko#howto)では仮想環境を使う詳細について説明されています。 + @@ -45,7 +46,7 @@ Pythonパッケージのインストールと管理は複雑なので、ほと Windows、macOS、Linuxのすべてのユーザー向けには: - [Anaconda](https://www.anaconda.com/distribution/) をインストールします（必要な パッケージと以下に挙げるすべてのツールがインストールされます）。 -- コードを書いたり、実行してみましょう。探索的・対話的コンピューティングには[JupyterLab](https://jupyterlab.readthedocs.io/en/stable/index.html)のノートブックが便利です。スクリプトやパッケージの作成には[Spyder](https://www.spyder-ide.org/)や[Visual Studio Code](https://code.visualstudio.com/)を利用できます。 +- コードを書いたり、実行してみましょう。 探索的・対話的コンピューティングには[JupyterLab](https://jupyterlab.readthedocs.io/en/stable/index.html)のノートブックが便利です。 スクリプトやパッケージの作成には[Spyder](https://www.spyder-ide.org/)や[Visual Studio Code](https://code.visualstudio.com/)を利用できます。 - [Anaconda Navigator](https://docs.anaconda.com/anaconda/navigator/) を使ってパッケージを管理し、JupyterLab、Spyder、Visual Studio Codeを使い始められます。 @@ -70,36 +71,37 @@ Windows、macOS、Linuxのすべてのユーザー向けには: ### Pipとconda -`pip` と `conda` がPythonパッケージをインストールするための2つの主要なツールです。 これら二つのツールの機能は部分的に重複しますが(例えば、両方とも `numpy`をインストールできます)、一緒に動作することもできます。ここでは、pip とcond の主要な違いについて説明します。これは、パッケージをどのように効果的に管理するかを理解したい場合、重要な知識です。 - -最初の違いは、condaは複数言語に対応可能で、Python自体をインストールできることです。pip はシステム上の特定の Python にインストールされ、パッケージはそのPython用にのみインストールします。そのため、condaはPython 以外のライブラリや必要なツール (コンパイラ、CUDA、HDF5など) をインストールできますが、pip はできません。 +`pip` と `conda` がPythonパッケージをインストールするための2つの主要なツールです。 これら二つのツールの機能は部分的に重複しますが(例えば、両方とも `numpy`をインストールできます)、一緒に動作することもできます。 ここでは、pip とcond の主要な違いについて説明します。 これは、パッケージをどのように効果的に管理するかを理解したい場合、重要な知識です。 2つ目の違いは、pipはPython Packaging Index(PyPI) からパッケージをインストールするのに対し、condaは独自のチャンネル(一般的には "defaults "や "conda-forge "など) からインストールすることです。 PyPIは最大のパッケージ管理システムですが、人気のある全てのパッケージがcondaでも利用可能です。 -3つ目の違いは、condaはパッケージ、依存関係、環境を管理するための統合されたソリューションであるのに対し、pipでは環境や複雑な依存関係を扱うために別のツール(たくさん存在しています！) が必要になるかもしれないということです。 +最初の違いは、condaは複数言語に対応可能で、Python自体をインストールできることです。 pip はシステム上の特定の Python にインストールされ、パッケージはそのPython用にのみインストールします。 PyPIは、最大のパッケージ管理システムですが、すべての代表的なパッケージは、condaにも利用可能です。 + +3つ目の違いは、condaはパッケージ、依存関係、環境を管理するための統合されたソリューションであるのに対し、pipでは環境や複雑な依存関係を扱うために別のツール(たくさん存在しています！ ### 再現可能なインストール -ライブラリが更新されると、コードの実行結果が変わったり、コードが完全に -壊れたりする可能性があります。なので重要なことは、使用しているパッケージの組み合わせと各バージョンのセットを再構築できるようにしておくことです。 ベストプラクティスは次の通りです: +ライブラリが更新されると、コードの実行結果が変わったり、コードが完全に 壊れたりする可能性があります。 なので重要なことは、使用しているパッケージの組み合わせと各バージョンのセットを再構築できるようにしておくことです。 ベストプラクティスは次の通りです: 1. プロジェクトごとに異なる仮想環境を使用して下さい。 -2. パッケージインストーラを使用してパッケージ名とバージョンを記録するようにして下さい。それぞれ、独自のメタデータフォーマットがあります: +2. パッケージインストーラを使用してパッケージ名とバージョンを記録するようにして下さい。 それぞれ、独自のメタデータフォーマットがあります: - condaの場合: [conda environmentsとenvironment.yml](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html) - pipの場合: [仮想環境](https://docs.python.org/3/tutorial/venv.html) と [requirements.txt](https://pip.readthedocs.io/en/latest/user_guide/#requirements-files) - poetryの場合: [仮想環境とpyproject.toml](https://python-poetry.org/docs/basic-usage/) + + ## NumPyパッケージと高速線形代数ライブラリ -NumPy は他の Python パッケージに依存していませんが、高速な線形代数ライブラリに依存しています。典型的には、[インテル® MKL](https://software.intel.com/en-us/mkl)や[OpenBLAS](https://www.openblas.net/)がこれにあたります。ユーザーは、これらの線形代数ライブラリのインストールを心配する必要はありません (NumPyのインストール方法に、あらかじめ含まれているためです)。 高度なユーザーは、使用されているBLASがパフォーマンスや、動作、ディスク上のサイズに影響を与えるため、より詳細を知りたがるかもしれません。 +NumPy は他の Python パッケージに依存していませんが、高速な線形代数ライブラリに依存しています。 典型的には、[インテル® MKL](https://software.intel.com/en-us/mkl)や[OpenBLAS](https://www.openblas.net/)がこれにあたります。 ユーザーは、これらの線形代数ライブラリのインストールを心配する必要はありません (NumPyのインストール方法に、あらかじめ含まれているためです)。 高度なユーザーは、使用されているBLASがパフォーマンスや、動作、ディスク上のサイズに影響を与えるため、より詳細を知りたがるかもしれません。 -- pipでインストールされるPyPI上の NumPy wheelは、OpenBLASを使ってビルドされます。つまりwheelにはOpenBLASライブラリが含まれています。そのため、ユーザが（例えば）SciPyも同じようにインストールした場合、ディスク上にOpenBLASのコピーをNumPyのものと2つ持つことになります +- pipでインストールされるPyPI上の NumPy wheelは、OpenBLASを使ってビルドされます。 つまりwheelにはOpenBLASライブラリが含まれています。 そのため、ユーザが（例えば）SciPyも同じようにインストールした場合、ディスク上にOpenBLASのコピーをNumPyのものと2つ持つことになります -- condaのデフォルトチャンネルでは、NumPy はインテル® MKLを使ってビルドされます。MKLはNumPyのインストール時に、独立したパッケージとしてユーザー環境にインストールされます。 +- condaのデフォルトチャンネルでは、NumPy はインテル® MKLを使ってビルドされます。 MKLはNumPyのインストール時に、独立したパッケージとしてユーザー環境にインストールされます。 -- conda-forgeのチャンネルでは、NumPyはダミーの「BLAS」パッケージを使ってビルドされています。 ユーザーがconda-forgeからNumPyをインストールすると、BLASパッケージが実際のライブラリと一緒にインストールされます。デフォルトはOpenBLASですが、MKL(default チャンネルの場合)や [BLIS](https://github.com/flame/blis)、またはBLASを利用することもできます。 +- conda-forgeのチャンネルでは、NumPyはダミーの「BLAS」パッケージを使ってビルドされています。 ユーザーがconda-forgeからNumPyをインストールすると、BLASパッケージが実際のライブラリと一緒にインストールされます。 デフォルトはOpenBLASですが、MKL(default チャンネルの場合)や [BLIS](https://github.com/flame/blis)、またはBLASを利用することもできます。 - OpenBLASは約30MBですが、MKLパッケージはOpenBLASよりもはるかに大きく、ディスク上の約700MBです。 @@ -108,14 +110,16 @@ NumPy は他の Python パッケージに依存していませんが、高速な インストールサイズ、パフォーマンスとロバスト性に加えて、考慮すべき2つの点があります: - インテル® MKL はオープンソースではありません。 通常の使用では問題ではありませんが、 ユーザーが NumPy で構築されたアプリケーションを再配布する必要がある場合、これは 問題が発生する可能性があります。 +- MKLとOpenBLASの両方とも、 np.dotのような関数呼び出しにマルチスレッドを使用し、スレッド数はビルド時オプションと環境変数の両方で決定されます。 多くの場合、すべての CPU コアが使用されます。 これにユーザーにとっては予想外のことかもしれません。 NumPy 自体は、関数呼び出しを自動的に並列化しないからです。 自動並列化により、一般にはパフォーマンスが向上しますが、逆にパフォーマンスが悪化する場合もあります。 例えば、Daskやscikit-learn、multiprocessingなど別のレベルの並列化を使用している場合です。 -- MKLとOpenBLASの両方とも、 np.dotのような関数呼び出しにマルチスレッドを使用し、スレッド数はビルド時オプションと環境変数の両方で決定されます。 多くの場合、すべての CPU コアが使用されます。 これにユーザーにとっては予想外のことかもしれません。NumPy 自体は、関数呼び出しを自動的に並列化しないからです。 自動並列化により、一般にはパフォーマンスが向上しますが、逆にパフォーマンスが悪化する場合もあります。例えば、Daskやscikit-learn、multiprocessingなど別のレベルの並列化を使用している場合です。 ## トラブルシューティング インストールに失敗した場合に、下記のエラーメッセージが表示される場合は、 トラブルシューティング ImportError を参照してください。 -
IMPORTANT: PLEASE READ THIS FOR ADVICE ON HOW TO SOLVE THIS ISSUE!
+```
+IMPORTANT: PLEASE READ THIS FOR ADVICE ON HOW TO SOLVE THIS ISSUE!
 
 Importing the numpy c-extensions failed. This error can happen for different reasons, often due to issues with your setup.
-
+``` + diff --git a/content/ja/learn.md b/content/ja/learn.md index e8c3b29503..7ab2dcde53 100644 --- a/content/ja/learn.md +++ b/content/ja/learn.md @@ -5,34 +5,29 @@ sidebar: false **公式の NumPy ドキュメント** については [numpy.org/doc/stable](https://numpy.org/doc/stable)を参照してください。 -## NumPyのチュートリアル - -[NumPyチュートリアル](https://numpy.org/numpy-tutorials)で、いくつかのチュートリアルと教育的資料を見ることができます。このページのゴールは、NumPyプロジェクトによる質のいい資料を提供することです。自習と講義形式の両方を想定しており、Jupyterノートブック形式で提供されます。もしあなた自身の資料を追加することに興味がある場合、[Github上のnumpy-tutorialsリポジトリ](https://github.com/numpy/numpy-tutorials)をチェックしてみて下さい。 - *** -以下は、キュレーションされた外部リソースのリストです。こちらのリストに貢献するには、 [このページの末尾](#add-to-this-list) を参照してください。 +以下は、Numpyへの貢献者とコミュニティによって開発された、NumPyの自己学習と他人への教育のための資料です。 ## 初心者向け NumPyについての資料は多数存在しています。 初心者の方にはこちらの資料を強くお勧めします： - **チュートリアル** + **動画** * [NumPy Quickstart チュートリアル](https://numpy.org/devdocs/user/quickstart.html) +* [NumPyチュートリアル](https://numpy.org/numpy-tutorials)で、いくつかのチュートリアルと教育的資料を見ることができます。 このページのゴールは、NumPyプロジェクトによる質のいい資料を提供することです。 自習と講義形式の両方を想定しており、Jupyterノートブック形式で提供されます。 もしあなた自身の資料を追加することに興味がある場合、[Github上のnumpy-tutorialsリポジトリ](https://github.com/numpy/numpy-tutorials)をチェックしてみて下さい。 +* [イラストで学ぶNumPy *by Lev Maximov*](https://betterprogramming.pub/3b1d4976de1d?sk=57b908a77aa44075a49293fa1631dd9b) * [SciPyレクチャー](https://scipy-lectures.org/) NumPyだけでなく、科学的なPythonソフトウェアエコシステムを広く紹介しています。 * [NumPy: 初心者のための基本](https://numpy.org/devdocs/user/absolute_beginners.html) -* [機械学習プラス - ndarray入門](https://www.machinelearningplus.com/python/numpy-tutorial-part1-array-python-examples/) -* [Edureka - NumPy配列を例題で学ぶ ](https://www.edureka.co/blog/python-numpy-tutorial/) -* [Dataquest - NumPyチュートリアル: Python を使ったデータ解析](https://www.dataquest.io/blog/numpy-tutorial-python/) * [NumPy チュートリアル *by Nicolas Rougier*](https://github.com/rougier/numpy-tutorial) -* [Stanford CS231 *by Justin Johnson*](http://cs231n.github.io/python-numpy-tutorial/) +* [スタンフォード大学 CS231 *by Justin Johnson*](http://cs231n.github.io/python-numpy-tutorial/) * [NumPyユーザーガイド](https://numpy.org/devdocs) - **書籍** + **チュートリアル** * [NumPガイド *Travelis E. Oliphant著*](http://web.mit.edu/dvp/Public/numpybook.pdf) これは2006年の無料版の初版です 最新版(2015年)については、こちら [を参照ください](https://www.barnesandnoble.com/w/guide-to-numpy-travis-e-oliphant-phd/1122853007). -* [PythonからNumPyまで *Nicolas P. Rougier著*](https://www.labri.fr/perso/nrougier/from-python-to-numpy/) +* [PythonにおけるNumPy (発展編)](https://www.labri.fr/perso/nrougier/from-python-to-numpy/) * [エレガントなSciPy](https://www.amazon.com/Elegant-SciPy-Art-Scientific-Python/dp/1491922877) *Juan Nunez-Iglesias・Stefan van der Walt・Harriet Dashnow 著* また、「Python+SciPy」を題材にした[推薦本リスト](https://www.goodreads.com/shelf/show/python-scipy) もチェックしてみてください。 ほとんどの本にはNumPyを核とした「SciPyエコシステム」が説明されています。 @@ -47,43 +42,35 @@ NumPyについての資料は多数存在しています。 初心者の方に 高度なインデックス指定、分割、スタッキング、線形代数など、NumPyの概念をより深く理解するためには、これらの上級者向け資料を試してみてください。 - **チュートリアル** + **書籍** -* [NumPyエクササイズ100](http://www.labri.fr/perso/nrougier/teaching/numpy.100/index.html) *Nicolas P. Rougier* +* https://www.tutorialspoint.com/numpy/numpy_advanced_indexing.htm * [NumPyとSciPyへのイントロダクション](https://engineering.ucsb.edu/~shell/che210d/numpy.pdf) *M. Scott Shell著* * [NumPy救急キット](http://mentat.za.net/numpy/numpy_advanced_slides/) *Stéfan van der Walt著* -* [PythonにおけるNumPy (発展編)](https://www.geeksforgeeks.org/numpy-python-set-2-advanced/) -* [高度なインデックス指定](https://www.tutorialspoint.com/numpy/numpy_advanced_indexing.htm) -* [NumPyによる機械学習とデータ分析](https://www.machinelearningplus.com/python/numpy-tutorial-python-part2/) +* [NumPyチュートリアル](https://numpy.org/numpy-tutorials)で、いくつかのチュートリアルと教育的資料を見ることができます。 このページのゴールは、NumPyプロジェクトによる質のいい資料を提供することです。 自習と講義形式の両方を想定しており、Jupyterノートブック形式で提供されます。 もしあなた自身の資料を追加することに興味がある場合、[Github上のnumpy-tutorialsリポジトリ](https://github.com/numpy/numpy-tutorials)をチェックしてみて下さい。 - **書籍** + **チュートリアル** * [Pythonデータサイエンスハンドブック](https://www.amazon.com/Python-Data-Science-Handbook-Essential/dp/1491912057) *Jake Vanderplas著* * [Pythonデータ解析](https://www.amazon.com/Python-Data-Analysis-Wrangling-IPython/dp/1491957662) *Wes McKinney著* * [数値解析Python: NumPy, SciPy, Matplotlibによる数値計算とデータサイエンスアプリケーション](https://www.amazon.com/Numerical-Python-Scientific-Applications-Matplotlib/dp/1484242459) *Robert Johansson著* - **動画** + **書籍** * [アドバンスドNumPy - ブロードキャストルール・ストライド・高度なインデックス指定](https://www.youtube.com/watch?v=cYugp9IN1-Q) *Fan Nunuz-Iglesias著* -* [NumPy配列における高度なインデクシング処理](https://www.youtube.com/watch?v=2WTDrSkQBng) *by Amuls Academy* *** -## NumPyに関するトーク +## NumPyに関する講演 * [NumPyにおけるインデックス指定の未来](https://www.youtube.com/watch?v=o0EacbIbf58) *Jaime Fernadezによる* (2016) * [Pythonにおける配列計算の進化](https://www.youtube.com/watch?v=HVLPJnvInzM&t=10s) *Ralf Gommersによる* (2019) -* [NumPy: 今までどう変わってきて、今後どう変わっていくのか？](https://www.youtube.com/watch?v=YFLVQFjRmPY) *Matti Picusによる* (2019) +* [NumPy: 今までどう変わってきて、今後どう変わっていくのか？ ](https://www.youtube.com/watch?v=YFLVQFjRmPY) *Matti Picusによる* (2019) * [NumPyの内部](https://www.youtube.com/watch?v=dBTJD_FDVjU) *Ralf Gommers, Sebastian Berg, Matti Picus, Tyler Reddy, Stefan van der Walt, Charles Harrisによる* (2019) * [Pythonにおける配列計算の概要](https://www.youtube.com/watch?v=f176j2g2eNc) *Travis Oliphantによる* (2019) *** -## NumPy を引用する場合 +## NumPyを引用する もし、あなたの研究においてNumPyが重要な役割を果たし、論文でこのプロジェクトについて言及したい場合は、こちらの[ページ](/ja/citing-numpy)を参照して下さい。 - -## このページへの貢献 - - -このページのリストに新しいリンクを追加するには、[プルリクエスト](https://github.com/numpy/numpy.org/blob/main/content/en/learn.md)を使って提案してみて下さい。 あなたが推薦するものがこのページで紹介するに値する理由と、その情報によりどのような人が最も恩恵を受けるかを説明して下さい。 diff --git a/content/ja/news.md b/content/ja/news.md index 8152792994..8d0cfb5b39 100644 --- a/content/ja/news.md +++ b/content/ja/news.md @@ -1,12 +1,173 @@ --- title: ニュース sidebar: false +newsHeader: "NumPy 2.0 リリース日: 6月16日" +date: 2023-09-16 --- -### NumPy 1.20.0 リリース +### NumPy 2.0 リリース日: 6月16日 + +_ 2024年5月23日_ -- NumPy 2.0が2024年6月16日にリリースされる予定になりました！ このリリースは1年以上かけて我々が準備してきたもので、2006年以来のメジャーリリースとなります。 このリリースで重要なことは、多くの新機能とパフォーマンスの向上に加えて、 このリリースは、 **破壊的な変更** である Python と C API を含む、ABI への変更 が含まれています。 NumPyに依存しているパッケージやエンドユーザーのコードがこのは破壊的変更に適応する必要がある可能性があります。可能であれば、あなたのコードがNumPy `2.0.0rc2`で動作するかどうか確認をお願いします。 **詳細は下記をご覧ください:** + +- [NumPy 2.0移行ガイド](https://numpy.org/devdocs/numpy_2_0_migration_guide.html) +- [2.0.0 リリース ノート](https://numpy.org/devdocs/release/2.0.0-notes.html) +- ステータスアップデートお知らせに関する問題: [numpy#24300](https://github.com/numpy/numpy/issues/24300) + + +### NumFOCUSの年末の資金調達 +_2023年12月19日_ -- NumFOCUSは、年末キャンペーンでPyCharmチームと協力し、PyCharmライセンスの初回購入に30%の割引を提供しています。 2023年12月23日までのPyCharm購入による1年目の収益は全てNumFOCUSのプログラムに直接寄付されます。 + +購入される方はこちらのURLか: https://lp.jetbrains.com/support-data-science/ こちらのクーポンコードを利用してください: ISUPPORTDATASCIENCE  + +### NumPy 1.26.0 がリリースされました。 + +_2023年9月16日_ -- [NumPy 1.26.0](https://numpy.org/doc/stable/release/1.26.0-notes.html)がリリースされました。 今回のリリースの目玉機能は次のとおりです。 + +* Python 3.12.0 のサポート +* Cython 3.0.0 への互換性 +* Mesonビルドシステムの利用 +* SIMD サポートの改善 +* f2py のバグ修正, meson と bind(x) のサポート +* 更新された BLAS/LAPACK の高速化ライブラリのサポート + +Numpy 1.26.0 は 1.25 からの互換性を保持しています。Mesonビルドシステムへの移行とCython 3.0.0へのサポートが目的のリリースです。 合計20人がこのリリースに貢献し、59個のプルリクエストがマージされました。 + +このリリースでサポートされている Python のバージョンは3.9から 3.12 です。 + +### numpy.orgが日本語とポルトガル語で利用可能になりました + +_2023年4月2日_ -- numpy.orgが2つの言語で利用可能になりました： 日本語とポルトガル語。 熱心なボランティアがいなければ、このプロジェクトは不可能でした： + +_ポルトガル語_ +* Melissa Weber Mendonça (melissawm) +* Ricardo Prins (ricardoprins) +* Getúlio Silva (getuliosilva) +* Julio Batista Silva (jbsilva) +* Alexandre de Siqueira (alexdesiqueira) +* Alexandre B A Villares (villares) +* Vini Salazar (vinisalazar) + +_日本語：_ +* Atsushi Sakai (AtsushiSakai) +* KKunai +* Tom Kelly (TomKellyGenetics) +* Yuji Kanagawa (kngwyu) +* Tetsuo Koyama (tkoyama010) + +翻訳インフラストラクチャに関するプロジェクトは、CZIからの資金援助でサポートされています。 + +今後も、NumPyのウェブサイトをより多くの言語に翻訳したいと思っています。 もし手伝える場合は、Slack上のNumPy翻訳チームに連絡をお願います: https://join.slack.com/t/numpy-team/shared_invite/zt-1gokbq56s-bvEpo10Ef7aHbVtVFeZv2w. (#translation チャンネルを探してください) また、Scientific Pythonエコシステム全体のドキュメントや教育コンテンツのローカライズに取り組む翻訳チームも 立ち上げています。 このプロジェクトにも興味がある場合は、是非Scientific Python Discordに参加してください: https://discord.gg/khWtqY6RKr. (#translation チャンネルを探してください) + +### NumPy 1.25.0 リリース + +_2023年1月17日_ -- [Numpy 1.25.0](https://numpy.org/doc/stable/release/1.25.0-notes.html) がリリースされました。 今回のリリースの目玉機能は次のとおりです。 + +* MUSLのサポート。MUSLのWheelが準備されました。 +* 富士通のC/C++コンパイラサポート +* einsum でオブジェクト配列がサポートされるようになりました. +* 行列の置き換え(inplace)掛け算のサポート (`@=`). + +Numpy 1.25. リリースは引き続きdtypeの取り扱いと dtypeのプロモーションを改善し、実行速度を向上させ、 ドキュメントを明確化するための継続的な作業を続けて行く予定です。 将来の NumPy 2.0.0 に向けた準備作業も行われており、 多数の新規および期限切れの機能廃止が可能となってきています。 + +合計148人がこのリリースに貢献し、530個のプルリクエストが マージされました。 + +このリリースでサポートされている Python のバージョンは3.3.9 - 3.11 です。 + +### インクルーシブな文化の育成: 参加の募集 + +_2023年5月10日_ -- インクルーシブ・カルチャーの育成: 参加募集 + +NumPyプロジェクトの多様性とインクルージョンに関して、我々はどのようなことを実施すればいいでしょうか？ 興味がある方はこちらの [レポート](https://contributor-experience.org/docs/posts/dei-report/) を読んで参加する方法を確認してください。 + +### NumPy ドキュメンテーションチームのリーダーの変更 + +_2023年1月6日_ –- Mukulika PahariとRoss Barnowskiは、Melissa MendoncAudioに代わるNumPyドキュメンテーションチームの新しいリーダーとして任命されました。 私たちは、MelissaにNumPyの公式ドキュメントと教育資料に対するすべての貢献に感謝し、MukulikaとRossに新しい役割にステップアップしてもらったことに感謝します。 + +### NumPy 1.24.0 リリース + +_2022年12月18日_ -- [Numpy 1.24.0](https://numpy.org/doc/stable/release/1.24.0-notes.html) がリリースされました。 今回のリリースのハイライトは次のとおりです。 + +* スタッキング関数のための新しい"dtype"と"casting"キーワードの追加 +* F2PYの新機能追加とバグ修正 +* 多くの新しい非推奨(Deprecation)の追加 +* 多くの期限切れの非推奨(Deprecation)の削除 + +Numpy 1.25. リリースは引き続きdtypeの取り扱いと dtypeのプロモーションを改善し、実行速度を向上させ、 ドキュメントを明確化するための継続的な作業を続けて行く予定です。 dtype のプロモーションとクリーンアップの変更により、多数の新規と期限切れの非推奨が存在しています。 今回のリリースは、444個のプルリクエストと177人のコントリビューターによるものです。 サポートされている Python のバージョンは 3.8-3.11 です。 + +### Numpy 1.23.0 リリース + +_2022年1月22日_ -- [Numpy 1.23.0](https://numpy.org/doc/stable/release/1.23.0-notes.html) がリリースされました。 今回のリリースの目玉機能は次のとおりです。 + +* `loadtxt` がCで実装されたことによる、大幅なパフォーマンス向上 +* より簡単なデータ交換のためのPythonレベルでのDLPackの公開 +* 構造化されたdtypesのプロモーションと比較方法の変更 +* f2pyの改善 + +Numpy 1.23. リリースでは引き続きdtypeの取り扱いと dtypeのプロモーションを改善し、実行速度を向上させ、 ドキュメントを明確化するための継続的な作業を続けて行く予定です。 今回のリリースは、494個のプルリクエストと151人のコントリビューターによるものです。 このリリースでサポートされている Python のバージョンは 3.8 - 3.10 です。 Python 3.11がrc ステージに到達すると Python 3.11 もサポートされます。 + +### NumFOCUS DEI研究への参加募集 -_2021年1月30日_ -- [NumPy 1.20.0](https://numpy.org/doc/stable/release/1.20.0-notes.html) が利用可能になりました。 今回のリリースは180以上のコントリビューターのおかげで、これまでで最大の NumPyのリリースとなりました。 最も重要な2つの新機能は次のとおりです。 -- NumPyの大部分のコードに型注釈が追加されました。そして新しいサブモジュールである`numpy.typing`が追加されました。このサブモジュールは`ArrayLike` や`DtypeLike`という型注釈のエイリアスが定義されており、これによりユーザーやダウンストリームのライブラリはこの型注釈を使うことができます。 +_2022年4月13日_ -- NumPyは、[NumFOCUS](http://numfocus.org/)と協力して、[ある研究プロジェクト](https://numfocus.org/diversity-inclusion-disc/a-pivotal-time-in-numfocuss-project-aimed-dei-efforts?eType=EmailBlastContent&eId=f41a86c3-60d4-4cf9-86cf-58eb49dc968c)を進めており、これは[Gordon & Betty Moore Foundation](https://www.moore.org/)によって資金提供されています。このプロジェクトでは、オープンソースソフトウェアコミュニティにおいて、特に歴史的に代表されてこなかったグループからの貢献者が参加する際の障壁を理解することを目的としています。 この研究チームは、新しい貢献者、プロジェクトの開発者およびメンテナー、そして過去に貢献した方々に、NumPyに参加し貢献した経験について話を聞きたいと考えています。 + +**あなたの経験を共有することに興味がありますか?** + +もし興味がある場合は、研究目標、プライバシー、および 守秘義務に関する追加情報が記載されている、この簡単な[参加者の興味](https://numfocus.typeform.com/to/WBWVJSqe)フォームに記入をお願いします。 多様で包括的なオープンソースソフトウェアコミュニティの 成長と持続可能性のために、このプロジェクトへのあなたの参加は非常に大きな価値があります。 参加を受け入れられた人は、研究チームメンバーと30分間のインタビューに参加することになります。 + +### NumPy 1.19.2 リリース + +_2021年12月31日_ -- [Numpy 1.22.0](https://numpy.org/doc/stable/release/1.22.0-notes.html) がリリースされました。 今回のリリースの目玉機能は次のとおりです。 + +* メインの名前空間の型アノテーションは基本的に完了しました。 上流のコードは常に変化するものなので、さらなる改良が必要でしょうが、大きな作業は終わったと考えています。 これはおそらく、今回のリリースで最も目に見える改良でしょう。 +* 以前から提案されていた [array API 標準](https://data-apis.org/array-api/latest/) のベータ版が提供されています ( [NEP 47](https://numpy.org/neps/nep-0047-array-api-standard.html) を参照) 。 これは、CuPy や JAX などのライブラリで使用できる 関数の標準的なコレクションを作成するために必要なステップです。 +* NumPy に DLPack バックエンドが追加されました。 DLPack は、配列(テンソル) データ用の共通のデータ変換フォーマットを提供します。 +* `quantile`, `percentile`, および関連する関数に新しいメソッドが追加されました。 これらの新しいメソッドは、論文で一般的に見られる一通りの処理を提供します。 +* ユニバーサル関数は、[NEP 43](https://numpy.org/neps/nep-0043-extensible-ufuncs.html) の多くを実装するためにリファクタリングされました。 これにより将来の DType API の処理も可能にします。 +* ダウンストリームのプロジェクトで使用するための新しい設定可能なメモリー・アロケーターが追加されました。 + +NumPy 1.22.0は、153人の貢献者が609のプルリクエストを作成した 非常に大きなリリースです。 このリリースでサポートされている Python のバージョンは 3.8 - 3.10 です。 + +### 科学的なPythonエコシステムにおける包括的な文化の前進 + +_ 2021年8月31日_ -- この度、Chan Zuckerberg Initiativeより、科学的なPythonプロジェクトにおいて、歴史的に疎外されてきたグループの人々のオンボーディング、インクルージョン、リテンションを支援し、NumPy、SciPy、Matplotlib、Pandasのコミュニティダイナミクスを構造的に改善するための [ 助成金を授与されました ](https://chanzuckerberg.com/newsroom/czi-awards-16-million-for-foundational-open-source-software-tools-essential-to-biomedicine/) ことをお知らせします。 + +[ CZIのEssential Open Source Software for Scienceプログラム ](https://chanzuckerberg.com/eoss/)の一環として、この[ Diversity & Inclusion補助金 ](https://cziscience.medium.com/advancing-diversity-and-inclusion-in-scientific-open-source-eaabe6a5488b)は、開けたなオープンソースコミュニティを育成するためにやるべきことを特定したり、文書化したり、実施したりするためのコントリビュータ体験のリーダー専任職の創設を支援することになります。 このプロジェクトは、Melissa Mendonça (NumPy) が中心となって、下記の方々の追加のメンタリングとサポートにより実施されます。Ralf Gommers (NumPy、SciPy)、Hannah AizenmanとThomas Caswell (Matplotlib)、Matt Haberland (SciPy)、そして Joris Van den Bossche (Pandas)。 + +このプロジェクトは私たちのOSSプロジェクトのコミュニティダイナミクスを構造的に改善する方法を発見し、実施することを目指す野心的なプロジェクトです。 このような複数のプロジェクトの横断的な役割を確立することで、Scientific Pythonコミュニティに新しいコラボレーションモデルを導入し、エコシステム内のコミュニティ構築作業をより効率的に、より大きな成果を生めるようにしたいと考えています。 特にこのプロジェクトにより、歴史的にこれまで代表的ではなかったグループからの新しいコントリビュータを引き付け、貢献を維持するために、何がうまくいき、何がうまくいかないかを、より明確に把握できるようになると期待しています。 最後に、実施したアクションについて詳細な報告書を作成し、プロジェクトの代表者やコミュニティとの交流の面で、プロジェクトにどのような影響を与えたかを説明する予定です。 + +2021年11月から2年間のプロジェクトが始まると予想されており、このプロジェクトの成果を楽しみにしています! このプロジェクトの提案書に関しては、[こちら](https://figshare.com/articles/online_resource/Advancing_an_inclusive_culture_in_the_scientific_Python_ecosystem/16548063) から全文を読むことができます. + +### 2021年度NumPyアンケート + +_2021年7月12日_ -- NumPy ではコミュニティの力を信じています。 昨年の第1回アンケートには、75カ国から1,236名のNumPyユーザーが参加してくれました。 この調査結果により、今後12ヶ月間、私たちがどのようなことに集中すべきかを、非常に良く理解することができました。 + +今年もアンケートの時間が来ました。もう一度アンケートへの回答をお願いいたします。 アンケートへの回答は15分ほどで終了します。 アンケートは英語以外にも、ベンガル語、フランス語、ヒンディー語、日本語、マンダリン、ポルトガル語、ロシア語、スペイン語の8ヶ国語に対応しています。 + +こちらのリンク先から、アンケートを始めることができます: https://berkeley.qualtrics.com/jfe/form/SV_aaOONjgcBXDSL4q. + + +### NumPy 1.19.0 リリース + +_2021年1月23日_ -- [Numpy 1.21.0](https://numpy.org/doc/stable/release/1.21.0-notes.html) がリリースされました。 今回のリリースのハイライトは下記の通りです。 + +- より多くの機能やプラットフォームをカバーするためのSIMD関連の改善が実施されました。 +- dtypeのための新しいインフラとキャストの準備 +- Mac 版の Python 3.8 と Python 3.9 用 universal2 wheel +- ドキュメントの改善 +- アノテーションの改善 +- 乱数生成用の新しい `PCG64DXSM` ビット生成機 + +今回のNumpy リリースは、175人による581件のプルリクエストのマージの結果です。 このリリースでサポートされている Python のバージョンは 3.7-3.9 です。Python 3.10 がリリースされた後、Python 3.10 のサポートが追加されます。 + + +### 2020年度 NumPy アンケート結果 + +_2021年6月22日_ -- NumPyの調査チームは、2020年に ミシガン大学とメリーランド大学の学生や教員と協力して、最初の公式NumPyコミュニティ調査を実施しました。 アンケートの結果はこちらから確認できます。 https://numpy.org/user-survey-2020/ + + +### NumPy 1.18.0 リリース + +_2021年1月30日_ -- [NumPy 1.20.0](https://numpy.org/doc/stable/release/1.20.0-notes.html) がリリースされました。 今回のリリースは180 人以上のコントリビューターのおかげで、これまでで最大の NumPyのリリースとなりました。 最も重要な2つの新機能は次のとおりです。 +- NumPyの大部分のコードに型注釈が追加されました。 そして新しいサブモジュールである`numpy.typing`が追加されました。 このサブモジュールは`ArrayLike` や`DtypeLike`という型注釈のエイリアスが定義されており、これによりユーザーやダウンストリームのライブラリはこの型注釈を使うことができます。 - X86(SSE、AVX)、ARM64(Neon)、およびPowerPC (VSX) 命令をサポートするマルチプラットフォームSIMDコンパイラの最適化が実施されました。 これにより、多くの関数で大きく パフォーマンスが向上しました (例: [sin/cos](https://github.com/numpy/numpy/pull/17587), [einsum](https://github.com/numpy/numpy/pull/18194)). ### NumPyプロジェクトの多様性 @@ -16,73 +177,102 @@ _2020年9月20日に_ 、私たちは[ NumPyプロジェクトにおけるダイ ### Natureに初の公式NumPy論文が掲載されました! -_2020年9月16日_ -- [NumPyに関する初の公式論文] (https://www.nature.com/articles/s41586-020-2649-2) が査読付き論文として掲載されました。 これはNumPy 1.0のリリースから14年後のことになります。 この論文では、配列プログラミングのアプリケーションと基本的なコンセプト、NumPyの上に構築された様々な科学的Pythonエコシステム、そしてCuPy、Dask、JAXのような外部の配列およびテンソルライブラリとの相互運用を容易にするために最近追加された配列プロトコルについて説明しています。 +_2020年9月16日_ -- NumPyに関する [ 最初の公式の論文 ](https://www.nature.com/articles/s41586-020-2649-2)がNatureに査読付き論文として掲載されました。 これはNumPy 1.0のリリースから14年後のことになりました。 この論文では、配列プログラミングのアプリケーションと基本的なコンセプト、NumPyの上に構築された様々な科学的Pythonエコシステム、そしてCuPy、Dask、JAXのような外部の配列およびテンソルライブラリとの相互運用を容易にするために最近追加された配列プロトコルについて説明しています。 ### Python 3.9のリリースに伴い、いつNumPyのバイナリwheelがリリースされるのですか？ -_2020年9月14日_ -- Python 3.9 は数週間後にリリースされる予定です。 もしあなたが新しいPythonのバージョンをいち早く取り入れているのであれば、NumPy（およびSciPyのような他のパッケージ）がリリース当日にバイナリwheelを用意していないことを知ってがっかりしたかもしれません。 ビルドインフラを新しいPythonのバージョンに適応させるのは大変な作業で、PyPIやconda-forgeにパッケージが掲載されるまでには通常数週間かかります。wheelのリリースに備えて、以下を確認してください。 +_2020年9月14日_ -- Python 3.9 は数週間後にリリースされる予定です。 もしあなたが新しいPythonのバージョンをいち早く利用している場合、NumPy（およびSciPyのような他のパッケージ）がリリース当日にバイナリwheelを用意していないことを知ってがっかりしたかもしれませんね。 ビルド用のインフラを新しいPythonのバージョンに適応させるのは非常に大変な作業で、PyPIやconda-forgeにパッケージが掲載されるまでには通常数週間かかります。 今後のwheelのリリースに備えて、以下を確認してください。 - `pip` が`manylinux2010` と `manylinux2014` をサポートするためにpipを少なくともバージョン 20.1 に更新する。 - [`--only-binary=numpy`](https://pip.pypa.io/en/stable/reference/pip_install/#cmdoption-only-binary) または `--only-binary=:all:` を`pip`がソースからビルドしようとするのを防ぐために使用します。 ### NumPy 1.19.2 リリース -_2020年1月10日_ -- [NumPy 19.2.0](https://numpy.org/devdocs/release/1.19.2-notes.html) がリリースされました。 この 1.19 シリーズの最新リリースでは、いくつかのバグが修正され、[来るべき Cython 3.xリリース](http:/docs.cython.orgenlatestsrcchanges.html)への準備が行われ、アップストリームの修正が進行中の間も distutils の動作を維持するためのsetuptoolsの固定がされています。 aarch64 wheelは最新のmanylinux2014リリースで構築されており、異なるLinuxディストリビューションで使用される異なるページサイズの問題を修正しています。 +_2020年9月10日_ -- [NumPy 19.2.0](https://numpy.org/devdocs/release/1.19.2-notes.html) がリリースされました。 この 1.19 シリーズの最新リリースでは、いくつかのバグが修正され、[ 来るべき Cython 3.xリリース ](http:/docs.cython.orgenlatestsrcchanges.html)への準備が行われ、アップストリームの修正が進行中の間も distutils の動作を維持するためのsetuptoolsのバージョンの固定が実施されています。 aarch64 wheelは最新のmanylinux2014リリースでビルドされており、異なるLinuxディストリビューションで使用される異なるページサイズの問題が修正されています。 ### 初めてのNumPyの調査が公開されました!! -_2020年7月2日_ -- このサーベイは、ソフトウェアとして、またコミュニティとしてのNumPyの開発に関する意思決定の指針となり、優先順位を設定するためのものになりました。 この調査結果は英語以外の8つの言語で利用可能です: バングラ, ヒンディー語, 日本語, マンダリン, ポルトガル語, ロシア語, スペイン語とフランス語. +_2020年7月2日_ -- このアンケート調査は、NumPyにおける、ソフトウェアとしてとコミュニティの両方における意思決定の指針となり、優先順位を決定する役に立ちました。 この調査結果は英語以外のこれらの8つの言語で利用可能です: バングラ, ヒンディー語, 日本語, マンダリン, ポルトガル語, ロシア語, スペイン語とフランス語. -NumPy をより良くするために、こちらの [アンケート](https://umdsurvey. umd. edu/jfe/form/SV_8bJrXjbhXf7saAl) に協力してもらえると嬉しいです。 +NumPy をより良くするために、こちらの [アンケート](https://umdsurvey. umd. edu/jfe/form/SV_8bJrXjbhXf7saAl) に協力してもらえると助かります。 ### NumPy に新しいロゴができました! -_2020年6月24日_ -- NumPy に新しいロゴが作成されました: +_2020年6月24日_ -- NumPyのロゴが新しくなりました: -NumPyのロゴ +NumPyのロゴ -新しいロゴは、古いもの比べてモダンで、よりクリーンなデザインになりました。 新しいロゴをデザインしてくれたIsabela Presedo-Floydと、15年以上にわたって使用してきた旧ロゴをデザインしてくれたTravis Vaughtに感謝します。 +新しいロゴは、古いロゴに比べて、モダンでよりクリーンなデザインになりました。 新しいロゴをデザインしてくれたIsabela Presedo-Floydと、15年以上にわたって使用してきた旧ロゴをデザインしてくれたTravis Vaughtに感謝します。 -### NumPy 1.19.0 リリース +### NumPy 1.20.0 リリース -_2020年6月20日_ -- NumPy 1.19.0 が利用可能になりました。 これのリリースは Python 2系のサポートがない最初のリリースであり、"クリーンアップ用のリリース" です。 サポートされている一番古いPython のバージョンは Python 3.6 になりました。 今回の重要な新機能は、NumPy 1.17.0で導入された乱数生成用のインフラにCythonからアクセスできるようになったことです。 +_2020年6月20日_ -- NumPy 1.19.0 がリリースされました。 このバージョンは Python 2系のサポートがない最初のリリースであり、"クリーンアップ用のリリース" です。 サポートされている一番古いPython のバージョンは Python 3.6 になりました。 また、今回の重要な新機能はNumPy 1.17.0で導入された乱数生成用のインフラにCythonからアクセスできるようになったことです。 ### ドキュメント受諾期間 -_2020年5月11日_ -- NumPyは、 Googleのシーズンオブドキュメントプログラムのメンター団体の1つとして選ばれました。 NumPy のドキュメントを改善するために、テクニカルライターと協力する機会を楽しみにしています! 詳細については、 [公式ドキュメントサイト](https://developers.google.com/season-of-docs/) と [アイデアページ](https://github.com/numpy/numpy/wiki/Google-Season-of-Docs-2020-Project-Ideas) をご覧ください。 +_2020年5月11日_ -- NumPyは、 Googleのシーズンオブドキュメントプログラムのメンター団体の1つとして選ばれました。 NumPy のドキュメントを改善するために、テクニカルライターと協力するこの機会を楽しみにしています! 詳細については、 [シーズンオブドキュメント公式サイト](https://developers.google.com/season-of-docs/) と [アイデアページ](https://github.com/numpy/numpy/wiki/Google-Season-of-Docs-2020-Project-Ideas) をご覧ください。 -### NumPy 1.18.0 リリース +### Numpy 1.18.0 リリース -_2019年12月22日_ -- NumPy 1.18.0 が利用可能になりました。 このリリースは、1.17.0の主要な変更の後の、統合的なリリースです。 Python 3.5 をサポートする最後のマイナーリリースになります。 今回のリリースでは、64ビットのBLASおよびLAPACKライブラリとリンクするためのインフラの追加や、`numpy.random`のための新しいC-APIの追加などが行われました。 +_2019年12月22日_ -- NumPy 1.18.0 がリリースされました。 このリリースは、1.17.0での主要な変更の後の、まとめのようなリリースです。 Python 3.5 をサポートする最後のマイナーリリースになります。 今回のリリースでは、64ビットのBLASおよびLAPACKライブラリとリンクするためのインフラの追加や、`numpy.random`のための新しいC-APIの追加などが行われました。 -詳細については、 [リリース ノート](https://github.com/numpy/numpy/releases/tag/v1.18.0) を参照してください。 +詳細については、 [リリースノート](https://github.com/numpy/numpy/releases/tag/v1.18.0) を参照してください。 ### NumPyはChan Zuckerberg財団から助成金を受けました。 -_2019年11月15日_ -- NumPyと、NumPyの重要な依存関係の1つであるOpenBLASが、Chan Zuckerberg財団の[Essential Open Source Software for Scienceプログラム](https:/chanzuckerberg.comeoss)を通じて、科学に不可欠なオープンソースツールのソフトウェアのメンテナンス、成長、開発、コミュニティへの参加を支援する195,000ドルの共同助成金を獲得したことを発表しました。 +_2019年11月15日_ -- NumPyと、NumPyの重要な依存ライブラリの1つであるOpenBLASが、Chan Zuckerberg財団の[Essential Open Source Software for Scienceプログラム](https:/chanzuckerberg.comeoss)を通じて、科学に不可欠なオープンソースツールのソフトウェアのメンテナンス、成長、開発、コミュニティへの参加などを支援する195,000ドルの共同助成金を獲得したことを発表しました。 -この助成金は、NumPy ドキュメント、ウェブサイトの再設計の改善に向けた取り組みを促進するために使用されます。 大規模かつ急速に拡大するユーザー基盤をより良くし、プロジェクトの長期的な持続可能性を確保するためのコミュニティ開発を行っていきます。 OpenBLASチームは、技術的に重要な問題、特にスレッド安全性、AVX-512に対処することに焦点を当てます。 また、スレッドローカルストレージ(TLS) の問題や、OpenBLASが依存するReLAPACK(再帰的なLAPACK) のアルゴリズムの改善も行っています。 +この助成金は、Numpy ドキュメントやウェブサイトの再設計などの改善に向けた取り組みを促進するために使用されます。 大規模かつ急速に拡大するユーザーの体験をより良くし、プロジェクトの長期的な持続可能性を確保するためのコミュニティ開発を行っていきます。 OpenBLASチームは、技術的に非常に重要な問題である、スレッド安全性、AVX-512に対処することに注力します。 また、スレッドローカルストレージ(TLS) の問題や、OpenBLASが依存するReLAPACK(再帰的なLAPACK) のアルゴリズムの改善も実施します。 -提案されたイニシアチブと成果物の詳細については、 [フルグラントプロポーザル](https://figshare.com/articles/Proposal_NumPy_OpenBLAS_for_Chan_Zuckerberg_Initiative_EOSS_2019_round_1/10302167) を参照してください。 この取り組みは2019年12月1日から始まり、今後12ヶ月間継続される予定です。 +提案されたイニシアチブとその成果の詳細については、 [フルグラントプロポーザル](https://figshare.com/articles/Proposal_NumPy_OpenBLAS_for_Chan_Zuckerberg_Initiative_EOSS_2019_round_1/10302167) を参照してください。 この取り組みは2019年12月1日から始まり、今後12ヶ月間継続実施される予定です。 -## 過去のリリース + -こちらがより過去のNumPy リリースのリストで、各リリースノートへのリンクが記載されています。 全てのバグフィックスリリース(バージョン番号`x.y.z` の`z`だけが変更されたもの)は新しい機能追加はされず、マイナーリリース (`y` が増えたもの)は、新しい機能追加されています。 +## 過去のリリース -- NumPy 1.18.4 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.18.4)) -- _2020年5月3日_. +こちらは、より以前のNumPyリリースのリストで、各リリースノートへのリンクが記載されています。 全てのバグフィックスリリース(バージョン番号`x.y.z` の`z`だけが変更されたもの)は新しい機能追加はされず、マイナーリリース (`y` が増えたもの)は、新しい機能追加されています。 + +- NumPy 1.26.3 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.26.2)) -- _ 2024年1月2日_. +- NumPy 1.26.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.3)) -- _2 Jan 2024_. +- NumPy 1.26.2 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.26.2)) -- _2023年11月12日_. +- NumPy 1.26.1 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.26.1)) -- _2023年10月14日_. +- NumPy 1.26.0 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.26.0)) -- _2023年9月16日_. +- NumPy 1.25.2 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.25.2)) -- _2023年7月31日_. +- NumPy 1.25.1 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.25.1)) -- _2023年7月8日_. +- NumPy 1.24.4 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.24.4)) -- _2023年6月26日_. +- NumPy 1.25.0 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.25.0)) -- _2023年6月17日_. +- NumPy 1.24.3 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.24.3)) -- _2023年4月22日_. +- NumPy 1.24.2 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.24.2)) -- _2023年2月5日_. +- NumPy 1.24.1 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.24.1)) -- _2022年12月26日_. - NumPy 1.18.4 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.18.4)) -- _2020年4月19日_. -- NumPy 1.18.2 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.18.2)) -- _2020年3月17日_. -- NumPy 1.18.1 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.18.1)) -- _2020年1月6日_. +- NumPy 1.23.5 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.23.5)) -- _2022年11月19日_. +- NumPy 1.23.4 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.23.4)) -- _2022年10月12日_. +- NumPy 1.23.3 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.23.3)) -- _2022年9月9日_. +- NumPy 1.23.2 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.23.2)) -- _2022年8月14日_. +- NumPy 1.23.1 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.23.1)) -- _2022年7月8日_. +- NumPy 1.23.0 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.23.0)) -- _2022年6月22日_. +- NumPy 1.22.4 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.22.4)) -- _2022年5月20日_. +- NumPy 1.21.6 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.21.6)) -- _2022年4月12日_. +- NumPy 1.22.3 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.18.2)) -- _2022年3月7日_. +- NumPy 1.22.2 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.22.2)) -- _2022年2月3日_. +- NumPy 1.22.1 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.22.1)) -- _2022年1月14日_. +- NumPy 1.22.0 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.22.0)) -- _2021年12月31日_. +- NumPy 1.21.5 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.21.5)) -- _2021年12月19日_. +- NumPy 1.21.0 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.21.0)) -- _2021年6月22日_. +- NumPy 1.20.3 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.20.3)) -- _2021年5月10日_. +- NumPy 1.20.0 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.20.0)) -- _2021年1月30日_. +- NumPy 1.19.5 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.19.5)) -- _2021年1月5日_. +- NumPy 1.19.0 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.19.0)) -- _2020年6月20日_. +- NumPy 1.18.4 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.18.4)) -- _2020年5月3日_. - NumPy 1.17.5 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.17.5)) -- _2020年1月1日_. - NumPy 1.18.0 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.18.0)) -- _2019年12月22日_. -- NumPy 1.17.4 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.17.4)) -- _2019年10月11日_. - NumPy 1.17.0 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.17.0)) -- _2019年7月26日_. - NumPy 1.16.0 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.16.0)) -- _2019年1月14日_. - NumPy 1.15.0 ([リリースノート](https://github.com/numpy/numpy/releases/tag/v1.15.0)) -- _2018年7月23日_. diff --git a/content/ja/press-kit.md b/content/ja/press-kit.md index a0fb78219d..6d28214989 100644 --- a/content/ja/press-kit.md +++ b/content/ja/press-kit.md @@ -5,4 +5,4 @@ sidebar: false 私たちはユーザーの皆さんが次に書く学術論文や、コース教材、プレゼンテーションなどに、NumPyプロジェクトのロゴを簡単に盛り込めるようにしたいと考えています。 -こちらから、様々な解像度のNumPyロゴのファイルをダウンロードできます: [ロゴリンク](https://github.com/numpy/numpy/tree/main/branding/logo)。numpy.orgのリソースを使用することで、[NumPy行動規範](/code-of-conduct) を受け入れたことになることに注意してください。 +こちらから、様々な解像度のNumPyロゴのファイルをダウンロードできます: [ロゴリンク](https://github.com/numpy/numpy/tree/main/branding/logo)。 numpy.orgのリソースを使用することで、[NumPy行動規範](/code-of-conduct) を受け入れたことになることに注意してください。 diff --git a/content/ja/report-handling-manual.md b/content/ja/report-handling-manual.md index f048688b4b..b200124145 100644 --- a/content/ja/report-handling-manual.md +++ b/content/ja/report-handling-manual.md @@ -5,20 +5,20 @@ sidebar: false NumPyの行動規範委員会はこのマニュアルに従います。 このマニュアルは様々な問題に対応する際に使用され、一貫性と公平性を確保します。 -[行動規範](/ja/code-of-conduct) を施行することは、私たちのコミュニティの現在のため、未来のために重要です。この施行は、軽いものではありません。施行の基準を見直す際、行動規範委員会は以下の考え方とガイドラインに留意するようにします。 +[行動規範](/ja/code-of-conduct) を施行することは、私たちのコミュニティの現在のため、未来のために重要です。 この施行は、軽いものではありません。 施行の基準を見直す際、行動規範委員会は以下の考え方とガイドラインに留意するようにします。 -* 機械的ではなく、人間的に行動します。 委員会は、当事者のプライバシーと報告者の必要なだけの機密性を尊重しながら、状況を理解するように働きかけることができます. ただし、1人以上の個人と直接連絡を取る必要がある場合もあります。委員会の目標は正しい決定を下すのではなく、コミュニティの健康を改善することなのです。 -* 行動を判断するのではなく、個人への共感を強調し、「良い」と「悪い」の二値評価を避けます。 明確な攻撃性とハラスメントが存在した場合、私たちはそれらに対処します。 しかし、解決が困難なシナリオの多くは、通常の意見の相違が、複数の当事者による無益または有害な行動に発展した場合です。完全に文脈を理解し、すべてを再び元に戻す道を見つけることは困難ですが、コミュニティにとって最終的に最も有益な方法です。 -* 私たちは、電子メールが判断に困難な媒体であり、独立して利用できることを理解しています。 個人の情報なしに電子メール上で批判を受けることは、特に苦痛である場合もあります。 そこで、他者の見解に対して、開放的で、敬意を持った雰囲気を保つことが重要になります。 それはまた、私たちの行動が透明でなければならないことを意味します。全てのメンバーが公平かつ同情をもって扱われるようにするため、私たちは全力を尽くします。 -* 差別の境界は時に曖昧で、また無意識に行われている場合もあります。これは、いたって普通のコミュニケーションの中で、不公平感や敵意として現れてきます。こうしたことが起きうることはわかっているので、注意深く見ていきます。不当な扱いを受けたと思われる方は、ぜひご連絡ください。 +* 機械的ではなく、人間的に行動します。 委員会は、当事者のプライバシーと報告者の必要なだけの機密性を尊重しながら、状況を理解するように働きかけることができます. ただし、1人以上の個人と直接連絡を取る必要がある場合もあります。 委員会の目標は正しい決定を下すのではなく、コミュニティの健康を改善することなのです。 +* 行動を判断するのではなく、個人への共感を強調し、「良い」と「悪い」の二値評価を避けます。 明確な攻撃性とハラスメントが存在した場合、私たちはそれらに対処します。 しかし、解決が困難なシナリオの多くは、通常の意見の相違が、複数の当事者による無益または有害な行動に発展した場合です。 完全に文脈を理解し、すべてを再び元に戻す道を見つけることは困難ですが、コミュニティにとって最終的に最も有益な方法です。 +* 私たちは、電子メールが判断に困難な媒体であり、独立して利用できることを理解しています。 個人の情報なしに電子メール上で批判を受けることは、特に苦痛である場合もあります。 そこで、他者の見解に対して、開放的で、敬意を持った雰囲気を保つことが重要になります。 それはまた、私たちの行動が透明でなければならないことを意味します。 全てのメンバーが公平かつ同情をもって扱われるようにするため、私たちは全力を尽くします。 +* 差別の境界は時に曖昧で、また無意識に行われている場合もあります。 これにより、普通の人との関わりの中で、不公平感や敵意として現れてくるのです。 私達は、このようなことが起こることはわかっているので、気をつけて見ていきたいと思います。 不当な扱いを受けたと思われる方は、ぜひご連絡ください。 * 良い議論を実践することで、エンゲージメントの向上に取り組みます。例えば議論がどこで止まっているのかを特定したり、 実践的な情報、指針、資源を提供することで、これらの問題を前向きな方向に変えていきます。 -* 新しいメンバーが何を必要としているかに留意します。特に社会的地位の低いグループからの参加を増やすことを目的に、明確なサポートと配慮を提供していきます。 -* 一人一人の文化的背景や母国語は異なります。ネイティブでない人が起こした悪気のない誤解を確認し、問題を理解してもらい、不快感を与えないために何を変えればよいかを教えてあげてください。 外国語での複雑な議論はとても難しいものであり、国籍や文化を超えて多様性を育てていきたいと考えています。 +* 新しいメンバーが何を必要としているかに留意します。 特に社会的地位の低いグループからの参加を増やすことを目的に、明確なサポートと配慮を提供していきます。 +* 一人一人の文化的背景や母国語は異なります。 ネイティブでない人が起こした悪気のない誤解を確認し、問題を理解してもらい、不快感を与えないために何を変えればよいかを教えてあげてください。 外国語での複雑な議論はとても難しいものであり、国籍や文化を超えて多様性を育てていきたいと考えています。 ## 仲介 -自主的な非公式の調停は、私たちの重要な役割です。2つのグループ以上の当事者が不適切な行動をエスカレートした場合(人類の紛争では悲しいことに一般的なものですが)、調停プロセスを促進するのは非常に重要です。ちなみに、これは一例に過ぎません。委員会は、どのようなケースでも調停を検討することができますが、このプロセスはあくまでも自発的なものであり、当事者に参加を迫ることはできないことを念頭に置いて下さい。 委員会が調停を提案する場合は、次のようにすべきです。 +自主的な非公式の調停は、私たちの重要な役割です。 2つのグループ以上の当事者が不適切な行動をエスカレートした場合(人類の紛争では悲しいことに一般的なものですが)、調停プロセスを促進するは非常に重要です。 ちなみに、これは一例に過ぎません。委員会は、どのようなケースでも調停を検討することができますが、このプロセスはあくまでも自発的なものであり、当事者に参加を迫ることはできないことを念頭に置いて下さい。 委員会が調停を提案する場合は、次のようにすべきです。 * 調停者として役立つ候補者を見つけます。 * 報告者の合意を取得します。 報告者は、調停のアイデアを拒否したり、代替の調停者を提案する権利を持ちます。 @@ -36,13 +36,13 @@ NumPyの行動規範委員会はこのマニュアルに従います。 この ## 明確かつ深刻な違反行為の解決 -私たちは、インターネットでの会話が簡単にひどい誹謗中傷になってしまうことを、痛いほど知っています。個人的な脅迫、暴力的、性差別的、人種差別的な言葉など、明らかで深刻な違反に対しては、迅速に対処します。 +私たちは、インターネットでの会話が簡単にひどい誹謗中傷になってしまうことを、痛いほど知っています。 個人的な脅迫、暴力的、性差別的、人種差別的な言葉など、明らかで深刻な違反に対しては、迅速に対処します。 行動規範委員会のメンバーは、明確かつ深刻な違反に気づいた場合、以下のように行動します。 * 直ちにすべてのNumPyのオンラインコミュニティから違反者を排除します。 * 報告が受信され、違反者が排除されたことを報告者に連絡します。 -* どのような場合でも、モデレーターは違反者に連絡するための合理的な努力を行い、違反者の言葉や行動がどのように「明確かつ重大な違反」に該当するのかを具体的に伝えるべきです。モデレーターはまた、違反者の権利についても述べるべきです。違反者は、排除を不当だと思う場合やNumPyコミュニティへの復帰を望む場合、以下に述べる行動規範委員会による審査を求める権利があります。モデレータは、この説明を行動規範委員会に転送する必要があります。 +* どのような場合でも、モデレーターは違反者に連絡するための合理的な努力を行い、違反者の言葉や行動がどのように「明確かつ重大な違反」に該当するのかを具体的に伝えるべきです。 モデレーターは、違反者がこれは不当だと思う場合、あるいはNumPyチャンネルとの再接続を望む場合には、行動規範委員会による以下のような審査を求める権利があることも述べるべきです。 モデレータは、この説明を行動規範委員会に転送する必要があります。 * 行動規範委員会は、このプロセスが適用されたすべてのケースを正式にレビューし署名することで、よくある盛り上がりすぎた論争を諫めるためこのプロセスが使用されたのでないことを確認します。 @@ -50,7 +50,7 @@ NumPyの行動規範委員会はこのマニュアルに従います。 この 報告が委員会に送られると、直ちに報告者に返信して報告を受領したことを確認します。 この返信は72時間以内に送信される必要があり、委員会はそれよりもはるかに迅速に対応するよう努める必要があります。 -レポートに十分な情報が含まれていない場合、委員会は行動する前に、関連するすべてのデータを取得するようにします。委員会は、事件の状況を全て知るために関係する個人に連絡する際に、運営協議会に代わって行動する権限を与えられています。 +レポートに十分な情報が含まれていない場合、委員会は行動する前に、関連するすべてのデータを取得するようにします。 委員会は、事件の状況を全て知るために関係する個人に連絡する際に、運営協議会に代わって行動する権限を与えられています。 その後、委員会は今回の問題を見直し、効果を最大限に発揮する対策を決定します。 @@ -61,9 +61,9 @@ NumPyの行動規範委員会はこのマニュアルに従います。 この これらの情報は書面で収集され、可能な限りグループの審議が記録され、保持されます (例えば、チャットの記録、Eメールのディスカッション、会議通話の記録、音声会話の概要など)。 -行動の一貫性を確保し、プロジェクトのために記録を残すために、委員会のすべての活動のアーカイブを保持することが重要です。この活動を支援するため、委員会はデフォルトでプライベートメーリングリストを議論に使用します。このメーリングリストには、要求が正当なものなら、委員会の現在および将来のメンバー、および運営委員会のメンバーがアクセスできるにします。委員会がリストにはない連絡方法を使用する必要がある場合(例: 早期/迅速な対応を求める電話など)、そのプロセスの良い記録となるように、これらをリストにまとめて戻すべきです。 +行動の一貫性を確保し、プロジェクトのために記録を残すために、委員会のすべての活動のアーカイブを保持することが重要です。 この活動支援するために、委員会のデフォルトの議論チャネルは、正当化された要求に応じて、委員会の現在および将来のメンバー、および運営委員会のメンバーがアクセスできるプライベートメーリングリストにします。 委員会がリストにはない連絡方法を使用する必要がある場合(例: 早期/迅速な対応を求める電話など)、そのプロセスの良い記録となるように、これらをリストにまとめて戻すべきです。 -行動規範委員会は、2週間以内に決議の合意を目指すべきです。その期間内に決議が確定できない場合。委員会は、レポーターに対して現状の更新と今後のタイムラインを連絡します。 +行動規範委員会は、2週間以内に決議の合意を目指すべきです。 その期間内に決議が確定できない場合。 委員会は、レポーターに対して現状の更新と今後のタイムラインを連絡します。 ## 解決方法 @@ -76,18 +76,18 @@ NumPyの行動規範委員会はこのマニュアルに従います。 この - 違反が起きていないと判断された - 検討中に問題が明らかに解決された * 調停の調整。すべての関係者が合意した場合、委員会は上記のように調停プロセスを促進することができます。 -* 公の場における説明。どの行動・言動・言語が不適切で、現在の状況がなぜか引き起こされ、人々を傷つけたのかを説明し、コミュニティに自省を要求します。 +* 公の場における説明。 どの行動・言動・言語が不適切で、現在の状況がなぜか引き起こされ、人々を傷つけたのかを説明し、コミュニティに自省を要求します。 * 委員会から関係者(複数可) への非公開処分の実施。 この場合、委員会は、電子メールを介して、グループにccを入れながら、対象者に問題の指摘を連絡します。 -* 公の場における処分の実施。この場合、委員会の議長は、違反が発生したのと同じ場所で、可能な範囲内で叱責を行います。例えば、メール規約違反の発生したメーリングリストなどです。しかし、人や状況がかわるかもしれないチャットルームなどの場合、他の手段を利用する可能性もあります。文書化のため、この問題のメッセージを他の場所で公開することを対策グループが選択する場合もあります。 -* 報告者がこの考えに同意することを前提とした、公的または私的な謝罪の要求。報告者は自分の裁量で、違反者とのさらなる接触を拒否することもできます。委員会がこの要求を通達します。委員会は、必要に応じてこの要求に「条件」を付けることができます。例えば、メーリングリストの会員資格を維持するために、違反者に謝罪を求めることができます。 -* 「相互に合意した休止」の要求。これは、委員会から個人への、コミュニティへの参加を一時的に控えるような要請です。 対象者が自発的に一時的な休みを取らないことを選択した場合、委員会は「冷却期限」を準備することがあります。 -* これは、一部またはすべてのNumPyオンラインコミュニティ (メーリングリスト、gitter.im など) からの永続的または一時的な出入り禁止。将来的に禁止が見直されるのか、維持されるか決定できるよう、対策グループは出入り禁止の記録を全て保持します。 +* 公の場での指摘。 この場合、委員会の議長は、違反が発生したのと同じ場所で、実用性の範囲内で叱責を行います。 例えば、メールルールの違反の元のメーリングリストなどです。しかし、人や状況がかわるかもしれないチャットルームなどの場合、他の手段を利用する可能性もあります。 文書化のため、この問題のメッセージを他の場所で公開することを対策グループが選択する場合もあります。 +* 報告者がこの考えに同意することを前提とした、公的または私的な謝罪の要求。 報告者は自分の裁量で、違反者とのさらなる接触を拒否することもできます。 委員会がこの要求をお届けします。 委員会は、必要に応じてこの要求に「条件」を付けることができます。例えば、メーリングリストの会員資格を維持するために、違反者に謝罪を求めることができます。 +* 「相互に合意した休止」の要求。 これは、委員会から個人への、コミュニティへの参加を一時的に控えるような要請です。 対象者が自発的に一時的な休みを取らないことを選択した場合、委員会は「冷却期限」を準備することがあります。 +* これは、一部またはすべてのNumPyオンラインコミュニティ (メーリングリスト、gitter.im など) からの永続的または一時的な出入り禁止。 将来的に禁止が見直されるのか、維持されるか決定できるよう、対策グループは出入り禁止の記録を全て保持します。 -決議が合意されると制定される前に、委員会は、元の報告者およびその他の影響を受けた当事者に連絡し、提案された決議を説明します。 委員会は、この決議が受け入れられるかどうかを尋ねます。そして、記録のためのフィードバックに注意を払います。 +決議が合意されると制定される前に、委員会は、元の報告者およびその他の影響を受けた当事者に連絡し、提案された決議を説明します。 委員会は、この決議が受け入れられるかどうかを尋ねます。 そして、記録のためのフィードバックに注意を払います。 最後に 委員会は、NumPy Steering Councilに報告を行います(NumPy Coreチームにも、出入り禁止など進行中の出来事については報告します)。 -委員会はこの問題について公に議論することはありません。すべての公開声明は、行動規範委員会またはNumPy Steering Councilの議長によって行われます。 +委員会はこの問題について公に議論することはありません。 すべての公開声明は、行動規範委員会またはNumPy Steering Councilの議長によって行われます。 ## 利益相反 diff --git a/content/ja/tabcontents.yaml b/content/ja/tabcontents.yaml index 9f4deb6121..5a208d9f6b 100644 --- a/content/ja/tabcontents.yaml +++ b/content/ja/tabcontents.yaml @@ -1,189 +1,373 @@ -machinelearning: - paras: - - para1: NumPy forms the basis of powerful machine learning libraries like [scikit-learn](https://scikit-learn.org) and [SciPy](https://www.scipy.org). As machine learning grows, so does the list of libraries built on NumPy. [TensorFlow’s](https://www.tensorflow.org) deep learning capabilities have broad applications — among them speech and image recognition, text-based applications, time-series analysis, and video detection. [PyTorch](https://pytorch.org), another deep learning library, is popular among researchers in computer vision and natural language processing. [MXNet](https://github.com/apache/incubator-mxnet) is another AI package, providing blueprints and templates for deep learning. - para2: Statistical techniques called [ensemble](https://towardsdatascience.com/ensemble-methods-bagging-boosting-and-stacking-c9214a10a205) methods such as binning, bagging, stacking, and boosting are among the ML algorithms implemented by tools such as [XGBoost](https://github.com/dmlc/xgboost), [LightGBM](https://lightgbm.readthedocs.io/en/latest/), and [CatBoost](https://catboost.ai) — one of the fastest inference engines. [Yellowbrick](https://www.scikit-yb.org/en/latest/) and [Eli5](https://eli5.readthedocs.io/en/latest/) offer machine learning visualizations. - -arraylibraries: - intro: - - text: NumPy's API is the starting point when libraries are written to exploit innovative hardware, create specialized array types, or add capabilities beyond what NumPy provides. - - headers: - - text: Array Library - - text: Capabilities & Application areas - - libraries: - - title: Dask - text: Distributed arrays and advanced parallelism for analytics, enabling performance at scale. - img: /images/content_images/arlib/dask.png - alttext: Dask - url: https://dask.org/ - - title: CuPy - text: NumPy-compatible array library for GPU-accelerated computing with Python. - img: /images/content_images/arlib/cupy.png - alttext: CuPy - url: https://cupy.chainer.org - - title: JAX - text: "Composable transformations of NumPy programs differentiate: vectorize, just-in-time compilation to GPU/TPU." - img: /images/content_images/arlib/jax_logo_250px.png - alttext: JAX - url: https://github.com/google/jax - - title: Xarray - text: Labeled, indexed multi-dimensional arrays for advanced analytics and visualization - img: /images/content_images/arlib/xarray.png - alttext: xarray - url: https://xarray.pydata.org/en/stable/index.html - - title: Sparse - text: NumPy-compatible sparse array library that integrates with Dask and SciPy's sparse linear algebra. - img: /images/content_images/arlib/sparse.png - alttext: sparse - url: https://sparse.pydata.org/en/latest/ - - title: PyTorch - text: Deep learning framework that accelerates the path from research prototyping to production deployment. - img: /images/content_images/arlib/pytorch-logo-dark.svg - alttext: PyTorch - url: https://pytorch.org/ - - title: TensorFlow - text: An end-to-end platform for machine learning to easily build and deploy ML powered applications. - img: /images/content_images/arlib/tensorflow-logo.svg - alttext: TensorFlow - url: https://www.tensorflow.org - - title: MXNet - text: Deep learning framework suited for flexible research prototyping and production. - img: /images/content_images/arlib/mxnet_logo.png - alttext: MXNet - url: https://mxnet.apache.org/ - - title: Arrow - text: A cross-language development platform for columnar in-memory data and analytics. - img: /images/content_images/arlib/arrow.png - alttext: arrow - url: https://github.com/apache/arrow - - title: xtensor - text: Multi-dimensional arrays with broadcasting and lazy computing for numerical analysis. - img: /images/content_images/arlib/xtensor.png - alttext: xtensor - url: https://github.com/xtensor-stack/xtensor-python - - title: XND - text: Develop libraries for array computing, recreating NumPy's foundational concepts. - img: /images/content_images/arlib/xnd.png - alttext: xnd - url: https://xnd.io - - title: uarray - text: Python backend system that decouples API from implementation; unumpy provides a NumPy API. - img: /images/content_images/arlib/uarray.png - alttext: uarray - url: https://uarray.org/en/latest/ - - title: tensorly - text: Tensor learning, algebra and backends to seamlessly use NumPy, MXNet, PyTorch, TensorFlow or CuPy. - img: /images/content_images/arlib/tensorly.png - alttext: tensorly - url: http://tensorly.org/stable/home.html - -scientificdomains: - intro: - - text: Nearly every scientist working in Python draws on the power of NumPy. - - text: "NumPy brings the computational power of languages like C and Fortran to Python, a language much easier to learn and use. With this power comes simplicity: a solution in NumPy is often clear and elegant." - - librariesrow1: - - title: Quantum Computing - alttext: A computer chip. - img: /images/content_images/sc_dom_img/quantum_computing.svg - - title: Statistical Computing - alttext: A line graph with the line moving up. - img: /images/content_images/sc_dom_img/statistical_computing.svg - - title: Signal Processing - alttext: A bar chart with positive and negative values. - img: /images/content_images/sc_dom_img/signal_processing.svg - - title: Image Processing - alttext: An photograph of the mountains. - img: /images/content_images/sc_dom_img/image_processing.svg - - title: Graphs and Networks - alttext: A simple graph. - img: /images/content_images/sc_dom_img/sd6.svg - - title: Astronomy Processes - alttext: A telescope. - img: /images/content_images/sc_dom_img/astronomy_processes.svg - - title: Cognitive Psychology - alttext: A human head with gears. - img: /images/content_images/sc_dom_img/cognitive_psychology.svg - - librariesrow2: - - title: Bioinformatics - alttext: A strand of DNA. - img: /images/content_images/sc_dom_img/bioinformatics.svg - - title: Bayesian Inference - alttext: A graph with a bell-shaped curve. - img: /images/content_images/sc_dom_img/bayesian_inference.svg - - title: Mathematical Analysis - alttext: Four mathematical symbols. - img: /images/content_images/sc_dom_img/mathematical_analysis.svg - - title: Chemistry - alttext: A test tube. - img: /images/content_images/sc_dom_img/chemistry.svg - - title: Geoscience - alttext: The Earth. - img: /images/content_images/sc_dom_img/geoscience.svg - - title: Geographic Processing - alttext: A map. - img: /images/content_images/sc_dom_img/GIS.svg - - title: Architecture & Engineering - alttext: A microprocessor development board. - img: /images/content_images/sc_dom_img/robotics.svg - -datascience: - - intro: "NumPy lies at the core of a rich ecosystem of data science libraries. A typical exploratory data science workflow might look like:" - - image1: - - img: /images/content_images/ds-landscape.png - alttext: Diagram of Python Libraries. The five catagories are 'Extract, Transform, Load', 'Data Exploration', 'Data Modeling', 'Data Evaluation' and 'Data Presentation'. - - image2: - - img: /images/content_images/data-science.png - alttext: Diagram of three overlapping circle. The circles labeled 'Mathematics', 'Computer Science' and 'Domain Expertise'. In the middle of the diagram, which has the three circles overlapping it, is an area labeled 'Data Science'. - - examples: - - text: "Extract, Transform, Load: [Pandas](https://pandas.pydata.org),[ Intake](https://intake.readthedocs.io),[PyJanitor](https://pyjanitor.readthedocs.io/)" - - text: "Exploratory analysis: [Jupyter](https://jupyter.org),[Seaborn](https://seaborn.pydata.org),[ Matplotlib](https://matplotlib.org),[ Altair](https://altair-viz.github.io)" - - text: "Model and evaluate: [scikit-learn](https://scikit-learn.org),[ statsmodels](https://www.statsmodels.org/stable/index.html),[ PyMC3](https://docs.pymc.io),[ spaCy](https://spacy.io)" - - text: "Report in a dashboard: [Dash](https://plotly.com/dash),[ Panel](https://panel.holoviz.org),[ Voila](https://github.com/voila-dashboards/voila)" - - content: - - text: For high data volumes, [Dask](https://dask.org) and[Ray](https://ray.io/) are designed to scale. Stabledeployments rely on data versioning ([DVC](https://dvc.org)),experiment tracking ([MLFlow](https://mlflow.org)), andworkflow automation ([Airflow](https://airflow.apache.org) and[Prefect](https://www.prefect.io)). - -visualization: - images: - - url: https://www.fusioncharts.com/blog/best-python-data-visualization-libraries - img: /images/content_images/v_matplotlib.png - alttext: A streamplot made in matplotlib - - url: https://github.com/yhat/ggpy - img: /images/content_images/v_ggpy.png - alttext: A scatter-plot graph made in ggpy - - url: https://www.journaldev.com/19692/python-plotly-tutorial - img: /images/content_images/v_plotly.png - alttext: A box-plot made in plotly - - url: https://altair-viz.github.io/gallery/streamgraph.html - img: /images/content_images/v_altair.png - alttext: A streamgraph made in altair - - url: https://seaborn.pydata.org - img: /images/content_images/v_seaborn.png - alttext: A pairplot of two types of graph, a plot-graph and a frequency graph made in seaborn" - - url: https://docs.pyvista.org/examples/index.html - img: /images/content_images/v_pyvista.png - alttext: A 3D volume rendering made in PyVista. - - url: https://napari.org - img: /images/content_images/v_napari.png - alttext: A multi-dimensionan image made in napari. - - url: https://vispy.org/gallery/index.html - img: /images/content_images/v_vispy.png - alttext: A Voronoi diagram made in vispy. - - content: - - text: NumPy is an essential component in the burgeoning - [Python visualization landscape](https://pyviz.org/overviews/index.html), - which includes [Matplotlib](https://matplotlib.org), - [Seaborn](https://seaborn.pydata.org), [Plotly](https://plot.ly), - [Altair](https://altair-viz.github.io), [Bokeh](https://docs.bokeh.org/en/latest/), - [Holoviz](https://holoviz.org), [Vispy](http://vispy.org), [Napari](https://github.com/napari/napari), - and [PyVista](https://github.com/pyvista/pyvista), to name a few. - - text: NumPy's accelerated processing of large arrays allows researchers to visualize - datasets far larger than native Python could handle. \ No newline at end of file +params: + machinelearning: + paras: + - + para1: NumPyは、[scikit-learn](https://scikit-learn.org)や[SciPy](https://www.scipy.org)のような強力な機械学習ライブラリの基礎を形成しています。機械学習の技術分野が成長するにつれ、NumPyをベースにしたライブラリの数も増えています。[TensorFlow](https://www.tensorflow.org)の深層学習機能は、音声認識や画像認識、テキストベースのアプリケーション、時系列分析、動画検出など、幅広い応用用途があります。[PyTorch](https://pytorch.org)も、コンピュータビジョンや自然言語処理の研究者に人気のある深層学習ライブラリです。[MXNet](https://github.com/apache/incubator-mxnet)もAIパッケージの一つで、深層学習の設計図やテンプレート機能を提供しています。 + para2: '[ensemble](https://towardsdatascience.com/ensemble-methods-bagging-boosting-and-stacking-c9214a10a205)法と呼ばれる統計的手法であるビンニング、バギング、スタッキングや、[XGBoost](https://github.com/dmlc/xgboost)、[LightGBM](https://lightgbm.readthedocs.io/en/latest/)、[CatBoost](https://catboost.ai)などのツールで実装されているブースティングなどは、機械学習アルゴリズムの一つであり、最速の推論エンジンの一つです。[Yellowbrick](https://www.scikit-yb.org/en/latest/)や[Eli5](https://eli5.readthedocs.io/en/latest/)は機械学習の可視化機能を提供しています。' + arraylibraries: + intro: + - + text: NumPyのAPIは、革新的なハードウェアを利用したり、特殊な配列タイプを作成したり、NumPyが提供する以上の機能を追加するためにライブラリを作成する際の基礎となります。 + headers: + - + text: 配列ライブラリ + - + text: 機能と応用分野 + libraries: + - + title: Dask + text: 分析用の分散配列と高度な並列処理により、大規模な処理を可能にします。 + img: /images/content_images/arlib/dask.png + alttext: Dask + url: https://dask.org/ + - + title: CuPy + text: Python を使用した GPUによる高速計算用のNumPy互換配列ライブラリ + img: /images/content_images/arlib/cupy.png + alttext: CuPy + url: https://cupy.chainer.org + - + title: JAX + text: "NumPyコードの合成可能な変換ライブラリ: 微分、ベクトル化、GPU/TPUへのジャストインタイムコンパイル" + img: /images/content_images/arlib/jax_logo_250px.png + alttext: JAX + url: https://github.com/google/jax + - + title: Xarray + text: 高度な分析と視覚化のためのラベルとインデックス付き多次元配列 + img: /images/content_images/arlib/xarray.png + alttext: xarray + url: https://xarray.pydata.org/en/stable/index.html + - + title: Sparse + text: Dask と SciPy の疎行列の線形代数ライブラリを統合した、Numpy 互換の疎行列ライブラリ + img: /images/content_images/arlib/sparse.png + alttext: sparse + url: https://sparse.pydata.org/en/latest/ + - + title: PyTorch + text: 研究用のプロトタイピングから本番運用への展開を加速させる、深層学習フレームワーク + img: /images/content_images/arlib/pytorch-logo-dark.svg + alttext: PyTorch + url: https://pytorch.org/ + - + title: TensorFlow + text: 機械学習を利用したアプリケーションを簡単に構築・展開するための、エンド・ツー・エンドの機械学習プラットフォーム + img: /images/content_images/arlib/tensorflow-logo.svg + alttext: TensorFlow + url: https://www.tensorflow.org + - + title: MXNet + text: 柔軟や研究用のプロトタイピングから、実際の運用まで利用可能な深層学習フレームワーク + img: /images/content_images/arlib/mxnet_logo.png + alttext: MXNet + url: https://mxnet.apache.org/ + - + title: Arrow + text: 列型のインメモリーデータやその分析のための、複数の言語に対応した開発プラットフォーム + img: /images/content_images/arlib/arrow.png + alttext: arrow + url: https://github.com/apache/arrow + - + title: xtensor + text: 数値解析のためのブロードキャスティングと遅延計算を備えた多次元配列 + img: /images/content_images/arlib/xtensor.png + alttext: xtensor + url: https://github.com/xtensor-stack/xtensor-python + - + title: XND + text: Numpy のような イディオムを使って JSON のようなデータを操作するライブラリ + img: /images/content_images/arlib/xnd.png + alttext: awkward + url: https://awkward-array.org/ + - + title: uarray + text: APIを実装から切り離すPythonバックエンドシステム (unumpyはNumPy APIを提供しています) + img: /images/content_images/arlib/uarray.png + alttext: uarray + url: https://uarray.org/en/latest/ + scientificdomains: + intro: + - + text: Pythonを使って働くほとんどの科学者はNumPyの力を利用しています。 + - + text: "Numpy は、 C や Fortran のような言語の計算パフォーマンスを、Pythonにもたらします。 このパワーはNumPyのシンプルさから来ており、NumPyによるソリューションの多くは明確でエレガントになります。" + libraries: + - + title: 量子コンピューティング + alttext: コンピューターチップ + img: /images/content_images/sc_dom_img/quantum_computing.svg + links: + - + url: http://qutip.org + label: QuTiP + - + url: https://pyquil-docs.rigetti.com/en/stable + label: PyQuil + - + url: https://qiskit.org + label: Qiskit + - + url: https://pennylane.ai + label: PennyLane + - + title: 統計コンピューティング + alttext: 線グラフが上に移動します。 + img: /images/content_images/sc_dom_img/statistical_computing.svg + links: + - + url: https://pandas.pydata.org/ + label: Pandas + - + url: https://github.com/statsmodels/statsmodels + label: statsmodels + - + url: https://xarray.pydata.org/en/stable/ + label: Xarray + - + url: https://github.com/mwaskom/seaborn + label: Seaborn + - + title: 信号処理 + alttext: 正と負の値を持つ棒グラフ。 + img: /images/content_images/sc_dom_img/signal_processing.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://pywavelets.readthedocs.io/ + label: PyWavelets + - + url: https://python-control.org/ + label: python-control + - + url: https://hyperspy.org/ + label: HyperSpy + - + title: 画像処理 + alttext: 山々の写真 + img: /images/content_images/sc_dom_img/image_processing.svg + links: + - + url: https://scikit-image.org/ + label: Scikit-image + - + url: https://opencv.org/ + label: OpenCV + - + url: https://mahotas.rtfd.io/ + label: Mahotas + - + title: グラフとネットワーク + alttext: シンプルなグラフ + img: /images/content_images/sc_dom_img/sd6.svg + links: + - + url: https://networkx.org/ + label: NetworkX + - + url: https://graph-tool.skewed.de/ + label: graph-tool + - + url: https://igraph.org/python/ + label: igraph + - + url: https://pygsp.rtfd.io/ + label: PyGSP + - + title: 天文学 + alttext: 望遠鏡 + img: /images/content_images/sc_dom_img/astronomy_processes.svg + links: + - + url: https://www.astropy.org/ + label: AstroPy + - + url: https://github.com/sunpy/sunpy + label: SunPy + - + url: https://github.com/spacepy/spacepy + label: SpacePy + - + title: 認知心理学 + alttext: ギアをつけた人間の頭部 + img: /images/content_images/sc_dom_img/cognitive_psychology.svg + links: + - + url: https://www.psychopy.org/ + label: PsychoPy + - + title: 生命情報科学 + alttext: DNAの鎖 + img: /images/content_images/sc_dom_img/bioinformatics.svg + links: + - + url: https://biopython.org/ + label: BioPython + - + url: http://scikit-bio.org/ + label: Scikit-Bio + - + url: https://github.com/openvax/pyensembl + label: PyEnsembl + - + url: http://etetoolkit.org/ + label: ETE + - + title: ベイズ推論 + alttext: 鐘形の曲線のグラフ + img: /images/content_images/sc_dom_img/bayesian_inference.svg + links: + - + url: https://pystan.readthedocs.io/en/latest/ + label: PyStan + - + url: https://docs.pymc.io/ + label: PyMC3 + - + url: https://arviz-devs.github.io/arviz/ + label: ArviZ + - + url: https://emcee.readthedocs.io/ + label: emcee + - + title: 数学的分析 + alttext: ４つの数学記号 + img: /images/content_images/sc_dom_img/mathematical_analysis.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://www.sympy.org/ + label: SymPy + - + url: https://github.com/cvxgrp/cvxpy + label: cvxpy + - + url: https://fenicsproject.org/ + label: FEniCS + - + title: 化学 + alttext: 試験管 + img: /images/content_images/sc_dom_img/chemistry.svg + links: + - + url: https://cantera.org/ + label: Cantera + - + url: https://www.mdanalysis.org/ + label: MDAnalysis + - + url: https://github.com/rdkit/rdkit + label: RDKit + - + url: https://www.pybamm.org/ + label: PyBaMM + - + title: 地球科学 + alttext: 地球 + img: /images/content_images/sc_dom_img/geoscience.svg + links: + - + url: https://pangeo.io/ + label: Pangeo + - + url: https://simpeg.xyz/ + label: Simpeg + - + url: https://github.com/obspy/obspy/wiki + label: ObsPy + - + url: https://www.fatiando.org/ + label: Fatiando a Terra + - + title: 地理情報処理 + alttext: 地図 + img: /images/content_images/sc_dom_img/GIS.svg + links: + - + url: https://shapely.readthedocs.io/ + label: Shapely + - + url: https://geopandas.org/ + label: GeoPandas + - + url: https://python-visualization.github.io/folium + label: Folium + - + title: アーキテクチャとエンジニアリング + alttext: マイクロプロセッサ開発ボード + img: /images/content_images/sc_dom_img/robotics.svg + links: + - + url: https://compas.dev/ + label: COMPAS + - + url: https://cityenergyanalyst.com/ + label: 都市エネルギー分析 + - + url: https://nortikin.github.io/sverchok/ + label: Sverchok + datascience: + intro: "Numpy は豊富なデータサイエンスライブラリのエコシステムの中核にあります。一般的なデータサイエンスのワークフローは次のようになります。" + image1: + - + img: /images/content_images/ds-landscape.png + alttext: Python ライブラリの図 。5 つのカテゴリに分類され、「抽出、変換、読み込み」、「データ探索」、「モデリング」、「評価」、「可視化」です。 + image2: + - + img: /images/content_images/data-science.png + alttext: 三つの円が重なり合う図。円はそれぞれ「数学」、「コンピューターサイエンス」、「専門知識」でラベル付けされています。図の中心部には、三つの円が重なり合って形成されるエリアがあり、「データサイエンス」とラベル付けされています。 + examples: + - + text: "抽出, 変換, 読み込み: [Pandas](https://pandas.pydata.org), [Intake](https://intake.readthedocs.io), [PyJanitor](https://pyjanitor.readthedocs.io/)" + - + text: "Exploratory analysis: [Jupyter](https://jupyter.org),[Seaborn](https://seaborn.pydata.org),[ Matplotlib](https://matplotlib.org),[ Altair](https://altair-viz.github.io)" + - + text: "モデリングと評価: [scikit-learn](https://scikit-learn.org), [statsmodels](https://www.statsmodels.org/stable/index.html), [PyMC3](https://docs.pymc.io), [spaCy](https://spacy.io)" + - + text: "ダッシュボードでのレポート: [Dash](https://plotly.com/dash),[ Panel](https://panel.holoviz.org),[ Voila](https://github.com/voila-dashboards/voila)" + content: + - + text: 大規模データに対して、[Dask](https://dask.org)と[Ray](https://ray.io/)はスケールすることを目指して設計されています。安定したデプロイメントに関しては、データのバージョニング（[DVC](https://dvc.org)）、実験の追跡（[MLFlow](https://mlflow.org)）、ワークフローの自動化（[Airflow](https://airflow.apache.org)および[Prefect](https://www.prefect.io)が重要ですが様々なNumPyベースのツールが提供されています。 + visualization: + images: + - + url: https://www.fusioncharts.com/blog/best-python-data-visualization-libraries + img: /images/content_images/v_matplotlib.png + alttext: matplotlibで作られたストリームプロット + - + url: https://github.com/yhat/ggpy + img: /images/content_images/v_ggpy.png + alttext: ggpyで作られた散布図グラフ + - + url: https://www.journaldev.com/19692/python-plotly-tutorial + img: /images/content_images/v_plotly.png + alttext: plotyで作られた箱ひげ図 + - + url: https://alta-viz.github.io/gallery/streamgraph.html + img: /images/content_images/v_altair.png + alttext: altairで作られたストリームグラフ + - + url: https://seaborn.pydata.org + img: /images/content_images/v_seaborn.png + alttext: 2種類のグラフによるペアプロット。seabornで作られたプロットと周波数グラフ" + - + url: https://docs.pyvista.org/examples/index.html + img: /images/content_images/v_pyvista.png + alttext: PyVista製の3Dボリュームレンダリング + - + url: https://napari.org + img: /images/content_images/v_napari.png + alttext: napariで作られた多次元画像 + - + url: https://vispy.org/gallery/index.html + img: /images/content_images/v_vispy.png + alttext: vispyで作られたボロノイ図 + content: + - + text: NumPyは、[Matplotlib](https://matplotlib.org)、[Seaborn](https://seaborn.pydata.org)、[Plotly](https://plot.ly)、[Altair](https://altair-viz.github.io)、[Bokeh](https://docs.bokeh.org/en/latest/)、[Holoviz](https://holoviz.org)、[Vispy](http://vispy.org)、[Napari](https://github.com/napari/napari)、[PyVista](https://github.com/pyvista/pyvista)などの、急成長している[Python visualization landscape](https://pyviz.org/overviews/index.html)に欠かせないコンポーネントです。 + - + text: NumPy の大規模配列の高速処理により、研究者は、ネイティブの Python が扱うことができるよりもはるかに大きなデータセットを可視化することができます。 diff --git a/content/ja/teams.md b/content/ja/teams.md index 365aedcf19..c91e538a59 100644 --- a/content/ja/teams.md +++ b/content/ja/teams.md @@ -1,11 +1,9 @@ --- -title: NumPy Teams +title: NumPy開発チーム sidebar: false --- -We are an international team on a mission to support scientific and research -communities worldwide by building quality, open-source software. -[Join us]({{< relref "/contribute" >}})! +私たちは、高品質のオープンソースソフトウェアを構築することで、世界中の科学・研究コミュニティをサポートすることを使命とする国際的なチームです。 是非[参加してください]({{< relref "/contribute" >}})! {{< include-html "static/gallery/maintainers.html" >}} @@ -19,6 +17,6 @@ communities worldwide by building quality, open-source software. {{< include-html "static/gallery/emeritus-maintainers.html" >}} -# Governance +# ガバナンス For the list of people on the Steering Council, please see [here](https://numpy.org/devdocs/dev/governance/people.html). diff --git a/content/ja/user-survey-2020.md b/content/ja/user-survey-2020.md new file mode 100644 index 0000000000..a6533d3cb8 --- /dev/null +++ b/content/ja/user-survey-2020.md @@ -0,0 +1,18 @@ +--- +title: 2020年 NumPyコミュニティ調査 +sidebar: false +--- + +2020年に、NumPyの調査チームは、ミシガン大学とメリーランド大学が共同で開催した、調査方法学の修士コースの学生と教員と共同で、初めて公式のNumPyコミュニティ調査を実施しました。 75カ国から1,200人以上のNumPyユーザーが参加してくれました。NumPyコミュニティの全体像を描き、プロジェクトの未来像についての意見を述べてもらいました。 + +{{< figure src="/surveys/NumPy_usersurvey_2020_report_cover.png" class="fig-left" alt="Cover page of the 2020 Numpy User survey report, titled 'Numpyコミュニティ調査2020 - 結果'" width="250">}} +src = '/surveys/NumPy_usersurvey_2020_report_cover.png' alt = 'Cover page of the 2020 NumPy user survey report, titled "NumPy Community Survey 2020 - results"' width = '250' +{{< /figure >}} + +調査結果を詳細を知りたい場合は、**[こちらのレポート](/surveys/NumPy_usersurvey_2020_report.pdf)** をダウンロードしてください。 + + +結果の概要については、 **[こちらの図](https://github.com/numpy/numpy-surveys/blob/master/images/2020NumPysurveyresults_community_infographic.pdf)** をチェックしてください。 + +より詳細が知りたくなりましたか？ **https://numpy.org/user-survey-2020-details/** をご覧ください。 + diff --git a/content/ja/user-surveys.md b/content/ja/user-surveys.md new file mode 100644 index 0000000000..7be9979c3a --- /dev/null +++ b/content/ja/user-surveys.md @@ -0,0 +1,10 @@ +--- +title: NumPyユーザアンケート +sidebar: false +--- + +**2020** NumPY調査チームは、ミシガン大学とメリーランド大学の学生や教員と協力して、最初の公式NumPyコミュニティ調査を実施しました。 アンケートの結果は[こちら](https://numpy.org/user-survey-2020/)をご覧ください。 + +**2021** 収集された調査データは現在解析中です。 + +過去または今後のNumPyユーザ調査に関する質問や提案がある場合は、[こちら](https://github.com/numpy/numpy-surveys/issues)にイシューを作成してください。 diff --git a/content/ko/404.md b/content/ko/404.md new file mode 100644 index 0000000000..41504d0c8a --- /dev/null +++ b/content/ko/404.md @@ -0,0 +1,8 @@ +--- +title: 404 +sidebar: false +--- + +앗! 잘못된 접근입니다. + +만약 이곳에 어떤 페이지가 있어야 한다면 [Issue 열기](https://github.com/numpy/numpy.org/issues)에서 문제를 제기할 수 있습니다. diff --git a/content/ko/_index.md b/content/ko/_index.md new file mode 100644 index 0000000000..0f06c2c787 --- /dev/null +++ b/content/ko/_index.md @@ -0,0 +1,49 @@ +--- +title: null +--- + +{{< grid columns="1 2 2 3" >}} + +{{< card >}} +title = '강력한 N차원 배열' +body = ''' +빠르고 다재다능한 NumPy의 벡터화, 인덱싱, 전송 구성은 오늘날 배열 컴퓨팅의 사실상 표준입니다. +''' +{{< /card >}} + +{{< card >}} +title = '수치적 컴퓨팅 도구' +body = ''' +NumPy는 포괄적인 수학 함수, 난수 생성기, 선형 대수 루틴, 푸리에 변환 등을 제공합니다. +''' +{{< /card >}} + +{{< card >}} +title = '오픈 소스' +body = ''' +자유 [BSD 라이선스](https://github.com/numpy/numpy/blob/main/LICENSE.txt)에 따라, NumPy는 흥미에 찼으며, 반응이 빠르고, 다양성이 넘치는 [커뮤니티](/community)에 의하여 [GitHub](https://github.com/numpy/numpy)에서 공개적으로 개발되고 유지됩니다. +''' +{{< /card >}} + +{{< card >}} +title = '상호운용성' +body = ''' +NumPy는 광범위한 하드웨어 및 컴퓨팅 플랫폼을 지원합니다. 또 분산형, GPU, 희소 배열 라이브러리와도 잘 작동합니다. +''' +{{< /card >}} + +{{< card >}} +title = '효율성' +body = ''' +NumPy의 핵심은 최적화된 C 코드로 구성되어 있습니다. 컴파일된 코드의 속도와 함께 Python의 유연함을 즐기세요. +''' +{{< /card >}} + +{{< card >}} +title = '쉬운 사용법' +body = ''' +NumPy의 고수준 문법은 어떤 배경이나 수준을 가지고 있는 프로그래머든 쉽게 접근하여 생산적인 일을 할 수 있도록 만들어줍니다. +''' +{{< /card >}} + +{{< /grid >}} diff --git a/content/ko/about.md b/content/ko/about.md new file mode 100644 index 0000000000..c05b97d1b8 --- /dev/null +++ b/content/ko/about.md @@ -0,0 +1,90 @@ +--- +title: NumPy 정보 +sidebar: false +--- + +NumPy는 Python을 통해 수치적 컴퓨팅을 할 수 있도록 도와주는 오픈소스 프로젝트입니다. Numerical와 Numarray라는 라이브러리의 초기 작업을 기반으로 2005년에 만들어졌습니다. NumPy는 항상 100% 오픈 소스 소프트웨어이며 누구나 무료로 사용할 수 있습니다 [수정된 BSD 라이선스](https://github.com/numpy/numpy/blob/main/LICENSE.txt)의 자유로운 조건에 따라 릴리스됩니다. + +NumPy는 NumPy와 더 넓은 과학 Python 커뮤니티의 합의를 통해 GitHub의 공개적으로 개발되었습니다. 거버넌스 접근 방식에 대한 자세한 내용은 [거버넌스 문서](https://www.numpy.org/devdocs/dev/governance/index.html)를 참조하세요. + + +## 운영 위원회 + +NumPy 운영 위원회는 프로젝트를 관리하는 기관입니다. 그 역할은 더 넓은 NumPy 커뮤니티와 협력하고 서비스를 제공함으로써 소프트웨어 패키지와 커뮤니티로서 프로젝트의 장기적인 지속 가능성을 보장하는 것입니다. NumPy 운영 위원회는 현재 다음과 같은 회원들로 구성되어 있습니다. (성씨의 알파벳 순서) + +- Sebastian Berg +- Ralf Gommers +- Charles Harris +- Stephan Hoyer +- Inessa Pawson +- Matti Picus +- Stéfan van der Walt +- Melissa Weber Mendonça +- Eric Wieser + +명예 회원 + +- Alex Griffing (2015-2017) +- Allan Haldane (2015-2021) +- Marten van Kerkwijk (2017-2019) +- Travis Oliphant (프로젝트 설립자, 2005-2012) +- Nathaniel Smith (2012-2021) +- Julian Taylor (2013-2021) +- Jaime Fernández del Río (2014-2021) +- Pauli Virtanen (2008-2021) + +NumPy 운영 위원회에 문의하려면, numpy-team@googlegroups.com 주소로 이메일을 보내세요. + +## 팀 + +NumPy 프로젝트 리더십은 프로젝트에 대한 기여 경로를 다양화하기 위해 적극적으로 노력하고 있습니다.
NumPy에는 현재 다음 팀이 있습니다: + +- 개발 +- 문서 +- 심사 +- 웹사이트 +- 설문조사 +- 번역 +- 스프린트 멘토링 +- 최적화 +- 자원 및 보조금 + +스프린트 멘토링 + +## NumFOCUS 소위원회 + +- Charles Harris +- Ralf Gommers +- Inessa Pawson +- Sebastian Berg +- 외부 회원: Thomas Caswell + +## 스폰서 + +NumPy는 다음과 같은 곳들에서 직접적으로 자금을 받습니다. +{{< sponsors >}} + + +## 기관 파트너 + +기관 파트너는 그들의 업무의 일환으로 NumPy에 기여하는 직원을 고용하여 프로젝트를 지원하는 조직입니다. 현재 기관 파트너는 다음과 같습니다. + +- UC 버클리 (Stéfan van der Walt) +- Quansight (Nathan Goldbaum, Ralf Gommers, Matti Picus, Melissa Weber Mendonça) +- NVIDIA (Sebastian Berg) + +{{< partners >}} + + +## 후원 + +만약 NumPy가 당신의 업무, 연구 혹은 회사에서 유용하다고 판단된다면 당신의 자원에 맞는 프로젝트에 기여하는 것을 고려해보세요. 그것이 얼마든 도움이 됩니다! 모든 후원은 NumPy의 소프트웨어 개발, 문서 작성과 커뮤니티 운영의 자금으로 엄격하게 사용될 것입니다. + +NumPy는 미국의 501(c)(3) 비영리 단체인 NumFOCUS의 후원 프로젝트입니다. NumFOCUS는 NumPy에 재정적, 법적, 행정적 지원을 제공하고 프로젝트의 건강과 지속 가능성을 보장할 수 있도록 도와줍니다. 더 자세한 정보를 알고싶다면 [numfocus.org](https://numfocus.org)를 방문하세요. + +NumPy에 대한 후원은 [NumFOCUS](https://numfocus.org)가 관리합니다. 미국에 거주하는 후원자의 경우에는, 당신의 후원은 법이 제공하는 한도 내에서 세금 공제를 받을 수 있습니다. 기부와 마찬가지로 특정 세금 상황에 대해서는 세금 전문가와 상담해야합니다. + +NumPy 운영 위원회는 후원받은 후원금을 가장 잘 활용하는 방안을 결정합니다. 기술 및 인프라의 우선 순위는 NumPy [NumPy Roadmap](https://www.numpy.org/neps/index.html#roadmap)에 문서화되어 있습니다. + +{{}} + diff --git a/content/ko/arraycomputing.md b/content/ko/arraycomputing.md new file mode 100644 index 0000000000..9c661978d2 --- /dev/null +++ b/content/ko/arraycomputing.md @@ -0,0 +1,21 @@ +--- +title: 배열 연산 +sidebar: false +--- + +*배열 연산은 통계와 수학 뿐만 아니라 현대의 다양한 분야에 적용되는 데이터 사이언스, 데이터 시각화나 디지털 신호 처리, 영상 처리, 의생명 정보 공학, 기계학습, AI 등 다양한 분야에서 적용되는 데이터 분석 어플리케이션의 기반입니다.* + +대규모 데이터의 조작과 연산은 고효율, 고성능의 배열 연산에 달려있습니다. **Python**은 데이터 과학자, 머신 러닝 개발자, 그리고 효율적인 수치 계산을 필요로 하는 분야에서 선택되는 프로그래밍 언어입니다. + +**Num**erical **Py**thon 또는 NumPy 는 파이썬의 표준라이브러리에는 포함되지 않지만, 대규모, 다차원 행렬을 표현할 수 있고, 배열 연산을 위한 고수준의 수학 함수들을 포함한 라이브러리입니다. + +2006년에 NumPy가 출시된 이후로, 2008년에 이를 기반으로 Pandas가 나타났습니다. 그리고 몇년전까지도, 다양한 배열 연산 라이브러리가 잇따라 나오며 배열 연산 분야가 더욱 활발해 졌습니다. 최신의 라이브러리들중 대부분은 NumPy 같은 특징과 성능을 모방하고, 새로운 알고리즘이나 머신러닝이나 인공지능 어플리케이션을 위한 특화된 기능을 포함하고 있습니다. + +arraycl + +**배열 연산**의 기반은 **array ** 자료구조 입니다. *배열*은 대규모의 데이터를 정렬, 검색, 수학 계산, 그리고 변형을 쉽고 빠르게 처리하는데 사용됩니다. + +배열 연산은 *한번에 * 데이터 배열에 *모든 연산이* 계산 됩니다. 다시 말해서, 모든 배열 연산은 전체 데이터에 한번에 적용됩니다. 이 벡터화 접근법은 배열 연산을 위해 루프를 활용하여 개별적인 데이터에 접근하여 연산하는 코드를 작성하지 않고, 배열에 바로 연산하는 코드를 작성하여, 개발자가 보다 개발 빠르고 간단하게 할수 있게 해줍니다. diff --git a/content/ko/case-studies/blackhole-image.md b/content/ko/case-studies/blackhole-image.md new file mode 100644 index 0000000000..14ddb59afc --- /dev/null +++ b/content/ko/case-studies/blackhole-image.md @@ -0,0 +1,80 @@ +--- +title: "사례 연구: 최초의 블랙홀 사진" +sidebar: false +--- + +{{< figure src="/images/content_images/cs/blackhole.jpg" caption="**블랙홀 M87**" alt="블랙홀 사진" attr="*(사진 크레딧: Event Horizon Telescope Collaboration)*" attrlink="https://www.jpl.nasa.gov/images/universe/20190410/blackhole20190410.jpg" >}} +src = '/images/content_images/cs/blackhole.jpg' title = '블랙홀 M87' alt = '블랙홀 사진' attribution = '(사진 크레딧: Event Horizon Telescope Collaboration)' attributionlink = 'https://www.jpl.nasa.gov/images/universe/20190410/blackhole20190410.jpg' +{{< /figure >}} + +{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="Katie Bouman, *Assistant Professor, Computing & Mathematical Sciences, Caltech*" +> }} Imaging the M87 Black Hole is like trying to see something that is by definition impossible to see. +> +> {{< /blockquote >}} + +## 지구 크기의 망원경 + +[사건의 지평선 망원경(EHT)](https://eventhorizontelescope.org)은 8개의 지상 전파 망원경으로 구성된 지구 크기의 전산 망원경으로, 전례없는 감도와 해상도로 우주를 연구하는 데 쓰입니다. 초장기선 간섭 관측법(VLBI)이라는 기술을 사용하는 거대한 가상 망원경의 각해상도는 [20 마이크로각초][resolution]에 달하며 파리의 길거리 카페에서 뉴욕의 신문을 읽기에 충분한 정도입니다! + +### 주요 목표 및 결과 + +* **우주를 보는 새로운 방식:** EHT라는 획기적인 발상의 토대는 [아서 에딩턴 경][eddington]의 관측으로 아인슈타인의 일반 상대성이론이 최초로 관측적 지지를 받았던 시기인 100년 전에 마련되었습니다. + +* **블랙홀:** EHT는 처녀자리 은하단의 Messier 87(M87) 은하의 중심부에 있는 초대질량 블랙홀로 훈련되었으며 이는 지구에서 약 5500만 광년 떨어져 있습니다. 이 천체의 질량은 태양의 65억 배입니다. [100년 넘게](https://www.jpl.nasa.gov/news/news.php?feature=7385) 연구되었으나, 블랙홀을 시각적으로 볼 수 있게 구현한 바는 없었습니다. + +* **관찰과 이론의 비교:** 아인슈타인의 일반 상대성이론에 따라 과학자들은 중력의 시공간 왜곡이나 빛 흡수에 의해 어둡게 보이는 영역이 나타날 것으로 예측하였습니다. 과학자들은 이를 블랙홀의 엄청난 질량을 재는 데 이용할 수 있었죠. + +### 도전 + +* **계산의 규모** + + EHT는 급격한 대기 위상의 변동, 큰 기록 대역폭, 완전히 다르고 지리적으로 분산된 망원경 등의 문제를 포함한 막대한 데이터를 처리해야 하는 문제를 낳습니다. + +* **지나치게 많은 정보** + + EHT는 매일 350 테라바이트의 관측 결과를 생성하며, 이 정보는 헬륨으로 채운 하드 드라이브에 저장됩니다. 이토록 많은 데이터의 양과 복잡성을 줄여나가는 것은 지극히 어려운 일입니다. + +* **잘 알지 못함** + + 만약 목표가 이전에 본 적이 없는 것을 보는 것이라면, 과학자들은 어떻게 이 사진이 옳다고 입증할 수 있을까요? + +{{< figure >}} +src = '/images/content_images/cs/dataprocessbh.png' title = 'EHT 데이터 처리 파이프라인' alt = '데이터 파이프라인' align = 'center' attribution = '(다이어그램 크레딧: The Astrophysical Journal, Event Horizon Telescope Collaboration)' attributionlink = 'https://iopscience.iop.org/article/10.3847/2041-8213/ab0c57' +{{< figure src="/images/content_images/cs/bh_numpy_role.png" class="fig-center" alt="numpy의 역할" caption="**블랙홀 시각화에서 NumPy의 역할**" >}} + +## NumPy의 역할 + +데이터에 만약 문제가 있다면 어떨까요? 아니면 알고리즘이 특정 가정에 지나치게 의존할 수도 있습니다. 매개변수 하나만 달라져도 사진이 크게 바뀔까요? + +EHT는 기존 및 최첨된 이미지 재구성 기술을 모두 사용한 뒤, 개개의 팀이 데이터를 평가하도록 하여 이런 문제를 해결했습니다. 결과가 일관적이라는 것을 검증한 뒤, 이들을 결합해 최초의 블랙홀 이미지를 만들어내었습니다. + +그들의 연구는 협업 데이터 분석을 통해 과학을 발전시키는 과학적인 Python 생태계의 역할을 보여줍니다. + +{{< figure >}} +{{< figure src="/images/content_images/cs/numpy_bh_benefits.png" class="fig-center" alt="numpy를 통한 이익" caption="**활용된 주요 NumPy 기능**" >}} +{{< /figure >}} + +예를 들어, [`eht-imaging`][ehtim] Python 패키지는 VLBI 데이터를 통해 실험이나 이미지 재구성을 수행할 때 필요한 도구를 제공합니다. NumPy는 아래 소프트웨어 종속성 차트에 나와 있는 것처럼 이 패키지에서 사용되는 배열 데이터 처리의 핵심 역할을 합니다. + +{{< figure >}} +src = '/images/content_images/cs/ehtim_numpy.png' alt = 'numpy를 강조하는 ehtim의 종속성 지도' title = 'NumPy를 강조하는 ehtim 패키지의 소프트웨어 종속성 차트' +{{< /figure >}} + +NumPy 외에도 [SciPy](https://www.scipy.org)와 [Pandas](https://pandas.io) 등의 다른 많은 패키지가 블랙홀을 시각화하는 데이터 처리 파이프라인의 일부입니다. 표준 천문 파일 형식과 시간/좌표 변환에는 [Astropy][astropy]가 쓰였고 [Matplotlib][mpl]는 분석 과정 전체에서 블랙홀의 최종 사진을 생성하는 등 데이터를 시각화하는 데 쓰였습니다. + +## 요약 + +NumPy의 핵심 기능인 효율적이고 유용한 n차원 배열은 연구자들이 대규모 수치 데이터셋을 다룰 수 있도록 하여 최초의 블랙홀 사진을 만드는 데 토대를 제공했습니다. 이번 관측은 아인슈타인의 이론에 훌륭한 시각적 증거를 준 관측으로, 과학계에 한 획을 그은 순간이었습니다. 기술적 혁신뿐만 아니라 200명 이상의 과학자와 세계 최고의 전파 관측소 간의 국제 협력도 이루어 냈습니다. 기존의 천문학 모델을 개선한 혁신적인 알고리즘과 데이터 처리 기술이 우주의 비밀을 알아내는 데 도움을 주었습니다. + +{{< figure >}} +src = '/images/content_images/cs/numpy_bh_benefits.png' alt = 'numpy를 통한 이익' title = '활용된 주요 NumPy 기능' +{{< /figure >}} + +[resolution]: https://eventhorizontelescope.org/press-release-april-10-2019-astronomers-capture-first-image-black-hole + +[eddington]: https://ko.wikipedia.org/wiki/%EC%95%84%EC%84%9C_%EC%8A%A4%ED%83%A0%EB%A6%AC_%EC%97%90%EB%94%A9%ED%84%B4#%EC%9D%BC%EB%B0%98%EC%83%81%EB%8C%80%EC%84%B1%EC%9D%B4%EB%A1%A0%EC%9D%98_%EC%8B%A4%ED%97%98%EC%A0%81_%EA%B2%80%EC%A6%9D + +[ehtim]: https://github.com/achael/eht-imaging + +[astropy]: https://www.astropy.org/ +[mpl]: https://matplotlib.org/ diff --git a/content/ko/case-studies/cricket-analytics.md b/content/ko/case-studies/cricket-analytics.md new file mode 100644 index 0000000000..7dfb7ea1df --- /dev/null +++ b/content/ko/case-studies/cricket-analytics.md @@ -0,0 +1,72 @@ +--- +title: "사례 연구: 판도를 뒤집은 크리켓 통계!" +sidebar: false +--- + +{{< figure src="/images/content_images/cs/ipl-stadium.png" caption="**인도 최대의 크리켓 축제인 IPLT20**" alt="인도 프리미어 리그 크리켓 컵 및 경기장" attr="*(사진 출처: IPLT20 (컵 및 로고) & Akash Yadav (경기장))*" attrlink="https://unsplash.com/@aksh1802" >}} +src = '/images/content_images/cs/ipl-stadium.png' title = '인도 최대의 크리켓 축제인 IPLT20' alt = '인도 프리미어 리그 크리켓 컵 및 경기장' attribution = '(사진 크레딧: IPLT20 (컵 및 로고) & Akash Yadav (경기장))' attributionlink = 'https://unsplash.com/@aksh1802' +{{< /figure >}} + +{{< blockquote cite="https://www.scoopwhoop.com/sports/ms-dhoni/" by="M S Dhoni, *International Cricket Player, ex-captain, Indian Team, plays for Chennai Super Kings in IPL*" +> }} You don't play for the crowd, you play for the country. +> +> {{< /blockquote >}} + +## 크리켓이란 + +인도인들이 크리켓과 사랑에 빠졌다고 해도 과언이 아닙니다. 크리켓은 인도의 거의 모든 지역 구석구석에서 시골이든 도시든 상관없이 사랑받고 있습니다. 다른 스포츠와 달리 인도의 수십억 명을 연결하는 매개체 역할을 하는 데다 남녀노소 모두에게 인기가 있습니다. 크리켓은 많은 미디어의 관심을 받고 있기도 합니다. 엄청난 [돈](https://www.statista.com/topics/4543/indian-premier-league-ipl/)과 명성이 달려 있기도 하죠. 최근 몇 년 동안, 기술이 이 분야의 판도를 뒤집어 버렸습니다. 청중들은 스트리밍 미디어, 토너먼트, 모바일 기기를 통해 실시간 크리켓 경기를 저렴하게 볼 수 있습니다. + +인도 프리미어 리그(IPL)는 2008년 설립되어 20개 팀으로 구성된 프로 크리켓 리그입니다. 이는 세계에서 가장 참가자가 많은 크리켓 이벤트 중 하나로, 2019년에 [67억 달러](https://en.wikipedia.org/wiki/Indian_Premier_League)에 달하는 가치로 추산됩니다. + +크리켓은 숫자의 게임이다 - 타자가 득점하고, 위켓이 가져간다. 크리켓 팀이 이긴 경기, 타자가 이긴 횟수 일종의 볼링 공격 등에 특정한 방식으로 반응합니다. NumPy와 같은 수치 컴퓨팅 소프트웨어로 구동되는 강력한 분석 도구를 통해 크리켓의 성능 향상과 사업 기회, 전반적인 시장, 경제학을 연구하기 위한 크리켓 숫자를 파헤칠 수 있는 능력은 큰 일이다. 크리켓 분석은 게임에 대한 흥미로운 통찰력과 게임 결과에 대한 예측 지능을 제공한다. + +오늘날, 크리켓 경기 기록의 풍부하고 거의 무한한 트로브가 있습니다. 이용 가능한 통계, 예를 들어, [ ESPN cricinfo](https://stats.espncricinfo.com/ci/engine/stats/index.html) 및 [cricsheet](https://cricsheet.org). 이들 및 몇몇 그러한 크리켓 데이터베이스는 최신 머신 러닝 및 예측 모델링 알고리즘을 이용한 [크리켓 분석](https://www.researchgate.net/publication/336886516_Data_visualization_and_toss_related_analysis_of_IPL_teams_and_batsmen_performances)에 사용되어 왔습니다. 게임과 관련된 전문 스포츠 기관과 함께 미디어 및 엔터테인먼트 플랫폼은 경기 승리 기회를 향상시키기 위한 주요 측정 기준을 결정하기 위한 기술 및 분석을 사용합니다. + +* 타격 성적 이동 평균, +* 점수 예측, +* 다른 상대에 맞서 선수의 체력과 경기력에 대한 통찰력을 얻기, +* 팀 구성에 대한 전략적 결정을 내리는 플레이어의 승패 기여도 + +{{< figure >}} +src = '/images/content_images/cs/cricket-pitch.png' title = '경기장의 중심이 되는 크리켓 피치' alt = '볼러와 배트맨으로 이루어진 크리켓 피치' align = 'center' attribution = '(사진 크레딧: Debarghya Das)' attributionlink = 'http://debarghyadas.com/files/IPLpaper.pdf' +{{< figure src="/images/content_images/cs/numpy_ca_benefits.png" class="fig-center" alt="NumPy를 크리켓 분석에 사용했을 때의 이익을 보여주는 다이어그램" caption="**활용된 주요 NumPy 기능**" >}} + +### 데이터 분석의 주요 목표 + +* 스포츠 데이터는 크리켓에서뿐만 아니라 [다른 스포츠](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx)에서도 팀의 전체 역량과 승리 확률을 높이는 데 쓰입니다. +* 실시간 데이터 분석은 경기 중에도 팀과 관련 사업의 변화하는 전략에 대한 통찰력을 확보하여 경제적 이익과 성장을 도모하는 데 도움이 될 수 있습니다. +* 과거 분석 외에도 예측 모델을 활용하여 상당한 수의 크런칭과 데이터 과학 노하우, 시각화 도구 및 분석에 더 새로운 관찰을 포함시킬 수 있는 기능이 필요한 가능한 일치 결과를 결정합니다. + +{{< figure >}} +src = '/images/content_images/cs/player-pose-estimator.png' alt = '자세 예측' title = '크리켓 자세 예측' attribution = '(사진 크레딧: connect.vin)' attributionlink = 'https://connect.vin/2019/05/ai-for-cricket-batsman-pose-analysis/' +{{< /figure >}} + +### 도전 + +* **데이터 정리 및 전처리** + + IPL은 크리켓을 고전적인 테스트 매치 형식에서 훨씬 더 큰 규모로 확대시켰습니다. 매 시즌 다양한 형식으로 열리는 경기의 수가 증가하고 있으며, 데이터, 알고리즘, 최신 스포츠 데이터 분석 기술, 시뮬레이션 모델 또한 증가하고 있습니다. 크리켓 데이터 분석에는 필드 매핑, 플레이어 추적, 공 추적, 플레이어의 타격 분석 및 공이 어떻게 움직이는지에 대한 각도, 스핀, 속도, 궤도 등 다른 많은 종류의 데이터를 필요로 합니다. 이 수많은 인자들은 데이터 정리 및 전처리 과정의 복잡성을 증가시켰습니다. + +* **동적 모델링** + + 크리켓에서는 다른 스포츠와 마찬가지로 다양한 선수의 수, 선수의 속성, 공이나 잠재적 행동의 가능성 등 여러 가능성을 추적할 때 많은 변수가 작용합니다. 데이터 분석 및 모델링의 복잡성은 분석 중 제시되는 예측 질문의 종류에 비례하며, 데이터 표현 및 모델에 크게 의존합니다. 타자가 다른 각도나 속도로 공을 쳤을 때 일어날 일과 같은 동적인 크리켓 경기를 예측할 때, 계산이나 데이터 비교 측면에서 상황이 훨씬 더 어려워집니다. + +* **예측 분석의 복잡성** + + 크리켓에서 의사결정의 상당 부분은 '볼 전달이 특정 유형일 경우 타자가 얼마나 자주 특정 종류의 샷을 하느냐', '배트맨이 특정 방식으로 전달에 반응하면 볼러가 라인과 길이를 어떻게 바꾸느냐' 등의 질문에 따른 것입니다. 이러한 예측 분석 쿼리는 매우 세분화된 데이터셋 가용성과 데이터를 합성하고 정확도가 높은 생성 모델을 만들 수 있는 기능이 필요합니다. + +## 크리켓 분석에서 NumPy의 역할 + +스포츠 분석은 현재 매우 활발한 분야입니다. 많은 연구자들과 기업체에서는 최신 머신러닝 및 AI 기법을 쓰는 대신 [NumPy](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx)나 Scikit-learn, SciPy, Matplotlib, Jupyter같은 PyData 패키지를 이용합니다. NumPy는 크리켓과 관련된 여러 스포츠 통계에 다음과 같이 쓰였습니다. + +* **통계적 분석:** NumPy의 수치적 기능은 다양한 플레이어 및 게임 전술에서 관찰 데이터 또는 경기의 통계적 중요성을 추정하는 데 도움을 주거나, 생성적 또는 정적 모델과 비교하여 게임 결과를 추정합니다. 전술 분석에는 [인과 분석](https://amplitude.com/blog/2017/01/19/causation-correlation) 및 [빅데이터 접근법](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4996805/)이 쓰입니다. + +* **데이터 시각화:** 그래프 그리기 및 [시각화](https://towardsdatascience.com/advanced-sports-visualization-with-pandas-matplotlib-and-seaborn-9c16df80a81b)는 다양한 데이터셋 사이의 관계를 볼 수 있는 유용한 관점을 제공해 줍니다. + +## 요약 + +스포츠 분석은 프로 게임의 판도를 바꿀 것입니다. 특히 최근까지는 주로 "직감"이나 과거부터 내려오던 것을 답습하는 식으로 이뤄진 전략적 의사 결정에 대해서 말입니다. NumPy는 데이터 분석, 기계 학습 및 AI 알고리즘과 관련하여 더욱 높은 수준의 기능을 제공하는 Python 패키지들의 견고한 기반을 제공합니다. 이들 패키지는 크리켓 경기뿐 아니라 크리켓 관련 추론이나 사업을 추진하면서, 판도를 바꿀만한 결정을 이끌어 내는 영감을 실시간으로 제공하는 데 널리 이용되고 있습니다. 크리켓 경기의 결과로 이어지는 숨겨진 매개변수, 패턴이나 속성을 찾는 것은 관계자가 숫자와 통계에 숨겨진 게임을 분석하는 방법을 파악하는 데 도움이 됩니다. + +{{< figure >}} +src = '/images/content_images/cs/numpy_ca_benefits.png' alt = 'NumPy를 크리켓 분석에 적용했을 때의 이익을 보여주는 다이어그램' title = '활용된 주요 NumPy 기능' +{{< /figure >}} diff --git a/content/ko/case-studies/deeplabcut-dnn.md b/content/ko/case-studies/deeplabcut-dnn.md new file mode 100644 index 0000000000..217e1565f2 --- /dev/null +++ b/content/ko/case-studies/deeplabcut-dnn.md @@ -0,0 +1,102 @@ +--- +title: "사례 연구: DeepLabCut 3D 포즈 추정" +sidebar: false +--- + +{{< figure src="/images/content_images/cs/mice-hand.gif" class="fig-center" caption="**DeepLapCut을 활용한 쥐의 손 움직임 분석**" alt="쥐 손 애니메이션" attr="*(출처: www.deeplabcut.org )*" attrlink="http://www.mousemotorlab.org/deeplabcut">}} +src = '/images/content_images/cs/mice-hand.gif' title = 'DeepLapCut을 활용한 쥐의 손 움직임 분석' alt = '쥐 손 애니메이션' attribution = '(출처: www.deeplabcut.org )' attributionlink = 'http://www.mousemotorlab.org/deeplabcut' +{{< /figure >}} + +{{< blockquote cite="https://news.harvard.edu/gazette/story/newsplus/harvard-researchers-awarded-czi-open-source-award/" by="Alexander Mathis, *Assistant Professor, École polytechnique fédérale de Lausanne* ([EPFL](https://www.epfl.ch/en/))" +> }} 오픈소스 소프트웨어는 생물 의학의 발전을 가속화하고 있습니다. DeepLabCut은 딥 러닝을 사용하여 동물 행동에 대한 자동화된 비디오 분석을 가능하게 합니다. +> +> {{< /blockquote >}} + +## DeepLabCut 소개 + +[DeepLabCut](https://github.com/DeepLabCut/DeepLabCut) 은 전 세계 수백 개 기관의 연구원이 인간 수준의 정확도로 매우 적은 훈련 데이터로 실험실 동물의 행동을 추적할 수 있는 오픈 소스 도구 상자입니다. DeepLabCut 기술을 통해 과학자들은 동물 종과 시간 척도에 걸쳐 운동 제어 및 행동에 대한 과학적 이해를 더 깊이 탐구할 수 있습니다. + +신경 과학, 의학 및 생체 역학을 포함한 여러 연구 분야에서 동물의 움직임을 추적한 데이터를 사용합니다. DeepLabCut은 필름에 기록된 동작을 구문 분석하여 인간과 다른 동물이 하는 일을 이해하는 데 도움을 줍니다. 심층 신경망 기반 데이터 분석과 함께 태깅 및 모니터링의 힘든 작업에 자동화를 사용하는 DeepLabCut은 영장류, 생쥐, 물고기, 파리 등과 같은 동물 관찰과 관련된 과학적 연구를 훨씬 빠르고 정확하게 만듭니다. + +{{< figure >}} +src = '/images/content_images/cs/race-horse.gif' title = '경주마 신체 부위의 위치를 추적하는 색 점' alt = '경주마 애니메이션' attribution = '(출처: Mackenzie Mathis)' +{{< /figure >}} + +DeeDeepLabCut의 동물 자세 추출을 통한 동물의 비침습적 행동 추적은 생체 역학, 유전학, 행동학 & 신경 과학과 같은 영역에서 과학적 추구에 매우 중요합니다. 동적으로 변화하는 배경에서 마커 없이 비디오에서 동물 포즈를 비침습적으로 측정하는 것은 기술적으로 뿐만 아니라 필요한 리소스 요구 사항 및 필요한 훈련 데이터 측면에서 계산적으로 어려운 일입니다. + +DeepLabCut을 사용하면 연구자가 피험자의 자세를 추정하고 Python 기반 소프트웨어 툴킷을 통해 행동을 정량화할 수 있습니다. DeepLabCut을 사용하여 연구원은 비디오에서 고유한 프레임을 식별하고 맞춤형 GUI를 사용하여 수십 개의 프레임에서 특정 신체 부위에 디지털 레이블을 지정한 다음 DeepLabCut의 딥 러닝 기반 포즈 추정 아키텍처가 나머지 프레임에서 동일한 기능을 선택하는 방법을 학습할 수 있습니다. 그것은 파리와 생쥐와 같은 일반적인 실험실 동물에서 [치타][cheetah-movement]와 같은 좀 더 특이한 동물에 이르기까지 다양한 동물 종에 걸쳐 작동합니다. + +DeepLabCut은 [전이 학습](https://arxiv.org/pdf/1909.11229), 이라는 원리를 사용하여 필요한 훈련 데이터의 양을 크게 줄이고 훈련 기간의 수렴 속도를 높입니다. 필요에 따라 사용자는 실시간 실험 피드백과 결합할 수 있는 더 빠른 추론(예: MobileNetV2)을 제공하는 다양한 네트워크 아키텍처를 선택할 수 있습니다. DeepLabCut은 원래 [DeeperCut](https://arxiv.org/abs/1605.03170),이라는 최고 성능의 인간 포즈 추정 아키텍처의 특징 검출기를 사용했습니다. 이 이름에 영감을 받았습니다. 이제 패키지가 추가 아키텍처, 증강 방법 및 전체 프런트 엔드 사용자 경험을 포함하도록 크게 변경되었습니다. 또한 대규모 생물학적 실험을 지원하기 위해 DeepLabCut은 능동적 학습 기능을 제공하므로 사용자는 시간이 지남에 따라 트레이닝 세트를 늘려 엣지 케이스를 다루고 특정 상황 내에서 포즈 추정 알고리즘을 강력하게 만들 수 있습니다. + +최근에는, 영장류의 안면 분석부터 개 자세까지 다양한 종과 실험 조건에 대해 사전 훈련된 모델을 제공하는 [DeepLabCut 모델 동물원](http://www.mousemotorlab.org/dlc-modelzoo) 이 도입되었습니다. 예를 들어 새 데이터의 레이블 지정이나 신경망 교육 없이 클라우드에서 실행할 수 있으며 프로그래밍 경험이 필요하지 않습니다. + +### 주요 목표 및 결과 + +* **과학적 연구를 위한 동물 자세 분석 자동화:** + + DeepLabCut 기술의 주요 목표는 다양한 환경에서 동물의 자세를 측정하고 추적하는 것입니다. 예를 들어, 이 데이터는 뇌가 움직임을 제어하는 ​​방법을 이해하거나 동물이 사회적으로 상호 작용하는 방식을 설명하기 위해 신경 과학 연구에서 사용될 수 있습니다. 연구원들은 DeepLabCut으로 [성능이 10배 향상](https://www.biorxiv.org/content/10.1101/457242v1) 되는 것을 관찰했습니다. 포즈는 최대 1200 초당 프레임 수(FPS)로 오프라인에서 추론할 수 있습니다. + +* **포즈 추정을 위해 사용하기 쉬운 Python 툴킷 생성:** + + DeepLabCut은 연구자들이 쉽게 채택할 수 있는 사용하기 쉬운 도구 형태로 동물 자세 추정 기술을 공유하고 싶었습니다. 그래서 그들은 프로젝트 관리 기능도 포함된 완전하고 사용하기 쉬운 Python 툴킷을 만들었습니다. 이를 통해 포즈 추정의 자동화는 물론 DeepLabCut Toolkit 사용자가 데이터 세트 수집 단계부터 공유 가능하고 재사용 가능한 분석 파이프라인 생성에 이르기까지 프로젝트 전체를 관리할 수 있습니다. + + [툴킷][DLCToolkit]은 현재 오픈 소스로 제공됩니다. + + 일반적인 DeepLabCut 작업 흐름에는 다음이 포함됩니다. + + - 능동적 학습을 통한 훈련 세트 생성 및 정제 + - 특정 동물 및 시나리오를 위한 맞춤형 신경망 생성 + - 동영상에 대한 대규모 추론을 위한 코드 + - 통합 시각화 도구를 사용하여 추론 도출 + +{{< figure >}} +{{< figure src="/images/content_images/cs/deeplabcut-workflow.png" class="fig-center" caption="**DeepLabCut 워크플로우**" alt="워크플로우" attr="*(출처: Mackenzie Mathis)*" attrlink="https://www.researchgate.net/figure/DeepLabCut-work-flow-The-diagram-delineates-the-work-flow-as-well-as-the-directory-and_fig1_329185962">}} +{{< figure >}} + +### 도전 + +* **속도** + + 동물의 행동을 측정하고 동시에 과학 실험을 보다 효율적이고 정확하게 하기 위해 동물 행동 비디오를 빠르게 처리합니다. 동적으로 변화하는 배경에서 마커 없이 실험실 실험을 위해 상세한 동물 포즈를 추출하는 것은 기술적으로 뿐만 아니라 필요한 리소스 요구 사항 및 필요한 교육 데이터 측면에서 어려울 수 있습니다. 컴퓨터 비전 전문 지식과 같은 기술 없이도 사용하기 쉬운 도구를 생각해내어 과학자들이 보다 실제 상황에서 연구를 수행할 수 있도록 하는 것은 해결해야 할 사소한 문제가 아닙니다. + +* **조합론** + + 조합론은 여러 사지의 움직임을 개별 동물 행동으로 조립하고 통합하는 것을 포함합니다. 키포인트와 연결을 개별 동물 움직임으로 조립하고 시간에 따라 연결하는 것은 특히 실험 비디오에서 여러 동물 움직임을 추적하는 경우 강력한 수치 분석이 필요한 복잡한 프로세스입니다. + +* **데이터 처리** + + 마지막으로 배열 조작 - 다양한 이미지, 대상 텐서 및 키포인트에 해당하는 대규모 배열 스택을 처리하는 것은 상당히 어렵습니다. + +{{< figure src="/images/content_images/cs/numpy_dlc_benefits.png" class="fig-center" alt="numpy를 통한 이익" caption="**활용한 주요 NumPy 기능**" >}} +src = '/images/content_images/cs/pose-estimation.png' title = '포즈 추정 변수 및 복잡도' alt = '난점 설명' align = 'center' attribution = '(출처: Mackenzie Mathis)' attributionlink = 'https://www.biorxiv.org/content/10.1101/476531v1.full.pdf' +{{< /figure >}} + +## 포즈 추정 문제를 해결하는 NumPy의 역할 + +NumPy는 행동 분석을 위한 고속 수치 계산에 대한 DeepLabCut 기술의 핵심 요구 사항을 해결합니다. NumPy 외에도 DeepLabCut은 [SciPy](https://www.scipy.org), [Pandas](https://pandas.pydata.org), [matplotlib](https://matplotlib.org), [Tensorpack](https://github.com/tensorpack/tensorpack), [imgaug](https://github.com/aleju/imgaug), [scikit-learn](https://scikit-learn.org/stable/), [scikit-image](https://scikit-image.org) 그리고 [Tensorflow](https://www.tensorflow.org) 와 같이 핵심에서 NumPy를 활용하는 다양한 Python 소프트웨어를 사용합니다. + +NumPy의 다음 기능은 이미지 처리, 조합 요구 사항 및 DeepLabCut 포즈 추정 알고리즘의 빠른 계산 요구 사항을 해결하는 데 중요한 역할을 했습니다. + +* 벡터화 +* 마스킹된 어레이 작업 +* 선형 대수학 +* 무작위 샘플링 +* 큰 배열의 재구성 + +DeepLabCut은 툴킷에서 제공하는 워크플로 전체에서 NumPy의 어레이 기능을 활용합니다. 특히, NumPy는 사람의 주석 레이블 지정을 위한 개별 프레임 샘플링과 주석 데이터 작성, 편집 및 처리에 사용됩니다. TensorFlow 내에서 신경망은 DeepLabCut 기술로 수천 번의 반복을 통해 훈련되어 프레임에서 실측 주석을 예측합니다. 이를 위해 이미지 대 이미지 변환 문제로 포즈 추정을 캐스팅하기 위해 목표 밀도(점수 지도)가 생성됩니다. 신경망을 강력하게 만들기 위해 다양한 기하 및 이미지 처리 단계에 따라 대상 스코어맵을 계산해야 하는 데이터 확대가 사용됩니다. 학습을 빠르게 하기 위해 NumPy의 벡터화 기능이 활용됩니다. 추론을 위해 대상 스코어맵에서 가장 가능성이 높은 예측을 추출해야 하며 효율적으로 "개별 동물을 조립하기 위해 예측을 연결"해야 합니다. + +{{< figure >}} +src = '/images/content_images/cs/deeplabcut-workflow.png' title = 'DeepLabCut 워크플로우' alt = '워크플로우' attribution = '(출처: Mackenzie Mathis)' attributionlink = 'https://www.researchgate.net/figure/DeepLabCut-work-flow-The-diagram-delineates-the-work-flow-as-well-as-the-directory-and_fig1_329185962' +{{< /figure >}} + +## 요약 + +행동을 관찰하고 효율적으로 설명하는 것은 현대 행동학, 신경과학, 의학 및 기술의 핵심 테넌트입니다. [DeepLabCut](http://orga.cvss.cc/wp-content/uploads/2019/05/NathMathis2019.pdf)을 사용하면 연구자가 피험자의 자세를 추정하여 행동을 정량화할 수 있습니다. DeepLabCut Python 툴킷은 작은 훈련 이미지 세트만으로 인간 수준의 라벨링 정확도 내에서 신경망을 훈련할 수 있습니다. 따라서 실험실에서의 행동 분석뿐만 아니라 잠재적으로 스포츠, 보행 분석, 의학 및 재활 연구에도 응용 분야를 확장합니다. DeepLabCut 알고리즘이 직면한 복잡한 조합, 데이터 처리 문제는 NumPy의 배열 조작 기능을 사용하여 해결됩니다. + +{{< figure >}} +src = '/images/content_images/cs/numpy_dlc_benefits.png' alt = 'numpy를 통한 이익' title = '활용된 주요 NumPy 기능' +{{< /figure >}} + +[cheetah-movement]: https://www.technologynetworks.com/neuroscience/articles/interview-a-deeper-cut-into-behavior-with-mackenzie-mathis-327618 + +[DLCToolkit]: https://github.com/DeepLabCut/DeepLabCut diff --git a/content/ko/case-studies/gw-discov.md b/content/ko/case-studies/gw-discov.md new file mode 100644 index 0000000000..e72252a423 --- /dev/null +++ b/content/ko/case-studies/gw-discov.md @@ -0,0 +1,77 @@ +--- +title: "사례 연구: 중력파의 발견" +sidebar: false +--- + +{{< figure src="/images/content_images/cs/gw_sxs_image.png" class="fig-center" caption="**중력파**" alt="이항 결합하며 중력파를 생성하는 블랙홀" attr="*(사진 크레딧: LIGO의 Simulating eXtreme Spacetimes (SXS) 프로젝트)*" attrlink="https://youtu.be/Zt8Z_uzG71o" >}} +src = '/images/content_images/cs/gw_sxs_image.png' title = '중력파' alt = '이항 결합하며 중력파를 생성하는 블랙홀' attribution = '(사진 크레딧: LIGO의 Simulating eXtreme Spacetimes (SXS) 프로젝트)' attributionlink = 'https://youtu.be/Zt8Z_uzG71o' +{{< /figure >}} + +{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="David Shoemaker, *LIGO Scientific Collaboration*" >}} The scientific Python ecosystem is critical infrastructure for the research done at LIGO. +{{< /blockquote >}} + +## [중력파](https://www.nationalgeographic.com/news/2017/10/what-are-gravitational-waves-ligo-astronomy-science/) 그리고 [LIGO](https://www.ligo.caltech.edu)에 대해 + +중력파는 '시공간 천막'의 물결이라고 할 수 있으며, 두 블랙홀의 충돌이나 병합, 쌍성의 결합 혹은 초신성과 같이 우주가 대격변하는 사건으로부터 생성됩니다. 중력파를 관측하는 것은 비단 중력 연구에 도움을 줄 뿐만 아니라 먼 우주에서의 모호한 현상들과 이것이 미치는 영향에 대해서도 이해할 수 있게 해 줍니다. + +병합에서 방출되는 중력파는 슈퍼컴퓨터를 사용하는 무차별 대입 수치 상대성 이론을 제외하고는 어떤 기술로도 계산할 수 없습니다. LIGO가 수집하는 데이터의 양은 중력파 신호가 작은 만큼 이해할 수 없을 정도로 많습니다. 그들 각각은 레이저 간섭계를 사용하는 수 킬로미터 규모의 중력파 탐지기를 가지고 있습니다. LIGO Scientific Collaboration(LSC) 은 미국 전역과 90개 이상의 대학 및 연구 기관에서 지원하는 14개국의 대학에서 온 1000명 이상의 과학자 그룹입니다. 약 250명의 학생들이 협력에 적극적으로 기여하고 있습니다. LIGO의 새로운 발견은 중력파가 지구를 통과할 때 공간과 시간에 미치는 미세한 교란을 측정함으로써 중력파 자체를 처음으로 관찰한 것입니다. 그것은 우주의 뒤틀린 측면, 즉 뒤틀린 시공간에서 만들어진 물체와 현상을 탐구하는 새로운 천체물리학의 영역을 열었습니다. + + +### 주요 목표 + +* [임무](https://www.ligo.caltech.edu/page/what-is-ligo)는 우주에서 가장 격렬하고 에너지가 넘치는 일부 과정에서 발생하는 중력파를 감지하는 것이지만, LIGO가 수집하는 데이터는 중력, 상대성 이론, 천체 물리학, 우주론, 입자 물리학 및 핵 물리학을 포함한 많은 물리학 영역에 광범위한 영향을 미칠 수 있습니다. +* 노이즈에서 신호를 식별하고, 관련 신호를 필터링하고, 관찰된 데이터의 유의성을 통계적으로 추정하기 위해 복잡한 수학을 포함하는 수치 상대성 계산을 통해 관측된 데이터를 고속 처리합니다. +* 이진/수치 상의 결과를 이해할 수 있도록 데이터 시각화를 진행합니다. + + + +### 도전 + +* **계산** + + 중력파는 매우 작은 효과를 생성하고 물질과의 상호 작용이 작기 때문에 감지하기 어렵습니다. LIGO의 모든 데이터를 처리하고 분석하려면 방대한 컴퓨팅 인프라가 필요합니다. 신호의 수십억 배에 해당하는 노이즈를 처리한 후에도 여전히 매우 복잡한 상대성 방정식과 엄청난 양의 데이터가 있어 계산상의 어려움이 있습니다. [바이너리 병합 분석에 필요한 O(10^7) CPU 시간](https://youtu.be/7mcHknWWzNI)은 6개의 전용 LIGO 클러스터에 퍼져 있습니다. + +* **데이터 홍수** + + 관찰 장치가 더 민감하고 신뢰할 수 있게 됨에 따라 데이터 폭증과 건초더미에서 바늘 찾기로 인해 제기되는 문제가 여러 배로 증가합니다. LIGO는 매일 테라바이트의 데이터를 생성합니다! 이 데이터를 이해하려면 탐지할 때마다 막대한 노력이 필요합니다. 예를 들어, LIGO가 수집하는 신호는 슈퍼컴퓨터에서 수십만 개의 가능한 중력파 서명 템플릿과 일치해야 합니다. + +* **시각화** + + 아인슈타인의 방정식을 슈퍼컴퓨터를 사용하여 풀 수 있을 만큼 충분히 이해하는 것과 관련된 장애물이 처리되면 다음 큰 과제는 데이터를 인간의 두뇌가 이해할 수 있도록 만드는 것입니다. 시뮬레이션 모델링과 신호 감지에는 효과적인 시각화 기술이 필요합니다. 시각화는 또한 이미징 및 시뮬레이션을 통해 더 많은 청중이 결과를 더 쉽게 이해할 수 있게 될 때까지 수치 상대성에 충분한 중요성을 부여하지 않은 순수 과학 애호가의 눈에 수치 상대성에 더 많은 신뢰성을 제공하는 역할을 합니다. 최신 실험 입력 및 통찰력을 사용하여 복잡한 계산 및 렌더링, 이미지 다시 렌더링 및 시뮬레이션의 속도는 이 영역의 연구자에게 도전이 되는 시간 소모적인 활동이 될 수 있습니다. + +{{< figure >}} +{{< figure src="/images/content_images/cs/gwpy-numpy-dep-graph.png" class="fig-center" alt="gwpy-numpy depgraph" caption="**Dependency graph showing how GwPy package depends on NumPy**" >}} +{{< figure src="/images/content_images/cs/numpy_gw_benefits.png" class="fig-center" alt="numpy를 통한 이익" caption="**활용된 주요 NumPy 기능**" >}} + +## 중력파 검출에서 NumPy의 역할 + +병합에서 방출되는 중력파는 슈퍼컴퓨터를 사용하는 무차별 대입 수치 상대성 이론을 제외하고는 어떤 기술로도 계산할 수 없습니다. LIGO가 수집하는 데이터의 양은 중력파 신호가 작은 만큼 이해할 수 없을 정도로 많습니다. + +Python용 표준 수치 분석 패키지인 NumPy는 LIGO의 GW 탐지 프로젝트 동안 수행되는 다양한 작업에 사용되는 소프트웨어에서 활용되었습니다. NumPy는 복잡한 수학 및 데이터 조작을 고속으로 해결하는 데 도움이 되었습니다. 몇 가지 예시를 들자면: + +* [신호 처리](https://www.uv.es/virgogroup/Denoising_ROF.html): 글리치 검출, [잡음 식별 및 데이터 결정](https://ep2016.europython.eu/media/conference/slides/pyhton-in-gravitational-waves-research-communities.pdf) (NumPy, scikit-learn, scipy, matplotlib, pandas, pyCharm) +* 데이터 수집: 어떤 데이터를 분석할 수 있을지 결정하고, 모래 속 바늘과 같이 미미한 신호가 있는지 파악 +* 통계적 분석: 관측 데이터의 통계적 유의성 추정, 모델을 비교하여 신호 매개변수(예: 별의 질량, 회전 속도, 거리 등)를 추정 +* 데이터의 시각화 + - 시계열 데이터 + - 스펙트로그램 +* 상관 분석 연산 +* [GwPy](https://gwpy.github.io/docs/stable/overview.html) 및 [PyCBC](https://pycbc.org)와 같은 중력파 데이터 분석에서 개발된 주요 [소프트웨어](https://github.com/lscsoft)는 후드 아래에서 NumPy 및 AstroPy를 사용하여 중력파 검출기에서 데이터를 연구하기 위한 유틸리티, 도구 및 방법에 객체 기반 인터페이스를 제공합니다. + +{{< figure >}} +src = '/images/content_images/cs/gwpy-numpy-dep-graph.png' alt = 'gwpy-numpy 종속성' title = 'GwPy 패키지가 어떻게 NumPy에 종속하는지를 나타내는 종속성 그래프' +{{< /figure >}} + +---- + +{{< figure >}} +src = '/images/content_images/cs/PyCBC-numpy-dep-graph.png' alt = 'PyCBC-numpy 종속성' title = 'PyCBC 패키지가 NumPy에 종속하는지를 나타내는 종속성 그래프' +{{< /figure >}} + +## 요약 + +중력파 검출을 통하여 연구자들은 완전히 예상치 못한 현상을 발견하게 됨으로써, 알려진 것 중 가장 난해한 천체물리학적 현상에 대하여 많은 사람들에게 새로운 통찰을 주었습니다. 숫자 계산 및 데이터 시각화는 과학자가 과학적 관찰에서 수집한 데이터에 대한 통찰력을 얻고 결과를 이해하는 데 도움이 되는 중요한 단계입니다. 계산은 복잡하며 실제 관찰 데이터 및 분석을 제공하는 컴퓨터 시뮬레이션을 사용하여 시각화하지 않는 한 사람이 이해할 수 없습니다. NumPy는 matplotlib, pandas 및 scikit-learn과 같은 다른 Python 패키지와 함께 [연구원](https://www.gw-openscience.org/events/GW150914/)이 복잡한 질문에 답하고 우주에 대한 이해의 새로운 지평을 발견할 수 있도록 합니다. + +{{< figure >}} +src = '/images/content_images/cs/numpy_gw_benefits.png' alt = 'numpy를 통한 이익' title = '활용된 주요 NumPy 기능' +{{< /figure >}} diff --git a/content/ko/citing-numpy.md b/content/ko/citing-numpy.md new file mode 100644 index 0000000000..cf1458e657 --- /dev/null +++ b/content/ko/citing-numpy.md @@ -0,0 +1,35 @@ +--- +title: NumPy 인용하기 +sidebar: false +--- + +진행한 연구에서 NumPy가 중요한 부분을 차지하고 있고 학술지에 출판한다면, 아래의 논문을 참조문헌에 써주시길 바랍니다. + +* Harris, C.R., Millman, K.J., van der Walt, S.J. et al. _Array programming with NumPy_. Nature 585, 357–362 (2020). DOI: [10.1038/s41586-020-2649-2](https://doi.org/10.1038/s41586-020-2649-2). ([링크](https://www.nature.com/articles/s41586-020-2649-2)). + +_BibTeX 형식:_ + + ``` +@Article{ harris2020array, + title = {Array programming with {NumPy}}, + author = {Charles R. Harris and K. Jarrod Millman and St{\'{e}}fan J. + van der Walt and Ralf Gommers and Pauli Virtanen and David + Cournapeau and Eric Wieser and Julian Taylor and Sebastian + Berg and Nathaniel J. Smith and Robert Kern and Matti Picus + and Stephan Hoyer and Marten H. van Kerkwijk and Matthew + Brett and Allan Haldane and Jaime Fern{\'{a}}ndez del + R{\'{i}}o and Mark Wiebe and Pearu Peterson and Pierre + G{\'{e}}rard-Marchant and Kevin Sheppard and Tyler Reddy and + Warren Weckesser and Hameer Abbasi and Christoph Gohlke and + Travis E. Oliphant}, + year = {2020}, + month = sep, + journal = {Nature}, + volume = {585}, + number = {7825}, + pages = {357--362}, + doi = {10.1038/s41586-020-2649-2}, + publisher = {Springer Science and Business Media {LLC}}, + url = {https://doi.org/10.1038/s41586-020-2649-2} +} +``` diff --git a/content/ko/code-of-conduct.md b/content/ko/code-of-conduct.md new file mode 100644 index 0000000000..b8dea1fd9d --- /dev/null +++ b/content/ko/code-of-conduct.md @@ -0,0 +1,83 @@ +--- +title: NumPy 이용 약관 +sidebar: false +aliases: + - /conduct.html +--- + +### 소개 + +이 이용 약관은 모든 공개 및 비공개 메일링 리스트, 이슈 트래커, 위키, 블로그, 트위터 및 커뮤니티에서 사용하는 기타 커뮤니케이션 채널을 포함하여 NumPy 프로젝트에서 관리하는 모든 공간에 적용됩니다. NumPy 프로젝트는 대면 이벤트를 조직하지 않지만 커뮤니티와 관련된 이벤트에는 이와 유사한 약관이 있어야 합니다. + +이 이용 약관은 NumPy 커뮤니티에 공식적으로 또는 비공식적으로 참여하거나 프로젝트 관련 활동에서, 특히 프로젝트를 대표할 때 어떤 역할에서든 프로젝트와의 관계를 주장하는 모든 사람이 준수해야 합니다. + +이 약관는 철저하거나 완전하지 않습니다. 협력적이고 공유되는 환경과 목표에 대한 우리의 공통적인 이해를 요약하는데 도움이 됩니다. 문자 그대로 뿐만 아니라, 정신에도 이 약관을 따르려 노력해 주시기 바랍니다. 이를 통해 주변 커뮤니티를 더욱 풍요롭게 하는 친근하고 생산적인 환경을 조성해주세요. + +### 구체적인 가이드라인 + +우리는 다음을 위해 노력합니다: + +1. 개방적이 되세요. 우리 커뮤니티에 누구든 참여할 수 있습니다. 민감한 내용이 아닌 이상, 프로젝트 관련 메시지에는 공개적인 의사소통 방법을 사용하는 것을 선호합니다. 이는 도움이나 프로젝트 지원과 관련된 메시지에도 적용됩니다. 공개적인 지원 요청은 질문에 대한 대답이 더욱 쉽게 이루어지며, 대답에 대한 우발적인 오류가 보다 쉽게 감지되고 수정될 수 있도록 도와줍니다. +2. 공감하며, 환영하며, 친절하며, 인내심을 가지세요. 우리는 갈등을 해결하기 위해 함께 노력하며, 선의의 의도를 기대합니다. 때로는 모두가 어떤 형태의 좌절을 경험할 수 있지만, 우리는 그것이 개인적인 공격으로 이어지지 않도록 허용하지 않습니다. 사람들이 불편하거나 위협받는 느낌을 가지는 커뮤니티는 생산적이지 않습니다. +3. 협력적이 되세요. 우리의 작업은 다른 사람들에 의해 사용될 것이며, 우리 역시 다른 사람들의 작업에 의존할 것입니다. 프로젝트의 이익을 위해 무언가를 만들 때, 우리는 다른 사람들에게 어떻게 작동하는지 설명할 의향이 있습니다. 이를 통해 다른 사람들이 그 작업을 발전시켜 더욱 좋은 결과물을 만들 수 있도록 도와줄 수 있습니다. 우리가 내린 어떤 결정도 사용자와 동료에게 영향을 미치며, 이러한 결과를 신중하게 고려하여 결정을 내리게 됩니다. +4. 호기심을 가져보세요. 아무도 모든 것을 알수는 없습니다! 미리 질문을 하면 나중에 많은 문제를 예방할 수 있으므로, 우리는 질문하는 것을 장려합니다. 단, 적절한 포럼으로 질문을 옮길 수도 있습니다. 우리는 답변을 빠르고 유용하게 제공하기 위해 최선을 다할 것입니다. +5. 사용하는 말에 신중하게 대해주세요. 우리는 의사소통 시 신중하고 예의 바른 태도를 유지하며, 우리 자신의 언어에 대한 책임을 집니다. 다른 사람들에게 친절하게 대하세요. 다른 참가자들을 모욕하거나 비하하지 마세요. 우리는 괴롭힘 및 다른 형태의 배제 행위를 용인하지 않습니다. 이러한 행위에는 다음과 같은 것들이 포함됩니다: + * 다른 사람을 향한 폭력적인 위협이나 언어. + * 성 차별, 인종 차별 또는 기타 차별적인 농담 및 언어. + * 성적으로 노골적이거나 폭력적인 자료 게시. + * 다른 사람의 개인 식별 정보를 게시(또는 게시하겠다고 위협)("신상 털기"). + * 발신자의 동의 없이 비공개 또는 비공개로 전송된 이메일과 같은 비공개 콘텐츠 또는 IRC 채널 기록과 같은 기록되지 않은 포럼을 공유. + * 개인적인 모욕, 특히 인종차별적이거나 성차별적인 용어를 사용하는 경우. + * 달갑지 않은 성적 관심. + * 과도한 욕설. 욕설은 삼가해주세요; 사람들은 욕에 대한 민감도가 크게 다릅니다. + * 타인에 대한 반복적인 괴롭힘. 일반적으로 누군가가 중지를 요청하면 중지합니다. + * 위의 행동을 옹호하거나 장려하는 행위. + +### 다양성 선언 + +NumPy 프로젝트는 모든 사람의 참여를 환영하고 장려합니다. 우리는 모두가 참여하는 것을 즐기는 커뮤니티가 되기 위해 최선을 다하고 있습니다. 항상 개인의 취향을 수용할 수는 없지만 모든 분들께 친절하게 대할 수 있도록 최선을 다하고 있습니다. + +당신이 자신을 어떻게 식별하든 다른 사람들이 당신을 어떻게 인식하든 상관없이 우리는 당신을 환영합니다. 어떤 목록도 포괄적이기를 바랄 수는 없지만 다음과 같은 다양성을 명시적으로 존중합니다: 연령, 문화, 민족, 유전자형, 성 정체성 또는 성 표현, 언어, 출신 국가, 신경형, 표현형, 정치적 신념, 직업, 인종, 종교, 성적 지향, 사회 경제적 지위, 하위 문화 및 기술적 능력 등을 이용 약관과 상충하지 않는 범위 안에서. + +우리는 모든 언어에 능통한 사람들을 환영하지만 NumPy 개발은 영어로 진행됩니다. + +NumPy 커뮤니티의 행동 표준은 위의 이용 약관에 자세히 설명되어 있습니다. 커뮤니티의 참가자는 모든 상호 작용에서 이러한 표준을 유지하고 다른 사람들도 그렇게 하도록 도와야 합니다(다음 섹션 참조). + +### 신고 지침 + +우리는 인터넷 통신이 명백하고 노골적인 남용에서 시작되거나 악화되는 것이 매우 흔한 일이라는 것을 알고 있습니다. 우리는 또한 때때로 사람들이 나쁜 하루를 보내거나 이 이용 약관의 일부 지침을 인식하지 못할 수 있음을 알고 있습니다. 본 약관 위반에 대응하는 방법을 결정할 때 이 점을 염두에 두십시오. + +명백히 의도적인 위반의 경우 이용 약관 위원회(아래 참조) 에 보고하십시오. 의도하지 않은 위반일 가능성이 있는 경우, 그 사람에게 회신하고 이 이용 약관을 지적할 수 있습니다(공개적이든 비공개적이든 가장 적절한 방식으로). 그렇게 하고 싶지 않다면 이용 약관 위원회에 직접 보고하거나 비밀리에 위원회에 조언을 구하십시오. + +numpy-conduct@googlegroups.com 으로 NumPy 윤리 강령 위원회에 문제를 보고할 수 있습니다. + +현재 위원회는 다음과 같이 구성되어 있습니다. + +* Stéfan van der Walt +* Melissa Weber Mendonça +* Rohit Goswami + +귀하의 보고서가 위원회 구성원과 관련이 있거나 보고서를 처리하는 데 이해 상충이 있다고 생각하는 경우 귀하의 보고서를 고려하지 않을 것입니다. 또는 어떤 이유로든 위원회에 보고하는 것이 불편하다고 느끼면 NumFOCUS 고위 직원에게 연락하셔도 됩니다. [conduct@numfocus.org](https://numfocus.org/code-of-conduct#persons-responsible). + +### 신고 해결 & 이용약관 강령 + +_이 섹션은 가장 중요한 사항을 요약하고 자세한 내용_은 [NumPy 행동 강령 - 보고에 대한 후속 조치 방법](/report-handling-manual)에서 찾을 수 있습니다. + +우리는 모든 불만을 조사하고 대응할 것입니다. NumPy 행동 강령 위원회와 NumPy 운영 위원회(관련된 경우) 는 신고자의 신원을 보호하고 신고 내용을 기밀로 취급합니다(신고자가 달리 동의하지 않는 한). + +심각하고 명백한 위반의 경우, 예: 개인적인 위협이나 폭력적, 성차별적 또는 인종차별적 언어를 사용하는 경우 NumPy 통신 채널에서 발신자를 즉시 연결 해제합니다. 자세한 내용은 설명서를 참조하십시오. + +본 행동 강령의 명백하고 명백한 위반과 관련되지 않은 경우 접수된 행동 강령 위반 보고서에 따라 조치를 취하는 절차는 다음과 같습니다. + +1. 보고서를 수신 확인, +2. 합리적인 토론/피드백, +3. 중재(피드백이 도움이 되지 않은 경우, 신고자와 피신고자 모두 동의하는 경우에만), +4. 행동 강령 위원회의 투명한 결정([해결책](/report-handling-manual/#resolutions) 참조) 을 통한 집행. + +위원회는 가능한 한 빨리, 최대 72시간 이내에 보고에 응답할 것입니다. + +### 끝내며 + +내용과 영감을 얻은 다음 문서의 배후에 있는 그룹에 감사드립니다. + +- [SciPy 이용 약관](https://docs.scipy.org/doc/scipy/dev/conduct/code_of_conduct.html) diff --git a/content/ko/community.md b/content/ko/community.md new file mode 100644 index 0000000000..7a12d27e8d --- /dev/null +++ b/content/ko/community.md @@ -0,0 +1,66 @@ +--- +title: 커뮤니티 +sidebar: false +--- + +NumPy는 다양한 [기여자](/teams/) 집단이 개발하며 커뮤니티에 의해 유지되는 오픈소스 프로젝트입니다. NumPy 운영진들은 개방적이며 포용적이고 긍정적인 커뮤니티를 만들기 위해 상당한 노력을 기울여오고 있습니다. 커뮤니티를 발전시키기 위한 다른 이용자들과의 상호작용에 대한 가이드라인은 [NumPy 이용약관](/code-of-conduct)을 통해 확인하실 수 있습니다. + +NumPy 커뮤니티에서는 배우고, 지식을 공유하고, 다른 사람들과 협력할 수 있는 여러 커뮤니케이션 채널을 제공합니다. + + +## 온라인으로 참여하기 + +NumPy 프로젝트 및 커뮤니티에 곧장 참여할 수 있는 방법들입니다. _사용자와 커뮤니티 회원이 사용 중 질문에 대하여 서로 도움을 주고받기를 권장한다는 것을 명심하십시오. [도움말](/gethelp)을 참고하세요._ + + +### [NumPy 메일링 리스트](https://mail.python.org/mailman/listinfo/numpy-discussion) + +이 리스트는 NumPy 신기능 추가, NumPy 로드맵 변경 등 모든 종류의 프로젝트 전체 의사 결정과 같은 장기적인 토론을 이끄는 주요 포럼이라 할 수 있습니다. 출시, 개발자 모임, 일반 모임, 컨퍼런스 강연과 같은 NumPy에 대한 공지도 이 리스트를 통해 받아볼 수 있습니다. + +리스트에 회신하려면 (다른 발신자에게 회신하기보다는) 하단의 게시물을 이용하십시오. 또, 자동 발신 메일에 회신하지 마십시오. 검색 가능한 아카이브는 [여기](https://mail.python.org/archives/list/numpy-discussion@python.org/)에서 이용할 수 있습니다. + +*** + +### [GitHub issue tracker](https://github.com/numpy/numpy/issues) + +- 버그 제보 (예: "`np.arange(3).shape` returns `(5,)`, when it should return `(3,)`"); +- 문서 관련 문제점 (예: "I found this section unclear"); +- 기능 요청 (예: "I would like to have a new interpolation method in `np.percentile`"). + +_GitHub은 보안 취약점을 제보하는 곳이 아님을 명심하십시오. NumPy의 보안 취약점을 발견한 것 같으시다면, [여기](https://tidelift.com/docs/security)에서 제보하십시오._ + +*** + +### [Slack](https://numpy-team.slack.com) + +NumPy에 _기여하는_ 방법에 대하여 질문하는 실시간 채팅방입니다. 여기는 비공개 공간으로, 공용 메일링 리스트나 GitHub에 질문 또는 아이디어를 올리는 것을 주저하는 사람들을 위한 곳입니다. [여기](https://numpy.org/devdocs/dev/index.html#contributing-to-numpy)에서 자세한 내용과 초대를 받는 방법을 알아보세요. + + +## 학술 그룹 및 모임 + +NumPy와 데이터 과학 및 과학적 컴퓨팅을 위한 Python 패키지의 생태계에 대해 자세히 알아보기 위하여, 지역 모임이나 학술 그룹을 찾고 싶다면 [PyData 모임](https://www.meetup.com/pro/pydata/) (150개 이상의 모임, 10만 명 이상의 회원) 사이트를 돌아보시는 것을 추천해 드립니다. + +NumPy에서도 가끔 자체 팀이나 관심 있는 기여자들을 위하여 직접 모임을 조직하기도 합니다. 보통 몇 달 전부터 미리 계획되며 [메일링 리스트](https://mail.python.org/mailman/listinfo/numpy-discussion) 및 [트위터](https://twitter.com/numpy_team)로 해당 사실을 알립니다. + + +## 컨퍼런스 + +NumPy 프로젝트에서는 자체 컨퍼런스를 추진하지 않습니다. 보통 NumPy 관리자나 기여자, 사용자들에게 가장 인기 있는 컨퍼런스는 SciPy나 PyData 쪽 컨퍼런스입니다. + +- [SciPy US](https://conference.scipy.org) +- [EuroSciPy](https://www.euroscipy.org) +- [SciPy Latin America](https://www.scipyla.org) +- [SciPy India](https://scipy.in) +- [SciPy Japan](https://conference.scipy.org) +- [PyData 컨퍼런스](https://pydata.org/event-schedule/) (세계 곳곳의 여러 나라에서 1년에 15~20개의 이벤트를 개최합니다) + +이런 컨퍼런스 대부분에는 NumPy를 배우는 튜토리얼의 날이나 NumPy 혹은 관련 오픈소스 프로젝트에 기여하는 방법을 배울 수 있는 장이 마련되어 있습니다. + + +## NumPy 커뮤니티에 가입하기 + +더욱 성장하기 위해, NumPy 프로젝트에서는 당신의 경험과 의욕을 필요로 합니다. 프로그래머가 아니라고요? 걱정하지 마세요! NumPy에 기여하는 방법에는 여러 가지가 있습니다. + +NumPy 기여자가 되는 데 관심이 있으시다면 [기여](/contribute) 페이지를 방문하시는 것을 추천해 드립니다. + +또한, 부담없이 커뮤니티 미팅에 참석 해주시길 바랍니다. 정확한 날짜들은 [행사달력](https://scientific-python.org/calendars/)을 확인해주세요. diff --git a/content/ko/config.yaml b/content/ko/config.yaml new file mode 100644 index 0000000000..9f0b9cfdc3 --- /dev/null +++ b/content/ko/config.yaml @@ -0,0 +1,140 @@ +languageName: 한국어 +params: + description: 왜 NumPy인가? 강력한 n차원 배열. 수치 컴퓨팅 도구. 상호운용성. 고성능. 오픈소스. + navbarlogo: + image: logo.svg + text: NumPy + link: /ko/ + hero: + #Main hero title + title: NumPy + #Hero subtitle (optional) + subtitle: Python으로 과학적 컴퓨팅을 하기 위한 기초 패키지 + #Button text + buttontext: "최신 릴리: NumPy 1.26. 모든 릴리즈 보기" + #Where the main hero button links to + buttonlink: "/ko/news/#releases" + #Hero image (from static/images/___) + image: logo.svg + shell: + title: 자리 표시자 + intro: + - + title: NumPy 써 보기 + text: 대화형 셸을 이용해 브라우저에서 NumPy를 사용해보세요 + docslink: 문서도 한 번 열람해보세요. + casestudies: + title: 사례 연구 + features: + - + title: 최초의 블랙홀 사진 + text: NumPy 및 NumPy에 의존하는 SciPy, Matplotlib와 같은 라이브러리가 사건의 지평선 망원경으로 최초의 블랙홀 사진을 생성할 수 있었던 방법 + img: /images/content_images/case_studies/blackhole.png + alttext: 최초의 블랙홀 사진. 검은 배경의 주황색 원입니다. + url: /case-studies/blackhole-image + - + title: 중력파 검출 + text: 1916년, 알베르트 아인슈타인이 중력파를 예측했습니다. LIGO 과학자들이 NumPy를 이용하여 이것이 존재함을 증명하기 100년 전이었습니다. + img: /images/content_images/case_studies/gravitional.png + alttext: 서로의 궤도를 도는 두 구체. 주위의 중력을 변화시키고 있습니다. + url: /case-studies/gw-discov + - + title: 스포츠 분석 + text: 크리켓 분석은 통계적 모델링과 예측 분석을 통해 선수와 팀의 성과를 개선하여 게임을 바꾸고 있습니다. NumPy는 이런 많은 분석을 가능하게 합니다. + img: /images/content_images/case_studies/sports.jpg + alttext: 크리켓 공이 녹지 위에 있습니다. + url: /ko/case-studies/cricket-analytics + - + title: 딥러닝을 통한 자세 추정 + text: DeepLabCut은 동물의 행동을 관찰하는 과학 연구의 속도를 개선하기 위해, NumPy를 사용하여 종이나 시간에 따른 운동 제어 방식을 잘 이해할 수 있도록 하였습니다. + img: /images/content_images/case_studies/deeplabcut.png + alttext: 치타 자세 분석 + url: /ko/case-studies/deeplabcut-dnn + tabs: + title: 생태계 + section5: false + navbar: + - + title: 설치 + url: /ko/install + - + title: 문서 + url: https://numpy.org/doc/stable + - + title: 배움 + url: /ko/learn + - + title: 커뮤니티 + url: /ko/community + - + title: NumPy 정보 + url: /ko/about + - + title: 소식 + url: /ko/news + - + title: 기여 + url: /ko/contribute + footer: + logo: logo.svg + socialmediatitle: "" + socialmedia: + - + link: https://github.com/numpy/numpy + icon: github + - + link: https://www.youtube.com/channel/UCguIL9NZ7ybWK5WQ53qbHng + icon: youtube + - + link: https://twitter.com/numpy_team + icon: twitter + quicklinks: + column1: + title: "" + links: + - + text: 설치 + link: /ko/install + - + text: 문서 + link: https://numpy.org/doc/stable + - + text: 배움 + link: /ko/learn + - + text: Numpy 인용 + link: /ko/citing-numpy + - + text: 로드맵 + link: https://numpy.org/neps/roadmap.html + column2: + links: + - + text: 정보 + link: /ko/about + - + text: 커뮤니티 + link: /ko/community + - + text: 사용자 설문조사 + link: /ko/user-surveys + - + text: 기여 + link: /ko/contribute + - + text: 이용약관 + link: /ko/code-of-conduct + column3: + links: + - + text: 도움받기 + link: /ko/gethelp + - + text: 이용약관 + link: /ko/terms + - + text: 개인정보처리방침 + link: /ko/privacy + - + text: 홍보 자료 + link: /ko/press-kit diff --git a/content/ko/contribute.md b/content/ko/contribute.md new file mode 100644 index 0000000000..51ce2bd4aa --- /dev/null +++ b/content/ko/contribute.md @@ -0,0 +1,66 @@ +--- +title: NumPy 프로젝트에 기여하기 +sidebar: false +--- + +NumPy 프로젝트에서는 당신의 경험과 의욕을 환영합니다! 당신의 선택지는 프로그래밍에만 국한되어 있지않습니다. 아래의 참여방법들을 확인하면 많은 부분에서 **당신의 도움**이 필요한 것을 알수있습니다. + +시작점을 찾기 힘들거나 재능을 어떻게 활용해야 할지 잘 모르겠다면, _물어보세요!_ [메일링 리스트](https://mail.python.org/mailman/listinfo/numpy-discussion)나 [GitHub](http://github.com/numpy/numpy) ([이슈](https://github.com/numpy/numpy/issues)를 생성하거나 관련 이슈에 답글을 다세요)에서 질문하시면 됩니다. + +앞서 소개드린 것들이 저희가 선호하는 연락 채널입니다. (오픈 소스는 원래 개방되어 있으니까요) 하지만 비공개적으로 대화를 나누고 싶으시다면, 을 통해 커뮤니티 코디네이터로 연락하시거나 [Slack](https://numpy-team.slack.com)을 이용하시면 됩니다. (초대를 받으시려면 을 쓰시면 됩니다). + +또한 저희는 격주마다 _커뮤니티 연락_을 합니다. 자세한 정보는 [메일링 리스트](https://mail.python.org/mailman/listinfo/numpy-discussion)로 알립니다. 당신의 참여를 매우 환영합니다. 오픈소스에 기여하는 게 처음이시라면 [이 도움말](https://opensource.guide/how-to-contribute/)을 읽어 보시는 것을 적극 권장합니다. + +저희 커뮤니티는 모두를 평등하게 대하고 모든 기여의 가치를 인정하려는 뜻을 품고 있습니다. 개방적이고 참여를 환영하는 분위기를 조성하기 위해 [이용약관](/code-of-conduct)을 만들었습니다. + +### 코드 작성 + +프로그래머 여러분, 이 [도움말](https://numpy.org/devdocs/dev/index.html#development-process-summary)에 어떻게 코드베이스에 기여하는지 설명되어 있습니다.
저희의 [유투브 채널](https://www.youtube.com/playlist?list=PLCK6zCrcN3GXBUUzDr9L4__LnXZVtaIzS)을 통해서 추가적인 정보를 확인 해주세요. + + +### Pull Request 리뷰 +프로젝트의 열린 풀 요청만 250개가 넘습니다. 즉 많은 잠재적 개선점과 오픈소스 기여자들이 피드백을 기다리고 있다는 것입니다. NumPy를 알고 있는 개발자라면, 코드베이스에 대해 잘 알지 못해도 기여할 수 있습니다. 아래와 같은 기여를 해 보십시오. +* 늘어지는 토론 요약 +* 문서의 풀 요청 심사 +* 제안된 변경 사항 테스트 + + +### 교육 자료 개발 + +NumPy의 [사용자 도움말](https://numpy.org/devdocs)은 현재 대규모로 재구성되고 있습니다. 현재 새로운 튜토리얼, 방법, 심층적 설명이 필요하고, 사이트의 구조를 다시 짜야 합니다. 글을 쓰는 사람에게만 기회가 주어지는 것은 아닙니다. 코드 예제와 노트북, 동영상 등을 통한 기여도 환영합니다. [NEP 44 — NumPy 문서의 재구성](https://numpy.org/neps/nep-0044-restructuring-numpy-docs.html)에 사이트 재구성에 대하여 자세한 내용이 설명되어 있습니다. + + +### 이슈 확인 + +[NumPy 이슈 트래커](https://github.com/numpy/numpy/issues)에는 _정말 많은_ 이슈들이 현재 열린 상태로 있습니다. 일부는 더 이상 유효하지 않은 이슈고, 일부는 우선 순위를 지정해야 하며, 일부는 새로운 기여자들이 볼 만한 좋은 이슈가 될 수 있을 것입니다. 아래와 같은 기여를 해 보십시오. + +* 오래된 버그가 현재도 남아 있는지 확인 +* 중복된 이슈를 찾아 하나로 묶기 +* 이슈를 재현하는 코드를 추가 +* 이슈를 올바르게 라벨링 (이 작업에는 심사 권한이 필요합니다. 필요한 경우 요청하십시오) + +한 번 참여해 보시길 바랍니다. + + +### 웹사이트 개발 + +사이트를 막 뜯어 고친 상태이지만, 아직 끝이라기엔 멀었습니다. 웹 개발을 좋아하신다면, [여기](https://github.com/numpy/numpy.org/issues?q=is%3Aissue+is%3Aopen+label%3Adesign)에서 저희가 이루지 못했던 사항의 목록을 볼 수 있습니다. 자신만의 아이디어를 마음껏 공유해 주십시오. + + +### 그래픽 디자인 + +그래픽 디자이너분들이 할 수 있는 기여의 목록을 여기에 열거하는 건 어렵습니다. 저희 문서에는 일러스트가 많이 부족합니다. 성장하는 사이트에는 이미지가 필요하기 때문에, 기여할 수 있는 기회가 많을 것입니다. + + +### 웹사이트 번역 + +사용자가 모국어로 NumPy를 이용할 수 있도록 [numpy.org](https://numpy.org)의 여러 번역을 계획하고 있습니다. 이를 위해서는 자원봉사자분들의 통역이 필요합니다. 자세한 내용은 [여기](https://numpy.org/neps/nep-0028-website-redesign.html#translation-multilingual-i18n)를 참고하십시오. [이 GitHub 이슈](https://github.com/numpy/numpy.org/issues/55)에 댓글을 달아 번역에 참여하십시오. + + +### 커뮤니티 조직 및 확산 + +우리는 커뮤니티 연락처를 통해 작업물을 더 널리 공유하고 미흡한 부분을 배워 나갑니다. 우리는 [Twitter](https://twitter.com/numpy_team) 계정, NumPy [코드 스프린트](https://scisprints.github.io/) 개최, 뉴스레터 발행, 그리고 아마 블로그 등을 통해서 더 많은 사람들이 커뮤니티에 참여하기를 간절히 바라고 있습니다. + +### 모금 + +NumPy는 오랜 기간 동안 자원봉사의 형태로 유지되었으나, 그 중요성이 커짐에 따라 안정성 및 성장을 보장하려면 경제적 지원이 필요함이 분명해졌습니다. 이런 지원이 얼마나 큰 차이를 만들어 냈는지 [SciPy'19 강연](https://www.youtube.com/watch?v=dBTJD_FDVjU)에서 확인하실 수 있습니다. 모든 비영리 조직과 마찬가지로 저희는 지속적으로 보조금, 후원 및 기타 종류의 지원을 끊임없이 찾고 있습니다. 모금을 받을 아이디어가 몇 개 있지만 당연히 더 많은 자금을 받게 된다면 좋을 것입니다. 모금도 정말 희귀한 능력 중 하나입니다 - 도움을 주신다면 감사드리겠습니다. diff --git a/content/ko/gethelp.md b/content/ko/gethelp.md new file mode 100644 index 0000000000..f216b532d6 --- /dev/null +++ b/content/ko/gethelp.md @@ -0,0 +1,20 @@ +--- +title: 도움 구하기 +sidebar: false +--- + +**개발 이슈:** NumPy 개발 관련 문제(버그 제보 등)의 경우, [커뮤니티](/community)를 방문해주시기 바랍니다. + +**사용자 질문:** 도움을 받을 수 있는 가장 좋은 방법은 링크된 사이트에 질문을 게시하는 것입니다. [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) 또는 [Reddit](https://www.reddit.com/r/Numpy/). 저희가 직접 이런 사이트들을 주시하거나 질문에 대해 답해드리고 싶지만, 그러기에는 질문의 양이 너무 많습니다! + +### [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) + +"How do I do X in NumPy?”와 같이 사용 중 질문을 올리는 포럼입니다. [`#numpy` 태그를 사용](https://stackoverflow.com/help/tagging)해주세요. + +*** + +### [Reddit](https://www.reddit.com/r/Numpy/) + +사용 중 질문을 올리는 또다른 포럼입니다. + +*** diff --git a/content/ko/history.md b/content/ko/history.md new file mode 100644 index 0000000000..015495f328 --- /dev/null +++ b/content/ko/history.md @@ -0,0 +1,21 @@ +--- +title: NumPy의 역사 +sidebar: false +--- + +NumPy는 배열 데이터 구조와 이에 대한 빠른 수치적 루틴을 제공하는 Python의 기초적인 라이브러리입니다. 처음 시작했을 때는 라이브러리를 개발할 자금이 거의 없었고, 주로 컴퓨터 공학 교육을 받지 못했고, 교수의 승인조차 받지 못한 대학원생이 이를 제작해나갔습니다. 소규모 "불량" 학생 프로그래머 집단이 이미 잘 정립되었으며 엄청난 자본과 많은 우수한 기술자들이 뒷받침하는 연구 소프트웨어 생태계를 뒤바꾼다고 상상해보세요. 정말 터무니없는 일입니다. 그러나 완전 개방형 도구 속에 감추어졌던 철학적 동기들이, 친근하고 들떴으며 특별한 목표를 가진 공동체와 결합되어, 장기적으로 유의미한 것이 드러났습니다. 오늘날 NumPy는 전 세계의 과학자, 기술자 및 기타 많은 전문가들의 신뢰를 받고 있습니다. 예를 들어, 중력파 분석에 사용되며 출시된 스크립트는 NumPy 패키지를 가져 왔고, M87 블랙홀 시각화 프로젝트에서는 NumPy를 직접 인용하였습니다. + +NumPy 및 관련 라이브러리의 개발 단계에 대한 자세한 설명은 [arxiv.org](https://arxiv.org/abs/1907.10121)를 참고하십시오. + +원본 Numeric 및 Numarray 라이브러리의 사본을 얻으려면 아래 링크를 들어가십시오. + +[*Numeric* 다운로드 페이지](https://sourceforge.net/projects/numpy/files/Old%20Numeric/)* + +[*Numarray* 다운로드 페이지](https://sourceforge.net/projects/numpy/files/Old%20Numarray/)* + +*이런 오래된 배열 패키지는 더 이상 지원되지 않으며, 배열 관련 기능을 이용하기 위해서는 NumPy를 사용하거나 NumPy 라이브러리를 활용하기 위해서는 기존 코드를 리팩토링하는 것이 좋습니다. + +### 역사적 문서 + +[*`Numeric'* 메뉴얼 다운로드](static/numeric-manual.pdf) + diff --git a/content/ko/install.md b/content/ko/install.md new file mode 100644 index 0000000000..c3128cee25 --- /dev/null +++ b/content/ko/install.md @@ -0,0 +1,126 @@ +--- +title: NumPy 설치 +sidebar: false +--- + +NumPy를 설치하는 유일한 선행 조건은 Python 자체입니다. 만약 아직 Python을 설치하지 않았고 가장 간단한 방법으로 시작하려면, [Anaconda 배포판](https://www.anaconda.com/distribution) 을 사용하길 권장합니다. 이 배포판에는 Python 와 NumPy 및 과학 계산 및 데이터 사이언스에 자주 사용되는 다른 패키지들이 포함되어 있습니다. + +NumPy 는 `conda` 나 `pip` 를 통해서 사용하여 설치할 수 있고, 또한 macOS 및 Linux의 패키지 관리자나 [원본 코드](https://numpy.org/devdocs/user/building.html) 를 이용하여 설치할 수 있습니다. 더 자세한 설명은 아래의 [Python 및 NumPy 설치 가이드](#python-numpy-install-guide)를 확인해주세요. + +**CONDA** + +`conda`를 사용한다면, NumPy를 `defaults` 또는 `conda-forge` 채널에서 설치할 수 있습니다: + +```bash +# 가장 좋은 방법은 기본 환경 대신에 새로운 환경을 이용하는 것입니다 +conda create -n my-env +conda activate my-env +# my-env 에 conda-forge 채널을 더해줍니다 +conda config --env --add channels conda-forge +# my-env 에 NumPy 를 설치합니다 +conda install numpy +``` + +**PIP** + +`pip`를 사용한다면, NumPy를 다음과 같이 설치할 수 있습니다: + +```bash +pip install numpy +``` +또한 `pip`를 사용할 때, 가상 환경을 사용하는 것을 추천합니다. 가상 환경을 사용하는 이유는 [재현 가능한 설치방법들](#reproducible-installs)을 참조해주세요. 가상 환경 사용에 대한 자세한 내용은 [이 가이드](https://dev.to/bowmanjd/python-tools-for-managing-virtual-environments-3bko#howto)에서 확인하실 수 있습니다. + + + + +# Python 및 NumPy 설치 가이드 + +Python에서 패키지를 설치하고 관리하는 것은 복잡한 작업이기에, 대부분의 작업에 대해 다양한 대안적인 해결책이 있습니다. 이 가이드는 독자분들에게 가장 좋은 (또는 가장 인기 있는) 해결책을 이해시키고 명확한 권장 사항을 제공하려고 노력합니다. 이 가이드는 일반적인 운영 체제와 하드웨어에서 Python, NumPy 및 PyData (또는 숫자 계산) 사용자에 중점을 둡니다. + +## 권장 사항 + +사용자의 경험과 사용하는 운영체제를 기준으로 추천하는 방식을 먼저 이야기 하겠습니다. 만약 당신이 초심자 또는 숙련자범위에 속해있다면, 간단하게 설치하고 싶다면 초심자로, 추후에 작업을 위해서 보다 구체적인 연습을 하고 싶다면 숙련자 자료를 참고하십시오. + +### 초심자 유저 + +Windows, macOS, Linux 등 일반적인 운영체제: + +- [Anaconda](https://www.anaconda.com/distribution/) 를 설치하십시오.(당신이 필요로 하는 패키지를 설치하고, 아래에 언급될 다양한 도구들을 제공합니다.) +- 코드를 작성하거나 실행할 때, 데이터를 분석하거나 대화형으로 코드를 작업하는 경우에는 [JupyterLab](https://jupyterlab.readthedocs.io/en/stable/index.html) 의 notebooks를 사용하십시오. 그리고 코드를 작성하거나 패키지를 작성할 때는 [Spyder](https://www.spyder-ide.org/)나 [Visual Studio Code](https://code.visualstudio.com/)를 사용하십시오. +- 패키지를 관리하거나 JupyterLab, Spyder, Visual Studio Code 를 사용하는 경우 [Anaconda Navigator](https://docs.anaconda.com/anaconda/navigator/)를 사용하십시오. + + +### 숙련자 유저 + +#### Conda + +- [Miniforge](https://github.com/conda-forge/miniforge)를 설치하십시오. +- `base` 라는 이름의 콘다 가상환경은 최소 상태를 유지하고, [콘다 가상환경](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html)을 만들어서, 해당 가상환경에 필요한 패키지를 설치하십시오. + +#### Pip/PyPI를 활용하는 경우: + +개인적인 선호나 아래의 conda 와 pip의 차이점을 설명하는 글을 읽은 유저나 또는 pip/PyPI기반의 설치 방법을 선호하는 경우 참고하십시오. +- [python.org](https://www.python.org/downloads/), [Homebrew](https://brew.sh/), or Linux package manager를 활용해서 Python을 설치하십시오. +- Conda와 동일한 수준의 가상환경 관리와 패키지 의존성을 해결을 도와주는 [Poetry](https://python-poetry.org/)를 유지관리 도구로 사용하십시오. + + +## Python 패키지 관리 + +패키지 관리는 아주 중요하기 때문에, 사용할 수 있는 도구들이 많습니다. 웹 및 범용 Python 개발을 위해 Pip뿐만 아니라 [다양한 도구](https://packaging.python.org/guides/tool-recommendations/)들이 있습니다. 고성능 컴퓨터 (HPC) 를 사용하는 경우 [Spack](https://github.com/spack/spack)를 사용하는 것을 추천합니다. 대부분 Numpy를 사용하는 유저는, [conda](https://conda.io/en/latest/) 와 [pip](https://pip.pypa.io/en/stable/)를 가장 많이 사용합니다. + + +### Pip & conda + +Python 패키지를 설치하고 관리하는 주요 툴은 `pip` 과 `conda` 입니다. 그 도구들의 기능은 부분적으로 겹칩지만 (e.g. both can install `numpy`), 같이 쓰일 수도 있습니다. 곧 pip와 conda의 차이점에 대해서 논의해볼 것입니다. - 패키지 관리를 효율적으로 하기 위해서는 차이를 이해하는게 중요합니다. + +첫번째 차이점은, conda는 cross-language 를 지원하고, Python을 설치할 수 도 있지만, pip는 특정 Python에만 패키지를 설치하고 관리할 수 있다는 것 입니다. 또한 conda는 non-Python 라이브러리나 도구들을 설치할 수 있지만 (e.g. compilers, CUDA, HDF5), pip는 Python이 필요하기 때문에 설치할 수 없습니다. + +두번째 차이점은 pip는 Python Packaging Index(PyPI) 로 부터 패키지를 다운받아 설치하지만, conda는 conda 만의 채널로 설치합니다 (일반적으로 "defaults" or "conda-forge" 채널을 사용합니다). PyPI 가 가장 큰 패키지 저장소입니다만, 많은 사람들이 사용하는 패키지는 conda에서도 설치할 수 있습니다. + +세번째 차이점은 conda는 환경이나 패키지간 의존성을 해결하기 위한 해키지 관리 도구를 제공하지만, pip는 그를 위해서 (아주 많은) 추가적인 도구들이 필요하다는 것 입니다. + + + +### 재현 가능한 설치방법들 + +라이브러리가 업데이트되면, 코드의 실행 결과가 바뀌거나, 코드가 완전히 손상될 수 있습니다. 그러므로 사용중인 패키지 및 버전을 재구성할 수 있도록 하는 것이 중요합니다. 권장되는 방법으로는 + +1. 작업 중인 프로젝트마다 다른 환경을 이용하고, +2. 각각 자체 메타 데이터 형식이 있는 패키지 설치 프로그램을 통해 패키지 이름과 버전을 기록해둡니다. + - Conda: [conda 환경들 과 environment.yml](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html) + - Pip: [가상 환경들](https://docs.python.org/3/tutorial/venv.html) 과 [requirements.txt](https://pip.readthedocs.io/en/latest/user_guide/#requirements-files) + - Poetry: [가상 환경들 과 pyproject.toml](https://python-poetry.org/docs/basic-usage/) + + + +## NumPy 패키지 & 고속 선형 대수 라이브러리 + +NumPy는 다른 Python 패키지에 의존하지 않습니다. 그러나 고속 선형 대수 라이브러리, 일반적으로 [Inter MKL](https://software.intel.com/en-us/mkl) 또는 [OpenBLAS](https://www.openblas.net/)에 의존하고 있습니다. 사용자는 이를 설치하지 않아도 됩니다 (NumPy 설치 중 설치됩니다). 고급 사용자의 경우 사용한 BLAS가 디스크의 성능, 동작 및 크기에 영향을 끼칠 수 있기 때문에 세부 정보가 필요할 수도 있습니다. + +- Pip가 설치하는 PyPI의 휠 파일에 있는 NumPy의 경우는 OpenBLAS로 빌드되었습니다. OpenBLAS 라이브러리는 휠 파일에 포함되어 있습니다. 이는 휠 파일의 크기를 더 크게 만들고, 사용자가 (예를 들어) SciPy도 설치하게 되면 디스크에 2개의 OpenBLAS 사본을 가지게 됩니다. + +- Conda의 기본 채널 내 NumPy는 Interl MKL로 빌드되었습니다. MKL은 NumPy를 설치할 때 사용자의 환경에 같이 설치되는 분할 패키지입니다. + +- Conda-forge 채널 내 NumPy는 더미 "BLAS" 패키지로 빌드되었습니다. 사용자가 conda-forge에서 NumPy를 설치할 때 해당 BLAS 패키지가 실제 라이브러리와 함께 설치됩니다. 기본값은 OpenBLAS이나, (기본 채널에서는) MKL이 될 수도 있고, 심지어 [BLIS](https://github.com/flame/blis)나 Reference BLAS가 될 수도 있습니다. + +- MKL 패키지가 OpenBLAS에 비해 더욱 큽니다. OpenBLAS가 30MB를 차지하는 반면, MKL 쪽은 700MB에 달하는 디스크 공간을 차지합니다. + +- 보통 MKL이 OpenBLAS보다 더 빠르고 안정적입니다. + +설치 크기, 성능 및 안정성을 제쳐 두더라도, 고려할 사항이 2가지 더 있습니다. + +- Intel MKL은 오픈소스가 아닙니다. 일반적으로 사용할 때는 문제가 되지 않지만, 사용자가 NumPy로 빌드한 애플리케이션을 재배포하는 경우 문제가 될 수 있습니다. +- MKL과 OpenBLAS 모두 `np.dot`과 같이 함수를 호출하는 데 다중 스레드를 사용하며, 스레드의 수는 빌드 시간 설정과 환경 변수에 의해 결정됩니다. 보통은 모든 CPU 코어가 사용됩니다. 이로 인하여 예기치 않은 일이 발생할 수 있습니다. NumPy 자체적으로는 어떤 함수 호출도 병렬화하지 않습니다. 일반적으로 더 나은 성능을 제공해주지만, 예를 들어 Dask, scikit-learn 또는 멀티프로세싱과 함께 다른 수준의 병렬화를 사용하는 경우 좋지 않은 결과를 초래할 수 있습니다. + + +## 트러블슈팅 + +아래와 같은 응답과 함께 설치에 실패한다면, [Troubleshooting ImportError](https://numpy.org/doc/stable/user/troubleshooting-importerror.html)를 참고하시기 바랍니다. + +``` +IMPORTANT: PLEASE READ THIS FOR ADVICE ON HOW TO SOLVE THIS ISSUE! + +Importing the numpy c-extensions failed. This error can happen for +different reasons, often due to issues with your setup. +``` + diff --git a/content/ko/learn.md b/content/ko/learn.md new file mode 100644 index 0000000000..3f32c000f3 --- /dev/null +++ b/content/ko/learn.md @@ -0,0 +1,76 @@ +--- +title: 배우기 +sidebar: false +--- + +**공식 NumPy 문서**는 [numpy.org/doc/stable](https://numpy.org/doc/stable)에 있습니다. + +*** + +아래는 NumPy 기여자들이 개발하고 커뮤니티에서 승인한 자기 학습 및 교육 자원들을 선별한 컬렉션입니다. + +## 초심자 + +내외적으로 NumPy에 대한 정보가 많이 있습니다. 처음 시작하는 경우 다음 자료들을 강력히 추천합니다: + + **튜토리얼** + +* [NumPy 빠른 시작 튜토리얼](https://numpy.org/devdocs/user/quickstart.html) +* [NumPy 튜토리얼](https://numpy.org/numpy-tutorials) - NumPy 문서 팀에서 개발 및 유지보수하는 Jupyter 노트북 형식의 튜토리얼 및 교육 자료 모음입니다. 만약 추가하고 싶은 내용이 생기는 경우 [numpy-tutorials repository on GitHub](https://github.com/numpy/numpy-tutorials)를 확인해 주십시오. +* [NumPy Illustrated: The Visual Guide to NumPy - *Lev Maximov 저*](https://betterprogramming.pub/3b1d4976de1d?sk=57b908a77aa44075a49293fa1631dd9b) +* [SciPy Lectures](https://scipy-lectures.org/) - 여기서는 NumPy를 다루는 것 외에도 Python 생태계에 대하여 광범위한 소개를 볼 수 있습니다. +* [NumPy: the absolute basics for beginners](https://numpy.org/devdocs/user/absolute_beginners.html) +* [From Python to NumPy - *Nicolas P. Rougier 저*](https://github.com/rougier/numpy-tutorial) +* [Stanford CS231 - *Justin Johnson 저*](http://cs231n.github.io/python-numpy-tutorial/) +* [NumPy User Guide](https://numpy.org/devdocs) + + **도서** + +* [Guide to NumPy - *Travis E. Oliphant 저*](http://web.mit.edu/dvp/Public/numpybook.pdf) 이건 2006년의 무료 버전 초판입니다. 최근 판(2015)은 [여기에서](https://www.barnesandnoble.com/w/guide-to-numpy-travis-e-oliphant-phd/1122853007) 볼 수 있습니다. +* [From Python to NumPy - *Nicolas P. Rougier 저*](https://www.labri.fr/perso/nrougier/from-python-to-numpy/) +* [Elegant SciPy - ](https://www.amazon.com/Elegant-SciPy-Art-Scientific-Python/dp/1491922877) *Juan Nunez-Iglesias, Stefan van der Walt, Harriet Dashnow 저* + +"Python+SciPy" 주제에 관한 [Goodreads 목록](https://www.goodreads.com/shelf/show/python-scipy)도 확인해보시기를 권장합니다. 거기서 대부분의 책은 "SciPy 생태계"에 관한 것이며, 이 생태계의 핵심에는 NumPy가 포함되어 있습니다. + + **영상** + +* [Introduction to Numerical Computing with NumPy - ](http://youtu.be/ZB7BZMhfPgk) *Alex Chabot-Leclerc 저* + +*** + +## 숙련자 + +Indexing, Splitting, Stacking, 선형대수 등과 같은 NumPy의 개념을 더 잘 이해하러면 이 고급 자료들을 참조 해보세요. + + **튜토리얼** + +* [100 NumPy Exercises](http://www.labri.fr/perso/nrougier/teaching/numpy.100/index.html) *by Nicolas P. Rougier* +* [An Introduction to NumPy and Scipy](https://engineering.ucsb.edu/~shell/che210d/numpy.pdf) *by M. Scott Shell* +* [Numpy Medkits](http://mentat.za.net/numpy/numpy_advanced_slides/) *by Stéfan van der Walt* +* [NumPy 튜토리얼](https://numpy.org/numpy-tutorials) - NumPy 문서 팀에서 개발 및 유지보수하는 Jupyter 노트북 형식의 튜토리얼 및 교육 자료 모음입니다. 만약 추가하고 싶은 내용이 생기는 경우 [numpy-tutorials repository on GitHub](https://github.com/numpy/numpy-tutorials)를 확인해 주십시오. + + **도서** + +* [Python Data Science Handbook](https://www.amazon.com/Python-Data-Science-Handbook-Essential/dp/1491912057) *by Jake Vanderplas* +* [Python for Data Analysis](https://www.amazon.com/Python-Data-Analysis-Wrangling-IPython/dp/1491957662) *by Wes McKinney* +* [Numerical Python: Scientific Computing and Data Science Applications with Numpy, SciPy, and Matplotlib](https://www.amazon.com/Numerical-Python-Scientific-Applications-Matplotlib/dp/1484242459) *by Robert Johansson* + + **영상** + +* [Advanced NumPy - broadcasting rules, strides, and advanced indexing](https://www.youtube.com/watch?v=cYugp9IN1-Q) *by Juan Nunez-Iglesias* + +*** + +## NumPy 이야기 + +* [The Future of NumPy Indexing](https://www.youtube.com/watch?v=o0EacbIbf58) *by Jaime Fernández* (2016) +* [Evolution of Array Computing in Python](https://www.youtube.com/watch?v=HVLPJnvInzM&t=10s) *by Ralf Gommers* (2019) +* [NumPy: what has changed and what is going to change?](https://www.youtube.com/watch?v=YFLVQFjRmPY) *by Matti Picus* (2019) +* [Inside NumPy](https://www.youtube.com/watch?v=dBTJD_FDVjU) *by Ralf Gommers, Sebastian Berg, Matti Picus, Tyler Reddy, Stefan van der Walt, Charles Harris* (2019) +* [Brief Review of Array Computing in Python](https://www.youtube.com/watch?v=f176j2g2eNc) *by Travis Oliphant* (2019) + +*** + +## NumPy 인용하기 + +만약 당신의 연구에서 NumPy가 중요한 역할을 수행하였고 학술 간행물에서 출판하기 위해서는 [이 인용 정보](/citing-numpy)를 참조하세요. diff --git a/content/ko/news.md b/content/ko/news.md new file mode 100644 index 0000000000..6f52f07394 --- /dev/null +++ b/content/ko/news.md @@ -0,0 +1,279 @@ +--- +title: 소식 +sidebar: false +newsHeader: "NumPy 2.0 출시일: 6월 16일" +date: 2023-09-16 +--- + +### NumPy 2.0 출시일: 6월 16일 + +_2024년 5월 23일_ -- NumPy 2.0이 2024년 6월 16일에 출시할 예정이라는 소식을 발표하게 되어 기쁩니다. 이 릴리즈를 제작하는 데 1년이 넘게 걸렸고, 2006년 이후 첫 번째 메인 릴리즈입니다. 중요한 건 많은 기능과 성능 개선 외에도, ABI와 Python, C API에 대한 **획기적인 변화**를 이뤄냈다는 것입니다. 아마 의존하는 패키지와 최종 사용자의 코드를 수정해야 할 겁니다. 가능하다면 코드가 `2.0.0rc2`에서 잘 작동하는지 검증해 주세요. **자세한 내용은 아래 항목들을 확인해 주세요.** + +- [NumPy 2.0 이주 가이드](https://numpy.org/devdocs/numpy_2_0_migration_guide.html) +- [2.0.0 릴리즈 노트](https://numpy.org/devdocs/release/2.0.0-notes.html) +- 상태 업데이트 공지용 이슈: [numpy#24300](https://github.com/numpy/numpy/issues/24300) + + +### NumPy 1.26.0 출시 +_2023년 12월 19_ -- NumFOCUS에서 연말 캠페인 기간 동안 PyCharm과 협력해 최초 PyCharm 이용자의 라이선스를 30% 할인된 가격에 제공했습니다. 지금부터 2023년 12월 23일까지 PyCharm 구매로 발생한 모든 수익은 NumFOCUS 프로그램으로 직접 전달됩니다. + +구매를 추적할 수 있는 고유 URL을 이용하거나: https://lp.jetbrains.com/support-data-science/ 쿠폰 코드를 사용하세요: ISUPPORTDATASCIENCE  + +### NumPy 1.26.0 출시 + +_2023년 12월 16일_ -- [NumPy 1.26.0](https://numpy.org/doc/stable/release/1.26.0-notes.html)이 출시되었습니다. 주요 기능들은 다음과 같습니다: + +* 파이썬 3.12.0 지원 +* Cython 3.0.0 호환 +* Meson 빌드 시스템 사용 +* 업데이트된 SIMD 지원 +* f2py 수정, meson 및 bind(x) 지원 +* 업데이트된 Accelerate BLAS/LAPACK 라이브러리 지원 + +NumPy 1.26.0 릴리스는 Meson 빌드 시스템으로의 전환과 Cython 3.0.0 지원을 표시하는 1.25.x 시리즈의 연장입니다. 총 20명의 사람들이 이 릴리스에 기여하였으며 59개의 풀 리퀘스트가 병합되었습니다. + +본 릴리즈에서 지원하는 Python 버전은 3.3.9-3.12입니다. + +### numpy.org은 이제 일본어와 포르투갈어로도 이용 가능합니다. + +_2023년 8월 2일_ - numpy.org은 이제 추가로 일본어와 포르투갈어로 이용 가능합니다. 이는 다음의 헌신적인 자원봉사자들의 노력 없이는 가능하지 않았을 것입니다: + +_포르투갈어_ +* Melissa Weber Mendonça (melissawm) +* Ricardo Prins (ricardoprins) +* Getúlio Silva (getuliosilva) +* Julio Batista Silva (jbsilva) +* Alexandre de Siqueira (alexdesiqueira) +* Alexandre B A Villares (villares) +* Vini Salazar (vinisalazar) + +_일본어_ +* Atsushi Sakai (AtsushiSakai) +* KKunai +* Tom Kelly (TomKellyGenetics) +* Yuji Kanagawa (kngwyu) +* Tetsuo Koyama (tkoyama010) + +번역 인프라에 대한 작업은 CZI로부터의 자금 지원을 받아 진행되었습니다. + +나아가서 NumPy 웹사이트가 더 많은 언어로 번역되기를 바랍니다. 도움을 주시려면 다음 Slack 링크를 통해 NumPy Translations Team 에 연락을 주십시오: https://join.slack.com/t/numpy-team/shared_invite/zt-1gokbq56s-bvEpo10Ef7aHbVtVFeZv2w. (#translations 채널을 add 해주세요) 또한 과학적 파이썬 생태계 전반에서 문서 및 교육 콘텐츠를 지역화하는데 참여할 Translations Team을 구축하고 있습니다. 이에 흥미를 느낀다면 Scientific Python Discord에서 함께해 주세요: https://discord.gg/khWtqY6RKr. (#translation 채널을 찾아보세요) + +### NumPy 1.25.0 출시 + +_2023년 6월 17일_ -- 이제 [NumPy 1.25.0](https://numpy.org/doc/stable/release/1.25.0-notes.html)을 이용할 수 있습니다. 주요 기능들은 다음과 같습니다: + +* MUSL 지원, 이제 MUSL Wheel도 배포됩니다. +* Fujitsu C/C++ 컴파일러 지원 +* Einsum에서 객체 배열 지원 +* Inplace 행렬 곱셈 (`@=`) 지원 + +NumPy 1.25.0 릴리스에서는 dtype의 처리 및 형변환을 개선하고 실행 속도를 높이는 작업, 문서를 보다 명료하게 다듬는 작업을 계속하고 있습니다. 미래의 NumPy 2.0.0을 위한 준비 작업도 있었는데, 이로 인해 수많은 기능들이 지원 종료 예정에 새로 포함되거나 완전히 만료되었습니다. + +총 148명의 사람들이 이 릴리스에 기여하였으며 530개의 풀 리퀘스트가 병합되었습니다. + +본 릴리즈에서 지원하는 Python 버전은 3.9-3.11입니다. + +### 포용적인 문화 조성: 참여 요청 + +_2023년 5월 10일_ -- 포용적인 문화 조성: 참여 요청 + +다양성과 포용성의 측면에서 우리는 어떻게 더 나아질 수 있을까요? [여기](https://contributor-experience.org/docs/posts/dei-report/)에서 보고서를 읽고 함께 참여하는 방법을 알아보세요. + +### NumPy 문서 팀 리더 변경 + +_2023년 1월 6일_ –- Mukulika Pahari, Ross Barnowski가 Melissa Mendonça를 대신해 새 NumPy 문서 팀 리더로 임명되었습니다. NumPy 공식 문서와 교육 자료에 기여한 Melissa와 한 걸음 더 나아간 Mukulika, Ross에게 감사를 표합니다. + +### NumPy 1.24.0 출시 + +_2022년 12월 18일_ -- [NumPy 1.24.0](https://numpy.org/doc/stable/release/1.24.0-notes.html)이 출시되었습니다. 주요 기능들은 다음과 같습니다: + +* 스태킹 함수를 위한 새 "dtype" 및 "casting" 키워드. +* 새 F2PY 기능 및 수정. +* 수많은 지원 종료 예정 기능, 확인하세요. +* 수많은 만료된 기능, + +NumPy 1.24.0 릴리스에서는 dtype의 처리 및 형변환을 개선하고 실행 속도를 높이는 작업, 문서를 보다 명료하게 다듬는 작업을 계속하고 있습니다. dtype의 형변환 및 정리를 변경하는 과정에서 수많은 기능들이 지원 종료 예정에 새로 포함되거나 완전히 만료되었습니다. 177명의 기여자가 생성한 444개의 풀 요청을 바탕으로 한 성과입니다. 지원하는 Python 버전은 3.8-3.11입니다. + +### NumPy 1.23.0 출시 + +_2022년 6월 22일_ -- [NumPy 1.23.0](https://numpy.org/doc/stable/release/1.23.0-notes.html)이 출시되었습니다. 주요 기능들은 다음과 같습니다: + +* `loadtxt`를 C로 구현하여 성능이 크게 향상되었습니다. +* 데이터 교환을 쉽게 하기 위해 Python 수준에서 DLPack을 개방합니다. +* 구조화된 dtype의 형변환 및 비교 방법을 변경했습니다. +* f2py를 개선했습니다. + +NumPy 1.23.0 릴리스에서는 dtype의 처리 및 형변환을 개선하고 실행 속도를 높이는 작업, 문서를 보다 명료하게 다듬는 작업, 오래된 지원 종료 예정 기능을 완전히 만료시키는 작업을 계속하고 있습니다. 151명의 기여자가 생성한 494개의 풀 요청을 바탕으로 한 성과입니다. 본 릴리즈에서 지원하는 Python 버전은 3.8-3.10입니다. Python 3.11은 rc 단계에 다다르면 지원할 예정입니다. + +### NumFOCUS DEI 연구: 참여 요청 + +_2022년 4월 13일_ -- NumPy는 [NumFOCUS](http://numfocus.org/)와 협력하여 [Gordon & Betty Moore 재단](https://www.moore.org/)에서 기금을 제공하는 [연구 프로젝트](https://numfocus.org/diversity-inclusion-disc/a-pivotal-time-in-numfocuss-project-aimed-dei-efforts?eType=EmailBlastContent&eId=f41a86c3-60d4-4cf9-86cf-58eb49dc968c)를 진행합니다. 본 연구는 오픈 소스 소프트웨어 커뮤니티에 기여자, 특히 역사적으로 과소평가된 집단의 기여자가 참여할 때 직면하는 장벽을 이해하는 것을 목표로 합니다. 연구팀은 새 기여자, 프로젝트 개발자 및 유지관리자, 과거에 기여한 사람들과 NumPy에 참여하고 기여한 경험에 대해 이야기하고자 합니다. + +**경험을 공유하고 싶으신가요?** + +간단한 ["참여 희망" 양식](https://numfocus.typeform.com/to/WBWVJSqe)을 작성해주세요. 양식에서 연구 목표, 개인정보 보호, 기밀 유지 사항에 대한 추가 정보를 확인할 수 있습니다. 당신의 참여가 다양성과 포용성을 갖춘 오픈 소스 소프트웨어 커뮤니티의 성장과 지속 가능성에 도움이 될 것입니다. 승인된 참가자는 연구팀과 30분 면담을 진행하게 됩니다. + +### Numpy 1.22.0 출시 + +_2021년 12월 31일_ -- [NumPy 1.22.0](https://numpy.org/doc/stable/release/1.22.0-notes.html)이 출시되었습니다. 주요 기능들은 다음과 같습니다: + +* 기본 네임스페이스에 대해 유형 주석의 지원을 거의 완료했습니다. 업스트림 코드는 항상 변하므로 추가 개선이 있을 수 있지만 주요 작업은 완료되었습니다. 아마도 이 릴리스에서 가장 체감되는 개선 사항일 것입니다. +* 제안된 [배열 API 표준의 예비 버전](https://data-apis.org/array-api/latest/)이 제공됩니다([NEP 47](https://numpy.org/neps/nep-0047-array-api-standard.html) 참조). 이는 CuPy 및 JAX와 같은 라이브러리에서 사용할 수 있는 표준 함수 모음을 만드는 단계입니다. +* NumPy가 DLPack 백엔드로 구동됩니다. DLPack은 배열(텐서) 데이터에 대한 공통 교환 형식을 제공합니다. +* `quantile`, `percentile` 관련 함수를 위한 새 메서드를 추가했습니다. 새 메서드를 이용해 문헌에서 일반적으로 쓰이는 처리를 진행할 수 있습니다. +* 범용 함수가 대부분의 [NEP 43](https://numpy.org/neps/nep-0043-extensible-ufuncs.html)을 구현하도록 리팩터링되었습니다. 이를 통해 미래의 DType API를 실험할 수 있는 능력도 갖췄습니다. +* 새 구성 가능한 메모리 할당자를 다운스트림 프로젝트에서 사용할 수 있습니다. + +NumPy 1.22.0은 153명의 기여자가 생성한 609개의 풀 요청을 바탕으로 만들어진 대형 릴리즈입니다. 본 릴리즈에서 지원하는 Python 버전은 3.8-3.10입니다. + +### 과학 Python 생태계에서 포용적 문화 발전 + +_2021년 8월 31일_ -- Chan Zuckerberg Initiative가 과학적 Python 프로젝트에서 역사적으로 소외된 그룹의 사람들을 온보딩, 포함 및 유지하고 NumPy, SciPy, Matplotlib 그리고 Pandas 의 커뮤니티 역학을 구조적으로 개선하기 위한 [보조금을 수여](https://chanzuckerberg.com/newsroom/czi-awards-16-million-for-foundational-open-source-software-tools-essential-to-biomedicine/)했음을 발표하게 되어 기쁩니다. + +[CZI의 Essential Open Source Software for Science 프로그램](https://chanzuckerberg.com/eoss/)의 일환으로 이 [Diversity & 포함 추가 보조금](https://cziscience.medium.com/advancing-diversity-and-inclusion-in-scientific-open-source-eaabe6a5488b)은 포괄적인 오픈 소스 커뮤니티를 육성하기 위한 관행을 식별, 문서화 및 구현하기 위한 전담 기여자 경험 리드 직책 생성을 지원합니다. 이 프로젝트는 Melissa Mendonça(NumPy) 님이 이끌고 Ralf Gommers(NumPy, SciPy), Hannah Aizenman, Thomas Caswell(Matplotlib), Matt Haberland(SciPy), Joris Van den Bossche(Pandas) 님이 추가 멘토링 및 지침을 제공합니다. + +이것은 프로젝트의 커뮤니티 역학을 구조적으로 개선해야 하는 활동을 발견하고 구현하는 것을 목표로 하는 야심 찬 프로젝트입니다. 새로운 교차 프로젝트 역할을 설정함으로써 과학적 Python 커뮤니티에 새로운 협업 모델을 도입하여 생태계 내에서 커뮤니티 구축 작업을 보다 효율적으로 수행하고 더 큰 결과를 얻을 수 있을 것으로 기대됩니다. 또한 특히 역사적으로 과소대표된 집단의 새로운 기여자를 참여시키고 유지하기 위해, 프로젝트에서 효과적인 것과 그렇지 않은 것에 대한 명확한 그림을 구축할 것으로 기대합니다. 마지막으로, 시행된 조치에 대해 자세한 보고서를 작성하여 커뮤니티와의 대표 및 상호 작용 측면에서 프로젝트에 어떤 영향을 미쳤는지 설명할 계획입니다. + +2개년 프로젝트가 2021년 11월부터 시작될 예정입니다. 프로젝트의 결과를 볼 날이 기대되네요! [여기에서 전체 제안서를 열람할 수 있습니다](https://figshare.com/articles/online_resource/Advancing_an_inclusive_culture_in_the_scientific_Python_ecosystem/16548063). + +### 2021년도 NumPy 설문조사 + +_2021년 7월 12일_ -- NumPy에서, 우리는 커뮤니티의 힘을 믿습니다. 작년에 75개국에서 1,236명의 NumPy 사용자가 첫 번째 설문조사에 참여했습니다. 설문 조사 결과를 통해 다음 12개월 동안 우리가 어떤 것에 집중해야 할지 아주 잘 이해할 수 있었습니다. + +이제 또다른 설문 조사를 진행할 시간이고, 여러분의 도움이 다시 한 번 필요합니다. 완료하는 데 약 15분 정도 소요될 겁니다. 설문지는 영어 외에도 8개 국어로 제공됩니다: 벵골어, 프랑스어, 힌디어, 일본어, 중국 관화, 포르투갈어, 러시아어, 스페인어. + +시작하려면 아래 링크를 눌러 주세요: https://berkeley.qualtrics.com/jfe/form/SV_aaOONjgcBXDSl4q. + + +### Numpy 1.21.0 출시 + +_2021년 9월 23일_ -- [NumPy 1.1.21](https://numpy.org/doc/stable/release/1.21.0-notes.html)이 출시되었습니다. 주요 기능들은 다음과 같습니다: + +- 더 많은 기능과 플랫폼을 다루는 지속적인 SIMD 작업, +- 새로운 dtype 인프라 및 캐스팅에 대한 초기 작업, +- Mac의 Python 3.8 및 Python 3.9용 universal2 휠, +- 문서화 향상, +- 주석 향상, +- 난수 생성에 이용되는 새 `PCG64DXSM` 비트 생성기. + +이번 NumPy 릴리즈는 175명이 기여해주신 581개의 풀 리퀘스트가 합쳐진 결과입니다. 본 릴리즈에서 지원하는 Python 버전은 3.7-3.9입니다. Python 3.10은 Python 3.10 릴리즈 이후 지원할 예정입니다. + + +### 2020년도 NumPy 설문조사 결과 + +_2021년 6월 22일_ -- 2020년에, NumPy 조사 팀은 조사방법론 학사 과정의 학생 및 교수와 협력하여 미시간 대학과 매릴렌드 대학이 공동으로 개최한 첫 공식 NumPy 커뮤니티 조사를 실시했습니다. 여기서 조사 결과를 확인하세요: https://numpy.org/user-survey-2020/. + + +### Numpy 1.20.0 출시 + +_2021년 9월 30일_ -- [NumPy 1.1.20](https://numpy.org/doc/stable/release/1.20.0-notes.html)이 출시되었습니다. 역대 최대의 NumPy 릴리즈입니다. 180명이 넘는 기여자분들께 감사드립니다. 다음은 이번 출시에서 가장 흥미로운 두가지 기능들 입니다. +- NumPy의 많은 부분에 대한 유형 주석 및 사용자와 다운스트림 라이브러리가 추가할 때 사용할 수 있는 `ArrayLike` 및 `DtypeLike` 별칭을 포함하는 새로운 `numpy.typing` 하위 모듈 자체 코드에 주석을 입력합니다. +- x86(SSE, AVX), ARM64(Neon) 및 PowerPC(VSX) 명령을 지원하는 다중 플랫폼 SIMD 컴파일러 최적화 입니다. 이는 많은 함수들의 상당한 성능향상을 가져왔습니다 (예: [sin/cos](https://github.com/numpy/numpy/pull/17587), [einsum](https://github.com/numpy/numpy/pull/18194)). + +### NumPy 프로젝트 내 다양성 + +_2020년 9월 20일_ -- 우리는 [NumPy 프로젝트 안에서의 다양성과 포용성에 관한 소셜 미디어의 상태 및 토론에 대한 성명서를 작성했습니다](/diversity_sep2020). + + +### Nature에 첫 공식 NumPy 논문 발표! + +_2020년 9월 16일_ -- [NumPy에 대한 첫 번째 공식 논문](https://www.nature.com/articles/s41586-020-2649-2)이 Nature에 리뷰 기사로 게재되었음을 발표하게 되어 기쁩니다. NumPy 1.0이 나온 지 14년 만입니다. 이 백서에서는 배열 프로그래밍의 응용 프로그램 및 기본 개념, NumPy 위에 구축된 풍부한 과학적 Python 생태계, CuPy, Dask 및 JAX와 같은 외부 배열 및 텐서 라이브러리와의 상호 운용성을 촉진하기 위해 최근에 추가된 배열 프로토콜을 다룹니다. + + +### Python 3.9가 곧 출시하는데, NumPy는 바이너리 Wheel을 언제 출시합니까? + +_2020년 9월 14일_ -- Python 3.9가 몇 주 내로 출시될 것입니다. 만약 Python 얼리어답터라면, NumPy (그리고 SciPy 등 다른 바이너리 패키지) 가 릴리즈 시일에 바이너리 Wheel을 준비하지 못한다는 것을 알고 실망했을 수 있습니다. 새로운 Python 버전에 빌드 환경을 맞추는 것은 많은 노력을 요하고, 패키지가 PyPI 및 conda-forge에 배포되는 데에는 일반적으로 몇 주가 걸립니다. 출시를 대비하려면 아래 요건을 충족하도록 하십시오. +- `pip` 버전을 최소 20.1로 업데이트하여 `manylinux2010` 및 `manylinux2014`를 지원하도록 합니다 +- [`--only-binary=numpy`](https://pip.pypa.io/en/stable/reference/pip_install/#cmdoption-only-binary)를 사용하거나 또는 `--only-binary=:all:`을 사용하여 `pip`가 소스에서 빌드하는 것을 막아주세요. + + +### NumPy 1.19.2 출시 + +_2020년 9월 10일_ -- [NumPy 1.19.2](https://numpy.org/devdocs/release/1.19.2-notes.html)이 출시되었습니다. 1.19 시리즈의 이 최신 릴리스는 몇 가지 버그를 수정하고 [다가오는 Cython 3.x 릴리스](http://docs.cython.org/en/latest/src/changes.html)를 준비하며 setuptools를 고정하여 업스트림 수정이 진행되는 동안 distutils가 계속 작동하도록 합니다. aarch64 휠은 다양한 Linux 배포판에서 사용되는 다양한 페이지 크기 문제를 해결하는 최신 manylinux2014 릴리스로 제작되었습니다. + +### 최초의 NumPy 설문조사가 진행 중입니다! + +_2020년 7월 2일_ -- 본 설문조사는 소프트웨어 및 커뮤니티로서의 NumPy 개발에 대하여, 의사결정의 우선 순위를 안내하고 설정하기 위해 실시됩니다. 설문지는 영어 외에도 8개 국어로 제공됩니다: 벵골어, 프랑스어, 힌디어, 일본어, 중국 관화, 포르투갈어, 러시아어, 스페인어. + +NumPy를 개선하게 도와주시고 이를위해 설문조사에 참여해 주세요. [여기](https://umdsurvey.umd.edu/jfe/form/SV_8bJrXjbhXf7saAl). + + +### NumPy에 새로운 로고가 생겼습니다! + +_2020년 6월 24일_ -- NumPy에 새로운 로고가 생겼습니다. + +NumPy 로고 + +이전 로고를 깔끔하고 현대적으로 다시 디자인했습니다. 새 로고를 만들어 주신 Isabela Presedo-Floyd님께 감사드립니다. 또 15년이 넘는 기간 동안 저희가 사용했던 로고를 만들어 주신 Travis Vaught님께도 감사의 말씀을 드립니다. + + +### NumPy 1.19.0 출시 + +_2020년 6월 20일_ -- NumPy 1.19.0이 출시되었습니다. Python 2의 지원을 중단한 첫 릴리즈라서 "정리 릴리즈"라고도 불립니다. 이제 지원하는 Python 최소 버전은 3.6입니다. 중요한 새 기능을 꼽자면, NumPy 1.17.0에 도입된 난수 생성 인프라를 Cython에서 접근할 수 있게 되었다는 것입니다. + + +### Season of Docs 승인 + +_2020년 5월 11일_ -- NumPy가 Google Season of Docs 프로그램의 선도 조직으로 승인되었습니다. 테크니컬 라이터와 협력해서 NumPy 문서를 다시 한 번 개선할 수 있는 기회를 갖게 되어 좋습니다! 이상 자세한 내용은 [공식 문서 시즌 사이트](https://developers.google.com/season-of-docs/) 및 [아이디어 페이지](https://github.com/numpy/numpy/wiki/Google-Season-of-Docs-2020-Project-Ideas) 를 참조하세요. + + +### NumPy 1.18.0 출시 + +_2019년 12월 22일_ -- NumPy 1.18.0이 출시되었습니다. 1.17.0에서의 주요 변경점을 통합하는 릴리즈입니다. 본 릴리즈는 Python 3.5를 지원하는 마지막 마이너 릴리즈입니다. 릴리즈의 주요 내용으로는, 64비트 BLAS 및 LAPACK 라이브러리와 연결하기 위한 환경 조성, `numpy.random`을 위한 새로운 C-API 등이 있습니다. + +자세한 내용은 [출시 노트](https://github.com/numpy/numpy/releases/tag/v1.18.0)를 참조하세요. + + +### NumPy가 Chan Zuckerberg Initiative에서 보조금을 받았습니다 + +_2019년 11월 15일_ -- NumPy의 주요 종속 패키지 중 하나인 NumPy와 OpenBLAS가 챈 저커버그 이니셔티브의 [과학 프로그램용 중요 오픈소스 소프트웨어](https://chanzuckerberg.com/eoss/) 지원을 통해 19만 5천 달러에 달하는 공동 보조금을 받았다는 소식을 전할 수 있어 기쁩니다. 이곳에서는 과학에 중요한 오픈소스 도구에 대해 유지 관리, 성장, 개발 및 커뮤니티 참여를 지원합니다. + +이 보조금은 NumPy 문서, 웹사이트 재설계 및 커뮤니티 개발을 개선하여 빠르게 성장하는 대규모 사용자 기반에 더 나은 서비스를 제공하고 프로젝트의 장기적인 지속 가능성을 보장하는 데 사용될 것입니다. OpenBLAS 팀은 OpenBLAS가 의존하는 ReLAPACK(Recursive LAPACK) 의 알고리즘 개선뿐만 아니라 특히 스레드 안전성, AVX-512 및 스레드 로컬 스토리지(TLS) 문제와 같은 일련의 핵심 기술 문제를 해결하는 데 집중할 것입니다. + +제안된 계획 및 결과물에 대한 자세한 내용은 [전체 보조금 제안](https://figshare.com/articles/Proposal_NumPy_OpenBLAS_for_Chan_Zuckerberg_Initiative_EOSS_2019_round_1/10302167)에서 확인할 수 있습니다. 2019년 12월 1일부터 작업을 시작하여 다음 12개월 동안 진행할 예정입니다. + + + + +## 릴리즈 + +NumPy 릴리즈의 목록입니다. 릴리즈 노트로 링크도 걸려 있습니다. 버그 수정 릴리즈(`x.y.z`에서 `z`만 바뀐 경우)에는 새로운 기능이 없습니다. 마이너 릴리즈(`y`가 증가한 경우)에는 새로운 기능이 있습니다. + +- NumPy 1.26.4 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.26.4)) -- _ 2024년 2월 5일_. +- NumPy 1.26.3 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.26.3)) -- _2024년 1월 2일_. +- NumPy 1.26.2 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.26.2)) -- _2023년 1월 2일_. +- NumPy 1.26.1 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.26.1)) -- _2023년 10월 14일_. +- NumPy 1.26.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.26.0)) -- _2023년 16월 9일_. +- NumPy 1.25.2 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.25.2)) -- _2023년 7월 31일_. +- NumPy 1.25.1 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.25.1)) -- _2023년 7월 8일_. +- NumPy 1.24.4 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.24.4)) -- _2023년 6월 26일_. +- NumPy 1.25.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.25.0)) -- _2023년 6월 17일_. +- NumPy 1.24.3 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.24.3)) -- _2023년 4월 22일_. +- NumPy 1.24.2 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.24.2)) -- _2023년 2월 5일_. +- NumPy 1.24.1 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.24.1)) -- _2022년 12월 26일_. +- NumPy 1.24.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.24.0)) -- _2022년 12월 18일_. +- NumPy 1.23.5 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.23.5)) -- _2022년 11월 19일_. +- NumPy 1.23.4 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.23.4)) -- _2022년 10월 12일_. +- NumPy 1.23.3 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.23.3)) -- _2022년 9월 9일_. +- NumPy 1.23.2 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.23.2)) -- _2022년 8월 14일_. +- NumPy 1.23.1 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.23.1)) -- _2022년 7월 8일_. +- NumPy 1.23.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.23.0)) -- _2022년 6월 22일_. +- NumPy 1.22.4 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.22.4)) -- _2022년 5월 20일_. +- NumPy 1.21.6 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.21.6)) -- _2022년 4월 12일_. +- NumPy 1.22.3 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.22.3)) -- _2022년 3월 7일_. +- NumPy 1.22.2 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.22.2)) -- _2022년 2월 3일_. +- NumPy 1.22.1 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.22.1)) -- _2022년 1월 14일_. +- NumPy 1.22.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.22.0)) -- _2021년 12월 31일_. +- NumPy 1.21.5 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.21.5)) -- _2021년 12월 19일_. +- NumPy 1.21.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.21.0)) -- _2021년 6월 22일_. +- NumPy 1.20.3 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.20.3)) -- _2021년 5월 10일_. +- NumPy 1.20.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.20.0)) -- _2021년 1월 30일_. +- NumPy 1.19.5 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.19.5)) -- _2021년 1월 5일_. +- NumPy 1.19.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.19.0)) -- _2020년 6월 20일_. +- NumPy 1.18.4 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.18.4)) -- _2020년 5월 3일_. +- NumPy 1.17.5 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.17.5)) -- _2020년 1월 1일_. +- NumPy 1.18.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.18.0)) -- _2019년 12월 22일_. +- NumPy 1.17.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.17.0)) -- _2019년 7월 26일_. +- NumPy 1.16.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.16.0)) -- _2019년 1월 14일_. +- NumPy 1.15.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.15.0)) -- _2018년 7월 23일_. +- NumPy 1.14.0 ([릴리즈 노트](https://github.com/numpy/numpy/releases/tag/v1.14.0)) -- _2018년 1월 7일_. diff --git a/content/ko/press-kit.md b/content/ko/press-kit.md new file mode 100644 index 0000000000..ddce954013 --- /dev/null +++ b/content/ko/press-kit.md @@ -0,0 +1,8 @@ +--- +title: 홍보 자료 +sidebar: false +--- + +저희는 당신이 NumPy 프로젝트의 상징을 논문, 코스 자료, 발표 자료 등에 삽입하기 쉽도록 하고자 합니다. + +[여기에서](https://github.com/numpy/numpy/tree/main/branding/logo) 여러 버전의 고화질 NumPy 로고를 찾을 수 있습니다. numpy.org 자료를 이용하는 경우, [NumPy 이용약관](/code-of-conduct)에 동의하게 됨을 명심하십시오. diff --git a/content/ko/privacy.md b/content/ko/privacy.md new file mode 100644 index 0000000000..b2fc57fcfd --- /dev/null +++ b/content/ko/privacy.md @@ -0,0 +1,8 @@ +--- +title: 개인정보 정책 +sidebar: false +--- + +**numpy.org**는 NumPy 프로젝트의 재정적 후원자인 [NumFOCUS, Inc.](https://numfocus.org)가 관리합니다. 이 웹 사이트에 대한 개인정보 정책을 확인하려면 https://numfocus.org/privacy-policy를 참고하세요. + +NumFOCUS의 데이터 수집, 이용, 공개 정책에 대하여 아무 질문이나 있으시다면, NumFOCUS 스태프인 privacy@numfocus.org로 연락해주세요. diff --git a/content/ko/report-handling-manual.md b/content/ko/report-handling-manual.md new file mode 100644 index 0000000000..cdd686956e --- /dev/null +++ b/content/ko/report-handling-manual.md @@ -0,0 +1,95 @@ +--- +title: NumPy 이용 약관 - 보고서의 후속 조치 방법 +sidebar: false +--- + +NumPy 행동 강령 위원회는 본 설명을 따릅니다. 문제를 해결할 때 일관성과 공정성을 확보하기 위한 지침입니다. + +[행동 강령](/code-of-conduct)을 시행하면 현재와 미래의 커뮤니티에 영향을 미칩니다. 우리가 가볍게 받아들이지 않는 행동입니다. 집행 조치를 검토할 때 행동 강령 위원회는 다음 가치와 지침을 염두에 둘 것입니다. + +* 비인간적이기보다는 개인적인 방식으로 행동하십시오. 위원회는 신고자의 사생활과 필요한 기밀을 존중하면서 상황을 이해하도록 당사자들을 참여시킬 수 있습니다. 그러나 때때로 한 명 이상의 개인과 직접 소통해야 할 필요가 있습니다. 위원회의 목표는 단지 공식적인 결정을 내리는 것이 아니라 우리 지역 사회의 건강을 개선하는 것입니다. +* 행동을 판단하기보다는 개인에 대한 공감을 강조하고 "좋음"과 "나쁨/악"이라는 이분법적인 레이블을 피하십시오. 노골적이고 분명한 공격성과 괴롭힘이 존재하며 우리는 단호하게 대처할 것입니다. 그러나 해결하기 어려운 것으로 입증될 수 있는 많은 시나리오는 정상적인 불일치가 여러 당사자의 도움이 되지 않거나 해로운 행동으로 귀결되는 시나리오입니다. 전체 맥락을 이해하고 모두를 다시 참여시키는 경로를 찾는 것은 어렵지만 궁극적으로 우리 커뮤니티에 가장 생산적입니다. +* 우리는 이메일이 어려운 매체이며 고립될 수 있음을 이해합니다. 개인적인 연락 없이 이메일을 통해 비판을 받는 것은 특히 고통스러울 수 있습니다. 따라서 다른 사람의 견해를 열린 마음으로 존중하는 분위기를 유지하는 것이 특히 중요합니다. 또한 투명하게 행동해야 하며 모든 구성원이 공정하고 공감하는 대우를 받을 수 있도록 최선을 다하겠다는 의미이기도 합니다. +* 차별은 미묘할 수도 있고 무의식적일 수도 있습니다. 일상적인 상호 작용에서 불공평과 적대감으로 나타날 수 있습니다. 저희는 이런 차별이 발생한다는 점을 인지하고 있으며 주의를 기울이고 조심할 것입니다. 저희는 귀하가 부당한 대우를 받았다고 느끼시는 경우 귀하의 의견을 듣고 싶습니다. 귀하의 불만 사항을 듣고 해결하기 위한 절차를 밟을 것입니다. +* 좋은 토론 관행에 참여를 늘리도록 도와주세요: 토론이 중단되었을 수 있는 부분을 파악하고 이러한 점에서 긍정적인 변화를 가져올 수 있는 실행 가능한 정보, 포인터 및 리소스를 제공하세요. +* 신입 회원의 필요를 염두에 두십시오. 특히 소외된 그룹의 참여를 늘리는 것을 목표로 명시적인 지원과 배려를 제공하십시오. +* 개인은 서로 다른 문화적 배경과 모국어를 가지고 있습니다. 원어민이 아닌 사람으로 인한 정직한 오해를 식별하고 문제를 이해하고 불쾌감을 주지 않도록 변경할 수 있는 사항을 이해하도록 돕습니다. 외국어로 복잡한 토론을 하는 것은 매우 위협적일 수 있으며 국적과 문화를 넘어 다양성을 키우고자 합니다. + + +## 중재 + +자발적인 비공식 중재는 우리가 사용할 수 있는 도구입니다. 두 명 이상의 당사자가 모두 부적절한 행동 (인간 갈등에서 슬프게도 흔한 일) 의 지점까지 확대된 경우와 같은 맥락에서 중재 프로세스를 촉진하는 것이 유용할 수 있습니다. 이것은 단지 예일 뿐입니다. 위원회는 어떤 경우에도 중재를 고려할 수 있으며, 그 과정은 엄격하게 자발적으로 이루어지며 어느 당사자도 참여하도록 압력을 받을 수 없다는 점을 염두에 두어야 합니다. 위원회가 중재를 제안하는 경우 위원회는 다음을 수행해야 합니다. + +* 중재자 역할을 할 수 있는 후보를 찾으십시오. +* 신고자(들) 의 동의를 얻습니다. 신고자는 중재 아이디어를 거부하거나 대체 중재자를 제안할 수 있는 완전한 자유가 있습니다. +* 피신고자의 동의를 얻습니다. +* 중재자 결정: 당사자는 제안된 후보와 다른 중재자를 제안할 수 있지만 모든 조건에 대해 공통된 합의에 도달한 경우에만 프로세스를 진행할 수 있습니다. +* 중재가 완료되기 위한 일정을 설정합니다 (이상적으로는 2주 이내). + +중재자는 모든 당사자와 관계를 맺고 모두가 만족할 만한 해결책을 모색합니다. 완료되면 조정자는 추가 단계에 대한 권장 사항과 함께 보고서(프로세스의 모든 당사자가 심사한) 를 위원회에 제공합니다. 그런 다음 위원회는 이러한 결과를 평가하고(만족스러운 해결 여부에 관계없이) 필요하다고 판단되는 추가 조치를 결정합니다. + + +## 위원회가 신고에 응답하는 방법 + +위원회(또는 위원회 위원) 는 보고서를 받으면 먼저 보고서가 명확하고 심각한 위반에 관한 것인지 여부를 결정합니다(아래에 정의됨). 만약 그렇다면, 정기적인 신고 처리 프로세스와 더불어 즉각적인 조치가 취해져야 합니다. + + +## 명확하고 심각한 권리침해 행위 + +우리는 인터넷 통신이 명백하고 노골적인 남용에서 시작되거나 악화되는 것이 매우 흔한 일이라는 것을 알고 있습니다. 우리는 개인적인 위협, 폭력적, 성차별적 또는 인종차별적 언어와 같은 명확하고 심각한 위반을 신속하게 처리할 것입니다. + +행동 강령 위원회 위원이 명백하고 심각한 위반 사실을 알게 되면 다음과 같은 조치를 취합니다. + +* 모든 NumPy 통신 채널에서 발신자를 즉시 연결 해제합니다. +* 보고자에게 보고가 접수되었으며 발신자의 연결이 끊어졌다고 회신합니다. +* 모든 경우에 중재자는 작성자에게 연락하기 위해 합당한 노력을 기울여야 하며, 그들의 언어나 행동이 어떻게 "명백하고 심각한 위반"에 해당하는지 구체적으로 알려야 합니다. 중재자는 또한 작성자가 이것이 불공평하다고 생각하거나 NumPy에 다시 연결되기를 원하는 경우 아래와 같이 행동 강령 위원회에 검토를 요청할 권리가 있다고 말해야 합니다. 중재자는 이 설명을 행동 강령 위원회에 발송해야 합니다. +* 행동 강령 위원회는 이 메커니즘이 일반적인 열띤 의견 불일치를 통제하는 데 사용되지 않도록 하기 위해 이 메커니즘이 적용된 모든 사례를 공식적으로 검토하고 승인할 것입니다. + + +## 신고 대응 + +보고서가 위원회에 보내지면 위원회는 접수 확인을 위해 즉시 보고자에게 회신할 것입니다. 이 회신은 72시간 이내에 보내야 하며 그룹은 그보다 훨씬 빨리 회신하기 위해 노력해야 합니다. + +보고서에 충분한 정보가 포함되어 있지 않은 경우 위원회는 조치를 취하기 전에 모든 관련 데이터를 수집합니다. 위원회는 사건에 대한 보다 완전한 설명을 얻기 위해 관련된 모든 개인에게 연락할 때 운영 위원회를 대신하여 행동할 권한이 있습니다. + +그런 다음 위원회는 사건을 검토하고 최선을 다해 다음을 결정할 것입니다. + +* 무슨 일이 일어났는지 +* 이 이벤트가 행동 강령 위반에 해당하는지 여부 +* 책임자는 누구인지 +* 이것이 현재 진행 중인 상황인지, 누군가의 신체적 안전에 대한 위협이 있는지 여부 + +이 정보는 서면으로 수집되며 가능할 때마다 그룹의 심의가 녹음 및 보관됩니다(예: 채팅 내용, 이메일 토론, 녹음된 회의 통화, 음성 대화 요약 등). + +행동의 일관성을 보장하고 프로젝트에 대한 제도적 기억을 제공하기 위해 이 위원회의 모든 활동 기록을 보관하는 것이 중요합니다. 이를 지원하기 위해 이 위원회의 기본 토론 채널은 정당한 요청이 있는 경우 현재 및 미래의 위원회 구성원과 운영 위원회 구성원이 액세스할 수 있는 개인 메일링 리스트가 될 것입니다. 위원회가 목록 외 커뮤니케이션(예: 조기/신속한 응답을 위한 전화 통화)을 사용해야 할 필요성을 발견하는 경우 모든 경우에 이를 목록에 다시 요약하여 프로세스에 대한 좋은 기록이 남도록 해야 합니다. + +행동 강령 위원회는 2주 이내에 결의안에 동의하는 것을 목표로 해야 합니다. 그 시간 내에 해결 방법을 결정할 수 없는 경우 위원회는 보고자에게 해결을 위한 업데이트 및 예상 일정에 대해 응답합니다. + + +## 해결 + +위원회는 반드시 합의를 바탕으로 결의를 내야 합니다. 그룹에서 합의가 이루어지지 못하고 1주 넘게 교착 상태에 빠진 경우, 결의를 내기 위해 해당 의제는 조정위원회로 이양됩니다. + +가능한 응답은 다음과 같습니다. + +* 추가 조치 없음: + - 위반이 발생하지 않았다고 판단되는 경우 + - 위원회가 응답을 고려하는 동안 문제가 공개적으로 해결된 경우. +* 자발적 중재 조정: 모든 관련 당사자가 동의하는 경우 위원회는 위에서 설명한 대로 중재 절차를 촉진할 수 있습니다. +* 공개적으로 상기시키고 일부 행동/행동/언어가 부적절하다고 판단되었으며 그 이유를 현재 상황에서 지적하거나 일부 사람들에게 상처를 줄 수 있으며 커뮤니티가 자체 조정하도록 요청합니다. +* 관련된 개인에 대한 위원회의 사적인 견책. 이 경우 그룹 의장은 그룹을 참조하여 이메일을 통해 개인(들) 에게 질책을 전달합니다. +* 공개 질책. 이 경우 위원회 위원장은 위반이 발생한 동일한 장소에서 실행 가능한 범위 내에서 질책합니다. 예를 들어 이메일 위반에 대한 원래 메일링 리스트이지만 사람/컨텍스트가 없을 수 있는 채팅방 토론의 경우 다른 방법으로 도달할 수 있습니다. 그룹은 문서화 목적으로 이 메시지를 다른 곳에 게시하도록 선택할 수 있습니다. +* 신고자가 이 생각에 동의한다고 가정하고 공개 또는 비공개 사과 요청: 신고자는 재량에 따라 위반자와의 추가 연락을 거부할 수 있습니다. 의장이 이 요청을 전달할 것입니다. 위원회는 선택하는 경우 이 요청에 "문자열"을 첨부할 수 있습니다. 예를 들어 그룹은 위반자에게 메일링 리스트의 회원 자격을 유지하기 위해 사과를 요청할 수 있습니다. +* 위원회가 개인에게 일시적으로 커뮤니티 참여를 자제하도록 요청하는 "상호 합의 중단". 개인이 자발적으로 일시적인 휴식을 취하지 않기로 선택한 경우 위원회는 "필수 냉각 기간"을 발행할 수 있습니다. +* 일부 또는 모든 NumPy 공간(메일링 리스트, gitter.im 등)에서 영구적 또는 일시적 금지. 그룹은 이러한 모든 금지에 대한 기록을 유지하여 나중에 검토하거나 다른 방법으로 유지할 수 있습니다. + +일단 결의안이 합의되면, 그러나 그것이 시행되기 전에 위원회는 원래 신고자와 영향을 받는 다른 당사자들에게 연락하여 제안된 결의안을 설명합니다. 위원회는 이 결의안이 수용 가능한지 묻고 기록을 위해 피드백을 기록해야 합니다. + +최종적으로 위원회는 NumPy 조정위원회에 보고서를 만들어 제출하게 됩니다 (추방 등 효력이 지속되는 결의가 발생하는 경우 NumPy 핵심 팀에게도 보고합니다). + +위원회는 문제를 반드시 비공개 상태로 다룰 것입니다. 모든 공개 성명은 행동강령 위원회 혹은 NumPy 조정위원회에서 담당합니다. + + +## 이해관계 충돌 + +이해관계 충돌이 일어난 경우, 위원회 회원은 즉시 이 사실을 다른 회원에게 고지하고 필요한 경우 자진 사퇴해야 합니다. diff --git a/content/ko/tabcontents.yaml b/content/ko/tabcontents.yaml new file mode 100644 index 0000000000..4fa0164021 --- /dev/null +++ b/content/ko/tabcontents.yaml @@ -0,0 +1,373 @@ +params: + machinelearning: + paras: + - + para1: NumPy는 [scikit-learn](https://scikit-learn.org) 및 [SciPy](https://www.scipy.org)와 같은 강력한 기계 학습 라이브러리의 기반을 형성합니다. 기계 학습이 성장함에 따라 NumPy에 구축된 라이브러리 목록도 늘어납니다. [TensorFlow의](https://www.tensorflow.org) 딥 러닝 기능은 폭넓게 응용할 수 있습니다. — 그 중에는 음성 및 이미지 인식, 텍스트 기반 애플리케이션, 시계열 분석 및 비디오 감지가 있습니다. 또 다른 딥 러닝 라이브러리인 [PyTorch](https://pytorch.org)는 컴퓨터 비전 및 자연어 처리 연구자들 사이에서 인기가 있습니다. [MXNet](https://github.com/apache/incubator-mxnet)은 딥 러닝을 위한 청사진과 템플릿을 제공하는 또 다른 AI 패키지입니다. + para2: '[ensemble](https://towardsdatascience.com/ensemble-methods-bagging-boosting-and-stacking-c9214a10a205) 메서드라고 하는 비닝(binning), 배깅(bagging), 스태킹(stacking), 부스팅(boosting)과 같은 통계 기법은 다음과 같은 도구로 구현되는 ML 알고리즘에 속합니다. [XGBoost](https://github.com/dmlc/xgboost), [LightGBM](https://lightgbm.readthedocs.io/en/latest/) 및 [CatBoost](https://catboost.ai) — – 가장 빠른 추론 엔진 중 하나입니다. [Yellowbrick](https://www.scikit-yb.org/en/latest/) 및 [Eli5](https://eli5.readthedocs.io/en/latest/)는 기계 학습 시각화를 제공합니다.' + arraylibraries: + intro: + - + text: NumPy의 API는 라이브러리가 혁신적인 하드웨어를 활용하거나, 특수 배열 유형을 생성하거나, NumPy가 제공하는 것 이상의 기능을 추가하도록 작성되는 출발점입니다. + headers: + - + text: 배열 라이브러리 + - + text: 기능 및 응용 분야 + libraries: + - + title: Dask + text: 분석을 위한 분산 배열 및 고급 병렬 처리를 통해 규모에 맞는 성능을 구현합니다. + img: /images/content_images/arlib/dask.png + alttext: Dask + url: https://dask.org/ + - + title: CuPy + text: Python에서 GPU 가속 컴퓨팅을 구현해주며 NumPy와 호환되는 배열 라이브러리. + img: /images/content_images/arlib/cupy.png + alttext: CuPy + url: https://cupy.chainer.org + - + title: JAX + text: "NumPy 프로그램을 부분적으로 변환하여 벡터화, GPU/TPU의 적시 컴파일을 제공하는 라이브러리." + img: /images/content_images/arlib/jax_logo_250px.png + alttext: JAX + url: https://github.com/google/jax + - + title: Xarray + text: 고급 통계 및 시각화를 구동하기 위하여 라벨링 및 인덱싱이 이뤄진 다차원 배열을 제공 + img: /images/content_images/arlib/xarray.png + alttext: xarray + url: https://xarray.pydata.org/en/stable/index.html + - + title: Sparse + text: Dask 및 SciPy의 희소 선형 대수와 통합되는 NumPy 호환 희소 배열 라이브러리입니다. + img: /images/content_images/arlib/sparse.png + alttext: sparse + url: https://sparse.pydata.org/en/latest/ + - + title: PyTorch + text: 연구 프로토타이핑에서 프로덕션 배포로의 경로를 가속화하는 딥 러닝 프레임워크입니다. + img: /images/content_images/arlib/pytorch-logo-dark.svg + alttext: PyTorch + url: https://pytorch.org/ + - + title: TensorFlow + text: 기계 학습을 위한 엔드 투 엔드 플랫폼으로 ML 기반 애플리케이션을 쉽게 구축하고 배포할 수 있습니다. + img: /images/content_images/arlib/tensorflow-logo.svg + alttext: TensorFlow + url: https://www.tensorflow.org + - + title: MXNet + text: 유연한 연구 프로토타이핑 및 생산에 적합한 딥 러닝 프레임워크입니다. + img: /images/content_images/arlib/mxnet_logo.png + alttext: MXNet + url: https://mxnet.apache.org/ + - + title: Arrow + text: 열 기반 메모리 내 데이터 및 분석을 위한 교차 언어 개발 플랫폼입니다. + img: /images/content_images/arlib/arrow.png + alttext: arrow + url: https://github.com/apache/arrow + - + title: xtensor + text: 수치 분석을 위한 브로드캐스팅 및 지연 컴퓨팅이 포함된 다차원 배열. + img: /images/content_images/arlib/xtensor.png + alttext: xtensor + url: https://github.com/xtensor-stack/xtensor-python + - + title: Awkward Array + text: NumPy와 유사한 관용어로 JSON과 유사한 데이터를 조작합니다. + img: /images/content_images/arlib/awkward.svg + alttext: awkward + url: https://awkward-array.org/ + - + title: uarray + text: 구현에서 API를 분리하는 Python 백엔드 시스템; unumpy는 NumPy API를 제공합니다. + img: /images/content_images/arlib/uarray.png + alttext: uarray + url: https://uarray.org/en/latest/ + scientificdomains: + intro: + - + text: Python으로 작업하는 거의 모든 과학자는 NumPy의 힘을 이용합니다. + - + text: "NumPy는 C 및 Fortran과 같은 언어의 계산 능력을 배우고 사용하기 훨씬 쉬운 언어인 Python으로 가져옵니다. 이 힘에는 단순함이 있습니다. NumPy의 솔루션은 종종 명확하고 우아합니다." + libraries: + - + title: 양자 컴퓨팅 + alttext: 컴퓨터 칩 + img: /images/content_images/sc_dom_img/quantum_computing.svg + links: + - + url: http://qutip.org + label: QuTiP + - + url: https://pyquil-docs.rigetti.com/en/stable + label: PyQuil + - + url: https://qiskit.org + label: Qiskit + - + url: https://pennylane.ai + label: PennyLane + - + title: 통계적 컴퓨팅 + alttext: 선이 위로 이동하는 선그래프 + img: /images/content_images/sc_dom_img/statistical_computing.svg + links: + - + url: https://pandas.pydata.org/ + label: Pandas + - + url: https://github.com/statsmodels/statsmodels + label: statsmodels + - + url: https://xarray.pydata.org/en/stable/ + label: Xarray + - + url: https://github.com/mwaskom/seaborn + label: Seaborn + - + title: 신호 처리 + alttext: 양의 값과 음의 값을 가지는 막대 차트 + img: /images/content_images/sc_dom_img/signal_processing.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://pywavelets.readthedocs.io/ + label: PyWavelets + - + url: https://python-control.org/ + label: python-control + - + url: https://hyperspy.org/ + label: HyperSpy + - + title: 이미지 처리 + alttext: 산이 찍힌 사진 + img: /images/content_images/sc_dom_img/image_processing.svg + links: + - + url: https://scikit-image.org/ + label: Scikit-image + - + url: https://opencv.org/ + label: OpenCV + - + url: https://mahotas.rtfd.io/ + label: Mahotas + - + title: 그래프 및 네트워크 + alttext: 간단한 그래프. + img: /images/content_images/sc_dom_img/sd6.svg + links: + - + url: https://networkx.org/ + label: NetworkX + - + url: https://graph-tool.skewed.de/ + label: graph-tool + - + url: https://igraph.org/python/ + label: igraph + - + url: https://pygsp.rtfd.io/ + label: PyGSP + - + title: 천문학 + alttext: 망원경 + img: /images/content_images/sc_dom_img/astronomy_processes.svg + links: + - + url: https://www.astropy.org/ + label: AstroPy + - + url: https://github.com/sunpy/sunpy + label: SunPy + - + url: https://github.com/spacepy/spacepy + label: SpacePy + - + title: 인지심리학 + alttext: 톱니바퀴가 안에서 돌아가는 사람의 머리 + img: /images/content_images/sc_dom_img/cognitive_psychology.svg + links: + - + url: https://www.psychopy.org/ + label: PsychoPy + - + title: 생물정보학 + alttext: DNA 가닥 + img: /images/content_images/sc_dom_img/bioinformatics.svg + links: + - + url: https://biopython.org/ + label: BioPython + - + url: http://scikit-bio.org/ + label: Scikit-Bio + - + url: https://github.com/openvax/pyensembl + label: PyEnsembl + - + url: http://etetoolkit.org/ + label: ETE + - + title: 베이지안 추론 + alttext: 종 모양 곡선이 그려진 그래프 + img: /images/content_images/sc_dom_img/bayesian_inference.svg + links: + - + url: https://pystan.readthedocs.io/en/latest/ + label: PyStan + - + url: https://docs.pymc.io/ + label: PyMC3 + - + url: https://arviz-devs.github.io/arviz/ + label: ArviZ + - + url: https://emcee.readthedocs.io/ + label: emcee + - + title: 수학적 분석 + alttext: 수학 기호 4개 + img: /images/content_images/sc_dom_img/mathematical_analysis.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://www.sympy.org/ + label: SymPy + - + url: https://github.com/cvxgrp/cvxpy + label: cvxpy + - + url: https://fenicsproject.org/ + label: FEniCS + - + title: 화학 + alttext: 시험관 + img: /images/content_images/sc_dom_img/chemistry.svg + links: + - + url: https://cantera.org/ + label: Cantera + - + url: https://www.mdanalysis.org/ + label: MDAnalysis + - + url: https://github.com/rdkit/rdkit + label: RDKit + - + url: https://www.pybamm.org/ + label: PyBaMM + - + title: 지구과학 + alttext: 지구 + img: /images/content_images/sc_dom_img/geoscience.svg + links: + - + url: https://pangeo.io/ + label: Pangeo + - + url: https://simpeg.xyz/ + label: Simpeg + - + url: https://github.com/obspy/obspy/wiki + label: ObsPy + - + url: https://www.fatiando.org/ + label: Fatiando a Terra + - + title: 지리학적 처리 + alttext: 지도 + img: /images/content_images/sc_dom_img/GIS.svg + links: + - + url: https://shapely.readthedocs.io/ + label: Shapely + - + url: https://geopandas.org/ + label: GeoPandas + - + url: https://python-visualization.github.io/folium + label: Folium + - + title: 아키텍처 및 엔지니어링 + alttext: 마이크로프로세서 개발 보드 + img: /images/content_images/sc_dom_img/robotics.svg + links: + - + url: https://compas.dev/ + label: COMPAS + - + url: https://cityenergyanalyst.com/ + label: 도시 에너지 분석가 + - + url: https://nortikin.github.io/sverchok/ + label: Sverchok + datascience: + intro: "NumPy는 풍부한 데이터 과학 라이브러리 생태계의 핵심에 있습니다. 일반적인 탐색적 데이터 과학 워크플로는 다음과 같습니다." + image1: + - + img: /images/content_images/ds-landscape.png + alttext: Python 라이브러리 다이어그램. 다섯 가지 범주는 '추출, 변환, 로드', '데이터 탐색', '데이터 모델링', '데이터 평가' 및 '데이터 프레젠테이션'입니다. + image2: + - + img: /images/content_images/data-science.png + alttext: 세 개의 겹치는 원의 다이어그램입니다. 원에는 '수학', '컴퓨터 과학' 및 '영역 전문성'이라는 레이블이 지정되어 있습니다. 세 개의 원이 겹치는 다이어그램의 중간에는 'Data Science'라는 레이블이 붙은 영역이 있습니다. + examples: + - + text: "추출, 변환, 로드: [Pandas](https://pandas.pydata.org), [Intake](https://intake.readthedocs.io), [PyJanitor](https:/ /pyjanitor.readthedocs.io/)" + - + text: "탐색적 분석: [Jupyter](https://jupyter.org), [Seaborn](https://seaborn.pydata.org), [Matplotlib](https://matplotlib.org), [Altair](https://altair-viz.github.io)" + - + text: "모델링 및 평가: [scikit-learn](https://scikit-learn.org), [statsmodels](https://www.statsmodels.org/stable/index.html), [PyMC3](https://docs.pymc.io), [spaCy](https://spacy.io)" + - + text: "대시보드에서 보고: [Dash](https://plotly.com/dash), [Panel](https://panel.holoviz.org), [Voila](https://github.com/voila-dashboards/voila)" + content: + - + text: 대용량 데이터의 경우 [Dask](https://dask.org) 및 [Ray](https://ray.io/)가 확장되도록 설계되었습니다. 안정적인 배포는 데이터 버전 관리([DVC](https://dvc.org)), 실험 추적([MLFlow](https://mlflow.org)) 및 워크플로 자동화([Airflow](https:// airflow.apache.org), Dagster](https://dagster.io) 와 [Prefect](https://www.prefect.io)). + visualization: + images: + - + url: https://www.fusioncharts.com/blog/best-python-data-visualization-libraries + img: /images/content_images/v_matplotlib.png + alttext: Matplotlib으로 만든 streamplot + - + url: https://github.com/yhat/ggpy + img: /images/content_images/v_ggpy.png + alttext: ggpy로 만든 산점도 + - + url: https://www.journaldev.com/19692/python-plotly-tutorial + img: /images/content_images/v_plotly.png + alttext: plotly로 만든 상자 그림 + - + url: https://altair-viz.github.io/gallery/streamgraph.html + img: /images/content_images/v_altair.png + alttext: altair로 만든 스트림 그래프 + - + url: https://seaborn.pydata.org + img: /images/content_images/v_seaborn.png + alttext: Seaborn에서 만든 두 가지 유형의 그래프, 플롯 그래프와 빈도 그래프의 pairplot + - + url: https://docs.pyvista.org/ + img: /images/content_images/v_pyvista.png + alttext: PyVista로 만든 3D 볼륨 렌더링. + - + url: https://napari.org + img: /images/content_images/v_napari.png + alttext: napari로 만든 다차원 이미지. + - + url: https://vispy.org/gallery/index.html + img: /images/content_images/v_vispy.png + alttext: vispy로 만든 보로노이 다이어그램. + content: + - + text: 몇 가지만 예를 들자면 NumPy는 [Matplotlib](https://matplotlib.org), [Seaborn](https://seaborn.pydata.org), [Plotly](https://plot.ly), [Altair](https://altair-viz.github.io), [Bokeh](https://docs.bokeh.org/en/latest/), [Holoviz](https://holoviz.org), [Vispy](http://vispy.org), [Napari](https://github.com/napari/napari), [PyVista](https://github.com/pyvista/pyvista) 등이 포함되어 있으며 급격히 성장해나가고 있는 [Python visualization landscape](https://pyviz.org/overviews/index.html)의 핵심 구성 요소 중 하나입니다. + - + text: NumPy는 큰 배열을 고속으로 처리할 수 있어 연구자가 기존 Python이 처리할 수 있는 데이터셋보다 훨씬 큰 것도 시각화할 수 있도록 합니다. diff --git a/content/ko/teams.md b/content/ko/teams.md new file mode 100644 index 0000000000..5654f4d6c2 --- /dev/null +++ b/content/ko/teams.md @@ -0,0 +1,22 @@ +--- +title: NumPy 팀 +sidebar: false +--- + +저희는 양질의 오픈소스 소프트웨어를 구축하여 전 세계의 과학 및 연구 커뮤니티를 지원한다는 사명을 지닌 국제적 팀입니다. [팀에 참여하십시오]({{< relref "/contribute" >}})! + +{{< include-html "static/gallery/maintainers.html" >}} + +{{< include-html "static/gallery/docs-team.html" >}} + +{{< include-html "static/gallery/web-team.html" >}} + +{{< include-html "static/gallery/triage-team.html" >}} + +{{< include-html "static/gallery/survey-team.html" >}} + +{{< include-html "static/gallery/emeritus-maintainers.html" >}} + +# 운영 + +운영 위원회 구성원 목록은 [여기](https://numpy.org/about/)를 참고하십시오. diff --git a/content/ko/user-survey-2020.md b/content/ko/user-survey-2020.md new file mode 100644 index 0000000000..68505b30f9 --- /dev/null +++ b/content/ko/user-survey-2020.md @@ -0,0 +1,18 @@ +--- +title: 2020 NUMPY 커뮤니티 설문조사 +sidebar: false +--- + +2020년, NumPy 팀은 조사방법론 학사 과정의 학생 및 교수와 협력하여 미시간 대학과 매릴렌드 대학이 공동으로 개최한 첫 공식 NumPy 커뮤니티 조사를 실시했습니다. 75개국 내 1200명 이상의 사용자 여러분들께서 저희가 NumPy 커뮤니티의 가닥을 잡을 수 있도록 도와주기 위해 참여해주셨으며 프로젝트의 미래에 대한 생각을 표현해주셨습니다. + +{{< figure src="/surveys/NumPy_usersurvey_2020_report_cover.png" class="fig-left" alt="'NumPy Community Survey 2020 - results'라는 제목이 붙은 2020년 NumPy 사용자 설문조사 보고서 표지" width="250">}} +src = '/surveys/NumPy_usersurvey_2020_report_cover.png' alt = '"2020 NumPy 커뮤니티 설문조사 - 결과"라는 제목이 붙은 2020 NumPy 사용자 설문조사 보고서의 표지' width = '250' +{{< /figure >}} + +**[보고서를 내려받아서](/surveys/NumPy_usersurvey_2020_report.pdf)** 설문조사 결과를 자세히 들여다 보세요. + + +요점만 보시려면, **[이 인포그래픽](https://github.com/numpy/numpy-surveys/blob/master/images/2020NumPysurveyresults_community_infographic.pdf)**을 참고하시기 바랍니다. + +더욱 자세한 정보가 궁금하신가요? **https://numpy.org/user-survey-2020-details/** 페이지를 방문하세요. + diff --git a/content/ko/user-surveys.md b/content/ko/user-surveys.md new file mode 100644 index 0000000000..9fec78190d --- /dev/null +++ b/content/ko/user-surveys.md @@ -0,0 +1,10 @@ +--- +title: NUMPY 사용자 설문조사 +sidebar: false +--- + +**2020년** NumPy 조사 팀은 조사방법론 학사 과정의 학생 및 교수와 협력하여 미시간 대학과 매릴렌드 대학이 공동으로 개최한 첫 공식 NumPy 커뮤니티 조사를 실시했습니다. [여기](https://numpy.org/user-survey-2020/)서 조사 결과를 확인하세요. + +**2021년** 수집한 데이터가 현재 분석 중입니다. + +과거나 미래 설문조사에 대해 질문이나 제안 사항이 있으시면, [여기](https://github.com/numpy/numpy-surveys/issues)서 이슈를 생성하세요. diff --git a/content/pt/_index.md b/content/pt/_index.md new file mode 100644 index 0000000000..96a58860fc --- /dev/null +++ b/content/pt/_index.md @@ -0,0 +1,49 @@ +--- +title: null +--- + +{{< grid columns="1 2 2 3" >}} + +{{< card >}} +title = 'Potentes arrays N-dimensionais.' +body = ''' +Rápido e versátil, os conceitos de vetorização, indexação e broadcasting do NumPy são os padrões de fato da computação de arrays hoje em dia. +''' +{{< /card >}} + +{{< card >}} +title = 'Ferramentas numéricas de computação' +body = ''' +O NumPy oferece funções matemáticas abrangentes, geradores de números aleatórios, rotinas de álgebra linear, transformações de Fourier e muito mais. +''' +{{< /card >}} + +{{< card >}} +title = 'Código aberto' +body = ''' +Distribuído sob uma licença liberal [BSD](https://github.com/numpy/numpy/blob/main/LICENSE.txt), o NumPy é desenvolvido e mantido [publicamente no GitHub](https://github.com/numpy/numpy) por uma [comunidade](/community) vibrante, responsiva e diversa. +''' +{{< /card >}} + +{{< card >}} +title = 'Interoperável' +body = ''' +O NumPy suporta uma ampla gama de plataformas de hardware e computação, e se integra bem com bibliotecas distribuídas, de GPU e de arrays esparsos. +''' +{{< /card >}} + +{{< card >}} +title = 'Eficiente' +body = ''' +O núcleo do NumPy é composto por código C bem otimizado. Aproveite a flexibilidade do Python com a velocidade de código compilado. +''' +{{< /card >}} + +{{< card >}} +title = 'Fácil de usar' +body = ''' +A sintaxe de alto nível do NumPy o torna acessível e produtivo para programadores de qualquer formação ou nível de experiência. +''' +{{< /card >}} + +{{< /grid >}} diff --git a/content/pt/about.md b/content/pt/about.md index fd3fb1217a..30e0dbdd13 100644 --- a/content/pt/about.md +++ b/content/pt/about.md @@ -3,11 +3,9 @@ title: Quem Somos sidebar: false --- -_Algumas informações sobre o projeto NumPy e a comunidade_ +NumPy é um projeto de código aberto que visa possibilitar a computação numérica com Python. Foi criado em 2005, com base no trabalho inicial das bibliotecas Numeric e Numarray. O NumPy sempre será 100% software de código aberto, livre para que todos usem. É lançado sob os termos liberais da [licença BSD modificada](https://github.com/numpy/numpy/blob/main/LICENSE.txt). -NumPy é um projeto de código aberto visando habilitar a computação numérica com Python. Foi criado em 2005, com base no trabalho inicial das bibliotecas Numeric e Numarray. O NumPy sempre será um software 100% de código aberto, livre para que todos usem e disponibilizados sob os termos liberais da [licença BSD modificada](https://github.com/numpy/numpy/blob/main/LICENSE.txt). - -O NumPy é desenvolvido no GitHub, através do consenso da comunidade NumPy e de uma comunidade científica em Python mais ampla. Para obter mais informações sobre nossa abordagem de governança, por favor, consulte nosso [Documento de Governança](https://www.numpy.org/devdocs/dev/governance/index.html). +O NumPy é desenvolvido no GitHub, através do consenso da comunidade NumPy e de uma comunidade mais ampla de Python científico. Para obter mais informações sobre nossa abordagem de governança, por favor, consulte nosso [Documento de Governança](https://www.numpy.org/devdocs/dev/governance/index.html). ## Conselho Diretor (Steering Council) @@ -35,17 +33,31 @@ Membros Eméritos: - Jaime Fernández del Río (2014-2021) - Pauli Virtanen (2008-2021) +Para entrar em contato com o conselho diretor do NumPy, por favor envie um email para numpy-team@googlegroups.com. + ## Times -O projeto NumPy está crescendo; temos equipes para +A liderança do projeto NumPy trabalha ativamente na diversificação dos caminhos possíveis para contribuições.
Atualmente, o NumPy conta com os seguintes times: -- código +- desenvolvimento - documentação -- website - triagem +- website +- pesquisa +- traduções +- mentores para sprints de desenvolvimento +- otimização - financiamento e bolsas -Veja a página de [Times](/teams/) para membros individuais de cada time. +Veja a página sobre os [Times]({{< relref "/teams" >}}) para mais informações. + +## Subcomitê NumFOCUS + +- Charles Harris +- Ralf Gommers +- Melissa Weber Mendonça +- Sebastian Berg +- Membro externo: Thomas Caswell ## Patrocinadores @@ -56,6 +68,11 @@ O NumPy recebe financiamento direto das seguintes fontes: ## Parceiros Institucionais Os Parceiros Institucionais são organizações que apoiam o projeto, empregando pessoas que contribuem para a NumPy como parte de seu trabalho. Os parceiros institucionais atuais incluem: + +- UC Berkeley (Stéfan van der Walt) +- Quansight (Nathan Goldbaum, Ralf Gommers, Matti Picus, Melissa Weber Mendonça) +- NVIDIA (Sebastian Berg) + {{< partners >}} @@ -68,4 +85,6 @@ NumPy é um Projeto Patrocinado da NumFOCUS, uma instituição de caridade sem f Doações para o NumPy são gerenciadas pela [NumFOCUS](https://numfocus.org). Para doadores nos Estados Unidos, sua doação é dedutível para fins fiscais na medida oferecida pela lei. Como em qualquer doação, você deve consultar seu conselheiro fiscal sobre sua situação fiscal em particular. O Conselho Diretor da NumPy tomará as decisões sobre a melhor forma de utilizar os fundos recebidos. Prioridades técnicas e de infraestrutura estão documentadas no [NumPy Roadmap](https://www.numpy.org/neps/index.html#roadmap). -{{< numfocus >}} + +{{}} + diff --git a/content/pt/case-studies/blackhole-image.md b/content/pt/case-studies/blackhole-image.md index d84528f550..3f4031312d 100644 --- a/content/pt/case-studies/blackhole-image.md +++ b/content/pt/case-studies/blackhole-image.md @@ -4,11 +4,13 @@ sidebar: false --- {{< figure src="/images/content_images/cs/blackhole.jpg" caption="**Black Hole M87**" alt="black hole image" attr="*(Créditos: Event Horizon Telescope Collaboration)*" attrlink="https://www.jpl.nasa.gov/images/universe/20190410/blackhole20190410.jpg" >}} +src = '/images/content_images/cs/blackhole.jpg' title = 'Buraco Negro M87' alt = 'imagem do buraco negro' attribution = '(Créditos de Imagem: Event Horizon Telescope Collaboration)' attributionlink = 'https://www.jpl.nasa.gov/images/universe/20190410/blackhole20190410.jpg' +{{< /figure >}} -
-

Criar uma imagem do Buraco Negro M87 é como tentar ver algo que, por definição, é impossível de se ver.

-
Katie Bouman, Professora Assistente, Ciências da Computação e Matemática, Caltech
-
+{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="Katie Bouman, *Assistant Professor, Computing & Mathematical Sciences, Caltech*" +> }} Imaging the M87 Black Hole is like trying to see something that is by definition impossible to see. +> +> {{< /blockquote >}} ## Um telescópio do tamanho da Terra @@ -36,7 +38,9 @@ O [telescópio Event Horizon (EHT)](https://eventhorizontelescope.org), é um co Quando o objetivo é algo que nunca foi visto, como os cientistas podem ter confiança de que sua imagem está correta? -{{< figure src="/images/content_images/cs/dataprocessbh.png" class="csfigcaption" caption="**Etapas de Processamento de Dados do EHT**" alt="data pipeline" align="middle" attr="(Créditos do diagrama: The Astrophysical Journal, Event Horizon Telescope Collaboration)" attrlink="https://iopscience.iop.org/article/10.3847/2041-8213/ab0c57" >}} +{{< figure >}} +src = '/images/content_images/cs/dataprocessbh.png' title = 'EHT Pipeline de Processamento de Dados' alt = 'pipeline de dados' align = 'center' attribution = '(Créditos do Diagrama: The Astrophysical Journal, Event Horizon Telescope Collaboration)' attributionlink = 'https://iopscience.iop.org/article/10.3847/2041-8213/ab0c57' +{{< figure src="/images/content_images/cs/bh_numpy_role.png" class="fig-center" alt="role of numpy" caption="**O papel do NumPy na criação da primeira imagem de um Buraco Negro**" >}} ## O papel do NumPy @@ -46,11 +50,15 @@ A colaboração do EHT venceu esses desafios ao estabelecer equipes independente O trabalho desse grupo ilustra o papel do ecossistema científico do Python no avanço da ciência através da análise de dados colaborativa. -{{< figure src="/images/content_images/cs/bh_numpy_role.png" class="fig-center" alt="role of numpy" caption="**O papel do NumPy na criação da primeira imagem de um Buraco Negro**" >}} +{{< figure >}} +{{< figure src="/images/content_images/cs/numpy_bh_benefits.png" class="fig-center" alt="numpy benefits" caption="**Funcionalidades-chave do NumPy utilizadas**" >}} +{{< /figure >}} Por exemplo, o pacote Python [`eht-imaging`][ehtim] fornece ferramentas para simular e realizar reconstrução de imagem nos dados do VLBI. O NumPy está no coração do processamento de dados vetoriais usado neste pacote, como ilustrado pelo gráfico parcial de dependências de software abaixo. -{{< figure src="/images/content_images/cs/ehtim_numpy.png" class="fig-center" alt="ehtim dependency map highlighting numpy" caption="**Diagrama de dependência de software do pacote ehtim evidenciando o NumPy**" >}} +{{< figure >}} +src = '/images/content_images/cs/ehtim_numpy.png' alt = 'mapa de dependência ehtim destacando o numpy' title = 'Gráfico de dependência de software do pacote ehtim destacando o NumPy' +{{< /figure >}} Além do NumPy, muitos outros pacotes como [SciPy](https://www.scipy.org) e [Pandas](https://pandas.io) foram usados na *pipeline* de processamento de dados para criar a imagem do buraco negro. Os arquivos astronômicos de formato padrão e transformações de tempo/coordenadas foram tratados pelo [Astropy][astropy] enquanto a [Matplotlib][mpl] foi usada na visualização de dados em todas as etapas de análise, incluindo a geração da imagem final do buraco negro. @@ -58,7 +66,9 @@ Além do NumPy, muitos outros pacotes como [SciPy](https://www.scipy.org) e [Pan A estrutura de dados n-dimensional que é a funcionalidade central do NumPy permitiu aos pesquisadores manipular grandes conjuntos de dados, fornecendo a base para a primeira imagem de um buraco negro. Esse momento marcante na ciência fornece evidências visuais impressionantes para a teoria de Einstein. Esta conquista abrange não apenas avanços tecnológicos, mas colaboração científica em escala internacional entre mais de 200 cientistas e alguns dos melhores observatórios de rádio do mundo. Eles usaram algoritmos e técnicas de processamento de dados inovadores, que aperfeiçoaram os modelos astronômicos existentes, para ajudar a descobrir um dos mistérios do universo. -{{< figure src="/images/content_images/cs/numpy_bh_benefits.png" class="fig-center" alt="numpy benefits" caption="**Funcionalidades-chave do NumPy utilizadas**" >}} +{{< figure >}} +src = '/images/content_images/cs/numpy_bh_benefits.png' alt = 'benefícios do numpy' title = 'Principais recursos do NumPy utilizados' +{{< /figure >}} [resolution]: https://eventhorizontelescope.org/press-release-april-10-2019-astronomers-capture-first-image-black-hole diff --git a/content/pt/case-studies/cricket-analytics.md b/content/pt/case-studies/cricket-analytics.md index c2fbcac9c8..e7140e6c26 100644 --- a/content/pt/case-studies/cricket-analytics.md +++ b/content/pt/case-studies/cricket-analytics.md @@ -4,11 +4,13 @@ sidebar: false --- {{< figure src="/images/content_images/cs/ipl-stadium.png" caption="**IPLT20, o maior festival de Críquete da Índia**" alt="Copa e estádio da Indian Premier League Cricket" attr="*(Image credits: IPLT20 (cup and logo) & Akash Yadav (stadium))*" attrlink="https://unsplash.com/@aksh1802" >}} +src = '/images/content_images/cs/ipl-stadium.png' title = 'IPLT20, o maior Festival de Críquete da Índia' alt = 'Taça e estádio de críquete da Premier League indiana' attribution = '(Créditos de imagem: IPLT20 (taça e logo) & Akash Yadav (estádio))' attributionlink = 'https://unsplash.com/@aksh1802' +{{< /figure >}} -
-

Você não joga para a torcida, joga para o país.

-
—M S Dhoni, Jogador Internacional de Críquete, ex-capitão, Time Indiano, joga pelo Chennai Super Kings na IPL
-
+{{< blockquote cite="https://www.scoopwhoop.com/sports/ms-dhoni/" by="M S Dhoni, *International Cricket Player, ex-captain, Indian Team, plays for Chennai Super Kings in IPL*" +> }} You don't play for the crowd, you play for the country. +> +> {{< /blockquote >}} ## Sobre Críquete @@ -16,7 +18,7 @@ Dizer que os indianos adoram o críquete seria subestimar este sentimento. O jog A Primeira Liga Indiana (*Indian Premier League* - IPL) é uma liga profissional de críquete [Twenty20](https://pt.wikipedia.org/wiki/Twenty20), fundada em 2008. É um dos eventos de críquete mais assistidos no mundo, avaliado em [$6,7 bilhões de dólares](https://en.wikipedia.org/wiki/Indian_Premier_League) em 2019. -Críquete é um jogo dominado pelos números - as corridas executadas por um batsman, os wickets perdidos por um boleador, as partidas ganhas por uma equipe de críquete, o número de vezes que um batsman responde de certa maneira a um tipo de arremesso do boleador, etc. A capacidade de investigar números de críquete para melhorar o desempenho e estudar as oportunidades de negócio, mercado e economia de críquete através de poderosas ferramentas de análise, alimentadas por softwares numéricos de computação, como o NumPy, é um grande negócio. As análises de críquete fornecem informações interessantes sobre o jogo e informações preditivas sobre os resultados do jogo. +perdidos por um boleador, as partidas ganhas por uma equipe de críquete, o número de vezes que um batsman responde de certa maneira a um tipo de arremesso do boleador, etc. A capacidade de investigar números de críquete para melhorar o desempenho e estudar as oportunidades de negócio, mercado e economia de críquete através de poderosas ferramentas de análise, alimentadas por softwares numéricos de computação, como o NumPy, é um grande negócio. A capacidade de investigar estatísticas do críquete para melhorar a performance dos times e estudar oportunidades de negócios, o mercado em si, e a economia do críquete através de ferramentas de análise poderosas alimentadas por softwares de computação numérica como o NumPy é um grande negócio. As análises de críquete fornecem informações interessantes sobre o jogo e informações preditivas sobre os resultados do jogo. Hoje, existem conjuntos ricos e quase infinitos de estatísticas e informações sobre jogos de críquete, por exemplo, [ESPN cricinfo](https://stats.espncricinfo.com/ci/engine/stats/index.html) e [cricsheet](https://cricsheet.org). Estes e muitos outros bancos de dados de críquete foram usados para [análise de críquete](https://www.researchgate.net/publication/336886516_Data_visualization_and_toss_related_analysis_of_IPL_teams_and_batsmen_performances) usando os mais modernos algoritmos de aprendizagem de máquina e modelagem preditiva. Plataformas de mídia e entretenimento, juntamente com entidades de esporte profissionais associadas ao jogo usam tecnologia e análise para determinar métricas chave para melhorar as chances de vitória: @@ -25,7 +27,9 @@ Hoje, existem conjuntos ricos e quase infinitos de estatísticas e informações * ganho de informações sobre desempenho e condição física de um determinado jogador contra determinado adversário, * contribuições dos jogadores para vitórias e derrotas para a tomada de decisões estratégicas na composição do time -{{< figure src="/images/content_images/cs/cricket-pitch.png" class="csfigcaption" caption="**Pitch de críquete, o ponto focal do campo**" alt="Um pitch de críquete com um boleador e batsmen" align="middle" attr="*(Créditos de imagem: Debarghya Das)*" attrlink="http://debarghyadas.com/files/IPLpaper.pdf" >}} +{{< figure >}} +src = '/images/content_images/cs/cricket-pitch.png' title = 'Campo de críquete, o ponto focal no campo' alt = 'Campo de críquete com lançador e batedor' align = 'center' attribution = '(Crédito de imagem: Debarghya Das)' attributionlink = 'http://debarghyadas.com/files/IPLpaper.pdf' +{{< figure src="/images/content_images/cs/numpy_ca_benefits.png" class="fig-center" alt="Diagrama mostrando os benefícios de usar a NumPy para análise de críquete" caption="**Recursos principais da NumPy utilizados**" >}} ### Objetivos Principais da Análise de Dados @@ -33,7 +37,9 @@ Hoje, existem conjuntos ricos e quase infinitos de estatísticas e informações * A análise de dados em tempo real pode ajudar a obtenção de informações mesmo durante o jogo para orientar mudanças nas táticas da equipe e dos negócios associados para benefícios e crescimento econômicos. * Além da análise histórica, os modelos preditivos explorados para determinar os possíveis resultados das partidas requerem um conhecimento significativo sobre processamento numérico e ciência de dados, ferramentas de visualização e a possibilidade de incluir observações mais recentes na análise. -{{< figure src="/images/content_images/cs/player-pose-estimator.png" class="fig-center" alt="estimador de postura" caption="**Estimador de Postura de Críquete**" attr="*(Créditos de imagem: connect.vin)*" attrlink="https://connect.vin/2019/05/ai-for-cricket-batsman-pose-analysis/" >}} +{{< figure >}} +src = '/images/content_images/cs/player-pose-estimator.png' alt = 'estimador de poses' title = 'Estimador de poses no críquete' attribution = '(Créditos da imagem: connect.vin)' attributionlink = 'https://connect.vin/2019/05/ai-for-cricket-batsman-pose-analysis/' +{{< /figure >}} ### Desafios @@ -49,7 +55,7 @@ Hoje, existem conjuntos ricos e quase infinitos de estatísticas e informações Muito da tomada de decisões em críquete se baseia em questões como "com que frequência um batsman joga um certo tipo de lance se a recepção da bola for de um determinado tipo", ou "como um boleador muda a direção e alcance da sua jogada se o batsman responder de uma certa maneira". Esse tipo de consulta de análise preditiva requer a disponibilidade de conjuntos de dados altamente granulares e a capacidade de sintetizar dados e criar modelos generativos que sejam altamente precisos. -## Papel da NumPy na Análise de Críquete +## O papel do NumPy na análise de críquete A análise de dados esportivos é um campo próspero. Muitos pesquisadores e empresas [usam NumPy](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx) e outros pacotes PyData como Scikit-learn, SciPy, Matplotlib, e Jupyter, além de usar as últimas técnicas de aprendizagem de máquina e IA. O NumPy foi usado para vários tipos de análise esportiva relacionada a críquete, como: @@ -61,4 +67,6 @@ A análise de dados esportivos é um campo próspero. Muitos pesquisadores e emp A análise de dados esportivos é revolucionária quando se trata de como os jogos profissionais são jogados, especialmente se consideramos como acontece a tomada de decisões estratégicas, que até pouco tempo era principalmente feita com base na "intuição" ou adesão a tradições passadas. O NumPy forma uma fundação sólida para um grande conjunto de pacotes Python que fornecem funções de alto nível relacionadas à análise de dados, aprendizagem de máquina e algoritmos de IA. Estes pacotes são amplamente implantados para se obter informações em tempo real que ajudam na tomada de decisão para resultados decisivos, tanto em campo como para se derivar inferências e orientar negócios em torno do jogo de críquete. Encontrar os parâmetros ocultos, padrões, e atributos que levam ao resultado de uma partida de críquete ajuda os envolvidos a tomar nota das percepções do jogo que estariam de outra forma ocultas nos números e estatísticas. -{{< figure src="/images/content_images/cs/numpy_ca_benefits.png" class="fig-center" alt="Diagrama mostrando os benefícios de usar a NumPy para análise de críquete" caption="**Recursos principais da NumPy utilizados**" >}} +{{< figure >}} +src = '/images/content_images/cs/numpy_ca_benefits.png' alt = 'Diagrama mostrando os benefícios do uso do NumPy para análise de críquete' title = 'Principais recursos do NumPy utilizados' +{{< /figure >}} diff --git a/content/pt/case-studies/deeplabcut-dnn.md b/content/pt/case-studies/deeplabcut-dnn.md index 1dd02b9f92..299c321f1a 100644 --- a/content/pt/case-studies/deeplabcut-dnn.md +++ b/content/pt/case-studies/deeplabcut-dnn.md @@ -4,11 +4,13 @@ sidebar: false --- {{< figure src="/images/content_images/cs/mice-hand.gif" class="fig-center" caption="**Análise de movimentos de mãos de camundongos usando DeepLapCut**" alt="micehandanim" attr="*(Fonte: www.deeplabcut.org )*" attrlink="http://www.mousemotorlab.org/deeplabcut">}} +src = '/images/content_images/cs/mice-hand.gif' title = 'Analisando o movimento das mãos de ratos usando DeepLapCut' alt = 'micehandanim' attribution = '(Fonte: www.deeplabcut.org )' attributionlink = 'http://www.mousemotorlab.org/deeplabcut' +{{< /figure >}} -
-

Software de código aberto está acelerando a Biomedicina. DeepLabCut permite a análise automática de vídeos de comportamento animal usando Deep Learning.

-
—Alexander Mathis, Professor Assistente, École polytechnique fédérale de Lausanne (EPFL)
-
+{{< blockquote cite="https://news.harvard.edu/gazette/story/newsplus/harvard-researchers-awarded-czi-open-source-award/" by="Alexander Mathis, *Assistant Professor, École polytechnique fédérale de Lausanne* ([EPFL](https://www.epfl.ch/en/))" +> }} Open Source Software is accelerating Biomedicine. DeepLabCut permite a análise automática de vídeos de comportamento animal usando Deep Learning. +> +> {{< /blockquote >}} ## Sobre o DeepLabCut @@ -16,7 +18,9 @@ sidebar: false Várias áreas de pesquisa, incluindo a neurociência, a medicina e a biomecânica, utilizam dados de rastreamento da movimentação de animais. A DeepLabCut ajuda a compreender o que os seres humanos e outros animais estão fazendo, analisando ações que foram registradas em vídeo. Ao usar automação para tarefas trabalhosas de monitoramento e marcação, junto com análise de dados baseada em redes neurais profundas, a DeepLabCut garante que estudos científicos envolvendo a observação de animais como primatas, camundongos, peixes, moscas etc. sejam mais rápidos e precisos. -{{< figure src="/images/content_images/cs/race-horse.gif" class="fig-center" caption="**Pontos coloridos rastreiam as posições das partes do corpo de um cavalo de corrida**" alt="horserideranim" attr="*(Fonte: Mackenzie Mathis)*">}} +{{< figure >}} +src = '/images/content_images/cs/race-horse.gif' title = ‘Pontos coloridos rastreiam as posições de partes do corpo de um cavalo de corrida’ alt = 'horserideranim' attribution = '(Fonte: Mackenzie Mathis)' +{{< /figure >}} O rastreamento não invasivo dos animais pela DeepLabCut através da extração de poses é crucial para pesquisas científicas em domínios como a biomecânica, genética, etologia e neurociência. Medir as poses dos animais de maneira não invasiva através de vídeo - sem marcadores - com fundos dinâmicos é computacionalmente desafiador, tanto tecnicamente quanto em termos de recursos e dados de treinamento necessários. @@ -45,7 +49,9 @@ Recentemente, foi introduzido o [modelo DeepLabCut zoo](http://www.mousemotorlab - código para inferência em larga escala em vídeos - inferências de desenho usando ferramentas integradas de visualização -{{< figure src="/images/content_images/cs/deeplabcut-toolkit-steps.png" class="csfigcaption" caption="**Passos na estimação de poses com DeepLabCut**" alt="dlcsteps" align="middle" attr="(Fonte: DeepLabCut)" attrlink="https://twitter.com/DeepLabCut/status/1198046918284210176/photo/1" >}} +{{< figure >}} +{{< figure src="/images/content_images/cs/deeplabcut-workflow.png" class="fig-center" caption="**Fluxo de dados DeepLabCut**" alt="workflow" attr="*(Fonte: Mackenzie Mathis)*" attrlink="https://www.researchgate.net/figure/DeepLabCut-work-flow-The-diagram-delineates-the-work-flow-as-well-as-the-directory-and_fig1_329185962">}} +{{< /figure >}} ### Desafios @@ -61,7 +67,9 @@ Recentemente, foi introduzido o [modelo DeepLabCut zoo](http://www.mousemotorlab Por último, mas não menos importante, manipulação de matrizes - processar grandes conjuntos de matrizes correspondentes a várias imagens, tensores alvo e pontos-chave é bastante desafiador. -{{< figure src="/images/content_images/cs/pose-estimation.png" class="csfigcaption" caption="**Estimação de poses e complexidade**" alt="challengesfig" align="middle" attr="(Fonte: Mackenzie Mathis)" attrlink="https://www.biorxiv.org/content/10.1101/476531v1.full.pdf" >}} +{{< figure src="/images/content_images/cs/numpy_dlc_benefits.png" class="fig-center" alt="numpy benefits" caption="**Recursos chave do NumPy utilizados**" >}} +src = '/images/content_images/cs/pose-estimation.png' title = 'Variedade e complexidade da estimativa de pose' alt = 'challengesfig' align = 'center' attribution = '(Fonte: Mackenzie Mathis)' attributionlink = 'https://www.biorxiv.org/content/10.1101/476531v1.full.pdf' +{{< /figure >}} ## O papel da NumPy nos desafios da estimação de poses @@ -77,13 +85,17 @@ As seguintes características da NumPy desempenharam um papel fundamental para a A DeepLabCut utiliza as capacidades de manipulação de arrays da NumPy em todo o fluxo de trabalho oferecido pelo seu conjunto de ferramentas. Em particular, a NumPy é usada para amostragem de quadros distintos para serem rotulados com anotações humanas e para escrita, edição e processamento de dados de anotação. Dentro da TensorFlow, a rede neural é treinada pela tecnologia DeepLabCut em milhares de iterações para prever as anotações verdadeiras dos quadros. Para este propósito, densidades de alvo (*scoremaps*) são criadas para colocar a estimativa como um problema de tradução de imagem a imagem. Para tornar as redes neurais robustas, o aumento de dados é empregado, o que requer o cálculo de scoremaps alvo sujeitos a várias etapas geométricas e de processamento de imagem. Para tornar o treinamento rápido, os recursos de vectorização da NumPy são utilizados. Para inferência, as previsões mais prováveis de scoremaps alvo precisam ser extraídas e é necessário "vincular previsões para montar animais individuais" de maneira eficiente. -{{< figure src="/images/content_images/cs/deeplabcut-workflow.png" class="fig-center" caption="**Fluxo de dados DeepLabCut**" alt="workflow" attr="*(Fonte: Mackenzie Mathis)*" attrlink="https://www.researchgate.net/figure/DeepLabCut-work-flow-The-diagram-delineates-the-work-flow-as-well-as-the-directory-and_fig1_329185962">}} +{{< figure >}} +src = '/images/content_images/cs/deeplabcut-workflow.png' title = 'Fluxo de trabalho do DeepLabCut' alt = 'workflow' attribution = '(Fonte: Mackenzie Mathis)' attributionlink = 'https://www.researchgate.net/figure/DeepLabCut-work-flow-The-diagram-delineates-the-work-flow-as-well-as-the-directory-and_fig1_329185962' +{{< /figure >}} ## Resumo Observação e descrição eficiente do comportamento é uma peça fundamental da etologia, neurociência, medicina e tecnologia modernas. [DeepLabCut](http://orga.cvss.cc/wp-content/uploads/2019/05/NathMathis2019.pdf) permite que os pesquisadores estimem a pose do sujeito, permitindo efetivamente que o seu comportamento seja quantificado. Com apenas um pequeno conjunto de imagens de treinamento, o conjunto de ferramentas em Python da DeepLabCut permite treinar uma rede neural tão precisa quanto a rotulagem humana, expandindo assim sua aplicação para não só análise de comportamento dentro do laboratório, mas também potencialmente em esportes, análise de locomoção, medicina e estudos sobre reabilitação. Desafios complexos em combinatória e processamento de dados enfrentados pelos algoritmos da DeepLabCut são tratados através do uso de recursos de manipulação de matriz do NumPy. -{{< figure src="/images/content_images/cs/numpy_dlc_benefits.png" class="fig-center" alt="numpy benefits" caption="**Recursos chave do NumPy utilizados**" >}} +{{< figure >}} +src = '/images/content_images/cs/numpy_dlc_benefits.png' alt = 'benefícios do numpy' title = 'Principais recursos do NumPy utilizados' +{{< /figure >}} [cheetah-movement]: https://www.technologynetworks.com/neuroscience/articles/interview-a-deeper-cut-into-behavior-with-mackenzie-mathis-327618 diff --git a/content/pt/case-studies/gw-discov.md b/content/pt/case-studies/gw-discov.md index 6de7efed50..bf4be55545 100644 --- a/content/pt/case-studies/gw-discov.md +++ b/content/pt/case-studies/gw-discov.md @@ -4,17 +4,17 @@ sidebar: false --- {{< figure src="/images/content_images/cs/gw_sxs_image.png" class="fig-center" caption="**Ondas gravitacionais**" alt="binary coalesce black hole generating gravitational waves" attr="*(Créditos de imagem: O projeto Simulating eXtreme Spacetimes (SXS) no LIGO)*" attrlink="https://youtu.be/Zt8Z_uzG71o" >}} +src = '/images/content_images/cs/gw_sxs_image.png' title = 'Ondas Gravitacionais' alt = 'buraco negro coalesce binário gerando ondas gravitacionais' attribution = '(Créditos da Imagem: The Simulating eXtreme Spacetimes (SXS) Project at LIGO)' attributionlink = 'https://youtu.be/Zt8Z_uzG71o' +{{< /figure >}} -
-

O ecossistema científico Python é uma infraestrutura crítica para a pesquisa feita no LIGO.

-
David Shoemaker, Colaborador Científico no LIGO
-
+{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="David Shoemaker, *LIGO Scientific Collaboration*" >}} The scientific Python ecosystem is critical infrastructure for the research done at LIGO. +{{< /blockquote >}} ## Sobre [Ondas Gravitacionais](https://www.nationalgeographic.com/news/2017/10/what-are-gravitational-waves-ligo-astronomy-science/) e o [LIGO](https://www.ligo.caltech.edu) -Ondas gravitacionais são ondulações no tecido espaço-tempo, gerado por eventos cataclísmicos no universo, como colisão e fusão de dois buracos negros ou a coalescência de estrelas binárias ou supernovas. A observação de ondas gravitacionais pode ajudar não só no estudo da gravidade, mas também no entendimento de alguns dos fenômenos obscuros existentes no universo distante e seu impacto. +Gravitational waves are ripples in the fabric of space and time, generated by cataclysmic events in the universe such as collision and merging of two black holes or coalescing binary stars or supernovae. Observing GW can not only help in studying gravity but also in understanding some of the obscure phenomena in the distant universe and its impact. -O [Observatório Interferômetro Laser de Ondas Gravitacionais (LIGO)](https://www.ligo.caltech.edu) foi projetado para abrir o campo da astrofísica das ondas gravitacionais através da detecção direta de ondas gravitacionais previstas pela Teoria Geral da Relatividade de Einstein. O observatório consiste de dois interferômetros amplamente separados dentro dos Estados Unidos - um em Hanford, Washington e o outro em Livingston, Louisiana — operando em uníssono para detectar ondas gravitacionais. Cada um deles tem detectores em escala quilométrica de ondas gravitacionais que usam interferometria laser. A Colaboração Científica LIGO (LSC), é um grupo de mais de 1000 cientistas de universidades dos Estados Unidos e em 14 outros países apoiados por mais de 90 universidades e institutos de pesquisa; aproximadamente 250 estudantes contribuem ativamente com a colaboração. A nova descoberta do LIGO é a primeira observação de ondas gravitacionais em si, feita medindo os pequenos distúrbios que as ondas fazem ao espaço-tempo enquanto atravessam a Terra. A descoberta abriu novas fronteiras astrofísicas que exploram o lado "curvado" do universo - objetos e fenômenos que são feitos a partir da curvatura do espaço-tempo. +Ondas gravitacionais emitidas da fusão não podem ser calculadas usando nenhuma técnica a não ser relatividade numérica por força bruta usando supercomputadores. A quantidade de dados que o LIGO coleta é imensa tanto quanto os sinais de ondas gravitacionais são pequenos. Each of them has multi-kilometer-scale gravitational wave detectors that use laser interferometry. The LIGO Scientific Collaboration (LSC), is a group of more than 1000 scientists from universities around the United States and in 14 other countries supported by more than 90 universities and research institutes; approximately 250 students actively contributing to the collaboration. The new LIGO discovery is the first observation of gravitational waves themselves, made by measuring the tiny disturbances the waves make to space and time as they pass through the earth. It has opened up new astrophysical frontiers that explore the warped side of the universe—objects and phenomena that are made from warped spacetime. ### Objetivos @@ -39,13 +39,15 @@ O [Observatório Interferômetro Laser de Ondas Gravitacionais (LIGO)](https://w Uma vez que os obstáculos relacionados a compreender as equações de Einstein bem o suficiente para resolvê-las usando supercomputadores foram ultrapassados, o próximo grande desafio era tornar os dados compreensíveis para o cérebro humano. A modelagem de simulações, assim como a detecção de sinais, exigem técnicas de visualização efetiva. A visualização também desempenha um papel de fornecer mais credibilidade à relatividade numérica aos olhos dos aficionados pela ciência pura, que não dão importância suficiente à relatividade numérica até que a imagem e as simulações tornem mais fácil a compreensão dos resultados para um público maior. A velocidade da computação complexa, e da renderização, re-renderização de imagens e simulações usando as últimas entradas e informações experimentais pode ser uma atividade demorada que desafia pesquisadores neste domínio. -{{< figure src="/images/content_images/cs/gw_strain_amplitude.png" class="fig-center" alt="gravitational waves strain amplitude" caption="**Amplitude estimada da deformação das ondas gravitacionais do evento GW150914**" attr="(**Créditos do gráfico:** Observation of Gravitational Waves from a Binary Black Hole Merger, ResearchGate Publication)" attrlink="https://www.researchgate.net/publication/293886905_Observation_of_Gravitational_Waves_from_a_Binary_Black_Hole_Merger" >}} +{{< figure >}} +src = '/images/content_images/cs/gw_strain_amplitude.png' alt = 'gravitational waves strain amplitude' title = 'Estimated gravitational-wave strain amplitude from GW150914' attribution = '(Graph Credits: Observation of Gravitational Waves from a Binary Black Hole Merger, ResearchGate Publication)' attributionlink = 'https://www.researchgate.net/publication/293886905_Observation_of_Gravitational_Waves_from_a_Binary_Black_Hole_Merger' +{{< figure src="/images/content_images/cs/numpy_gw_benefits.png" class="fig-center" alt="numpy benefits" caption="**Recursos chave da NumPy utilizados**" >}} ## O papel da NumPy na detecção de ondas gravitacionais -Ondas gravitacionais emitidas da fusão não podem ser calculadas usando nenhuma técnica a não ser relatividade numérica por força bruta usando supercomputadores. A quantidade de dados que o LIGO coleta é imensa tanto quanto os sinais de ondas gravitacionais são pequenos. +Gravitational waves emitted from the merger cannot be computed using any technique except brute force numerical relativity using supercomputers. The amount of data LIGO collects is as incomprehensibly large as gravitational wave signals are small. -NumPy, o pacote padrão de análise numérica para Python, foi parte do software utilizado para várias tarefas executadas durante o projeto de detecção de ondas gravitacionais no LIGO. A NumPy ajudou a resolver problemas matemáticos e de manipulação de dados complexos em alta velocidade. Aqui estão alguns exemplos: +NumPy, the standard numerical analysis package for Python, was utilized by the software used for various tasks performed during the GW detection project at LIGO. NumPy helped in solving complex maths and data manipulation at high speed. Here are some examples: * [Processamento de sinais](https://www.uv.es/virgogroup/Denoising_ROF.html): Detecção de falhas, [Identificação de ruídos e caracterização de dados](https://ep2016.europython.eu/media/conference/slides/pyhton-in-gravitational-waves-research-communities.pdf) (NumPy, scikit-learn, scipy, matplotlib, pandas, PyCharm) * Recuperação de dados: Decidir quais dados podem ser analisados, compreender se os dados contém um sinal - como uma agulha em um palheiro @@ -56,14 +58,20 @@ NumPy, o pacote padrão de análise numérica para Python, foi parte do software * Cálculo de correlações * [Software](https://github.com/lscsoft) fundamental desenvolvido na análise de ondas gravitacionais, como [GwPy](https://gwpy.github.io/docs/stable/overview.html) e [PyCBC](https://pycbc.org) usam NumPy e AstroPy internamente para fornecer interfaces baseadas em objetos para utilidades, ferramentas e métodos para o estudo de dados de detectores de ondas gravitacionais. -{{< figure src="/images/content_images/cs/gwpy-numpy-dep-graph.png" class="fig-center" alt="gwpy-numpy depgraph" caption="**Grafo de dependências mostrando como o pacote GwPy depended da NumPy**" >}} +{{< figure >}} +src = '/images/content_images/cs/gwpy-numpy-dep-graph.png' alt = 'gwpy-numpy depgraph' title = 'Gráfico de dependência mostrando como o pacote GwPy depende do NumPy' +{{< /figure >}} ---- -{{< figure src="/images/content_images/cs/PyCBC-numpy-dep-graph.png" class="fig-center" alt="PyCBC-numpy depgraph" caption="**Grafo de dependências mostrando como o pacote PyCBC depended da NumPy**" >}} +{{< figure >}} +src = '/images/content_images/cs/PyCBC-numpy-dep-graph.png' alt = 'PyCBC-numpy depgraph' title = 'Gráfico de dependência mostrando como o pacote PyCBC depende do NumPy' +{{< /figure >}} ## Resumo -A detecção de ondas gravitacionais permitiu que pesquisadores descobrissem fenômenos totalmente inesperados ao mesmo tempo em que proporcionaram novas idéias sobre muitos dos fenômenos mais profundos conhecidos na astrofísica. O processamento e a visualização de dados é um passo crucial que ajuda cientistas a obter informações coletadas de observações científicas e a entender os resultados. Os cálculos são complexos e não podem ser compreendidos por humanos a não ser que sejam visualizados usando simulações de computador que são alimentadas com dados e análises reais observados. A NumPy, junto com outras bibliotecas Python, como matplotlib, pandas, e scikit-learn [permitem que pesquisadores](https://www.gw-openscience.org/events/GW150914/) respondam perguntas complexas e descubram novos horizontes em nossa compreensão do universo. +GW detection has enabled researchers to discover entirely unexpected phenomena while providing new insight into many of the most profound astrophysical phenomena known. Number crunching and data visualization is a crucial step that helps scientists gain insights into data gathered from the scientific observations and understand the results. The computations are complex and cannot be comprehended by humans unless it is visualized using computer simulations that are fed with the real observed data and analysis. NumPy along with other Python packages such as matplotlib, pandas, and scikit-learn is [enabling researchers](https://www.gw-openscience.org/events/GW150914/) to answer complex questions and discover new horizons in our understanding of the universe. -{{< figure src="/images/content_images/cs/numpy_gw_benefits.png" class="fig-center" alt="numpy benefits" caption="**Recursos chave da NumPy utilizados**" >}} +{{< figure >}} +src = '/images/content_images/cs/numpy_gw_benefits.png' alt = 'benefícios do numpy' title = 'Principais recursos do NumPy utilizados' +{{< /figure >}} diff --git a/content/pt/citing-numpy.md b/content/pt/citing-numpy.md index f6f9567450..f947689548 100644 --- a/content/pt/citing-numpy.md +++ b/content/pt/citing-numpy.md @@ -5,7 +5,7 @@ sidebar: false Se a NumPy é importante na sua pesquisa, e você gostaria de dar reconhecimento ao projeto na sua publicação acadêmica, sugerimos citar os seguintes documentos: -* Harris, C.R., Millman, K.J., van der Walt, S.J. et al. _Array programming with NumPy_. Nature 585, 357–362 (2020). DOI: [10.1038/s41586-020-2649-2](https://doi.org/10.1038/s41586-020-2649-2). ([Link da editora](https://www.nature.com/articles/s41586-020-2649-2)). +* Harris, C.R., Millman, K.J., van der Walt, S.J. Harris, C.R., Millman, K.J., van der Walt, S.J. et al. _Array programming with NumPy_. Nature 585, 357–362 (2020). DOI: [10.1038/s41586-020-2649-2](https://doi.org/10.1038/s41586-020-2649-2). ([Link da editora](https://www.nature.com/articles/s41586-020-2649-2)). _Em formato BibTeX:_ diff --git a/content/pt/code-of-conduct.md b/content/pt/code-of-conduct.md index 13ce6cb018..fe68237a92 100644 --- a/content/pt/code-of-conduct.md +++ b/content/pt/code-of-conduct.md @@ -43,7 +43,7 @@ Embora sejamos receptivos às pessoas fluentes em todas as línguas, o desenvolv Padrões de comportamento na comunidade NumPy estão detalhados no Código de Conduta acima. Os participantes da nossa comunidade devem se comportar de acordo com esses padrões em todas as suas interações e ajudar os outros a fazê-lo também (veja a próxima seção). -### Diretrizes de Resposta a Incidentes +### Diretrizes de resposta a incidentes Sabemos que é mais comum do que o desejado que a comunicação na Internet comece ou se transforme em abusos óbvios e flagrantes. Reconhecemos também que, por vezes, as pessoas podem ter um dia ruim, ou não conhecer algumas das orientações deste Código de Conduta. Tenha isto em mente ao decidir como responder a uma violação deste Código. @@ -59,7 +59,7 @@ Atualmente, o comitê é formato por: Se o seu relatório envolve algum membro da comissão, ou se você sentir que existe um conflito de interesses em tratá-lo, então os membros abster-se-ão de considerar o seu relatório. Como alternativa, se por qualquer razão você se sentir desconfortável em fazer um relatório à comissão, então você também pode entrar em contato com a equipe sênior da NumFOCUS em [conduct@numfocus.org](https://numfocus.org/code-of-conduct#persons-responsible). -### Resolução de Incidentes & Execução do Código de Conduta +### Resolução de Incidentes & Aplicação do Código de Conduta _Esta seção resume os pontos mais importantes, mais detalhes podem ser encontrados em_ [Código de Conduta do NumPy - Como dar seguimento a um relatório](/report-handling-manual). @@ -80,4 +80,4 @@ O comitê responderá a qualquer relatório o mais rapidamente possível e, no m Somos gratos aos grupos responsáveis pelos documentos abaixo, dos quais retiramos conteúdo e inspiração: -- [The SciPy Code of Conduct](https://docs.scipy.org/doc/scipy/reference/dev/conduct/code_of_conduct.html) +- [The SciPy Code of Conduct](https://docs.scipy.org/doc/scipy/dev/conduct/code_of_conduct.html) diff --git a/content/pt/community.md b/content/pt/community.md index c6b3115d76..7992ff2fd6 100644 --- a/content/pt/community.md +++ b/content/pt/community.md @@ -3,7 +3,7 @@ title: Comunidade sidebar: false --- -NumPy é um projeto de código aberto impulsionado pela comunidade desenvolvido por um grupo muito diversificado de [contribuidores](/pt/teams/). A liderança da NumPy assumiu um forte compromisso de criar uma comunidade aberta, inclusiva e positiva. Por favor, leia [o Código de Conduta NumPy](/pt/code-of-conduct) para orientações sobre como interagir com os outros de uma forma que faça a comunidade prosperar. +NumPy é um projeto de código aberto impulsionado pela comunidade desenvolvido por um grupo muito diversificado de [contribuidores](/pt/teams/). A liderança do NumPy assumiu um forte compromisso de criar uma comunidade aberta, inclusiva e positiva. Por favor, leia [o Código de Conduta NumPy](/pt/code-of-conduct) para orientações sobre como interagir com os outros de uma forma que faça a comunidade prosperar. Oferecemos vários canais de comunicação para aprender, compartilhar seu conhecimento e se conectar com outros dentro da comunidade NumPy. @@ -63,3 +63,4 @@ Para prosperar, o projeto NumPy precisa de sua experiência e entusiasmo. Não Se você está interessado em se tornar um contribuidor do NumPy (oba!) recomendamos que você confira nossa página sobre [Contribuições](/pt/contribute). +Além disso, sinta-se à vontade para passar por aqui e dizer oi em uma de nossas reuniões da comunidade. Para acompanhá-las, confira nosso calendário de eventos [aqui](https://scientific-python.org/calendars/). diff --git a/content/pt/config.yaml b/content/pt/config.yaml index df96976a9e..fadb422094 100644 --- a/content/pt/config.yaml +++ b/content/pt/config.yaml @@ -1,140 +1,140 @@ -# FOR TRANSLATORS: this is a YAML file, with lines being of the form: -# -# key: value -# -# Please translate the `value`s, not the `key`s! -# Comments (starting with `#`) and `url:` or `link:` lines also do not -# need to be translated. `title:` and `text:` lines do need translation. -# languageName: Português params: description: Por que NumPy? Arrays n-dimensionais poderosas. Ferramentas para computação numérica. Interoperabilidade. Alto desempenho. Código aberto. navbarlogo: image: logo.svg + text: NumPy link: /pt/ - hero: + #Main hero title title: NumPy + #Hero subtitle (optional) subtitle: A biblioteca fundamental para computação científica com Python - buttontext: Comece aqui - buttonlink: "/pt/install" + #Button text + buttontext: "Última versão: NumPy 1.26. Veja todas as versões" + #Where the main hero button links to + buttonlink: "/pt/news/#releases" + #Hero image (from static/images/___) image: logo.svg - - news: - title: NumPy v1.20.0 - content: Suporte a anotações de tipos - Melhorias no desempenho através de SIMD multi-plataformas - url: /pt/news - shell: - title: placeholder # do not translate - + title: placeholder intro: - - title: Try NumPy - text: Use the interactive shell to try NumPy in the browser - - docslink: Don't forget to check out the docs. - + - + title: Experimentar o NumPy + text: Use o shell interativo para testar o NumPy no navegador + docslink: Não se esqueça de conferir a documentação. casestudies: title: ESTUDOS DE CASO features: - - title: A Primeira Imagem de um Buraco Negro - text: Como o NumPy, junto com outras bibliotecas como SciPy e Matplotlib que dependem do NumPy, permitiram ao Event Horizon Telescope gerar a primeira imagem de um buraco negro da história. - img: /images/content_images/case_studies/blackhole.png - alttext: Primeira imagem de um buraco negro. É um círculo laranja em um fundo preto. - url: /pt/case-studies/blackhole-image - - title: Descoberta de Ondas Gravitacionais - text: Em 1916, Albert Einstein previu ondas gravitacionais; 100 anos depois, sua existência foi confirmada pelos cientistas do LIGO usando NumPy. - img: /images/content_images/case_studies/gravitional.png - alttext: Duas esferas orbitando a si mesmas. Elas deslocam a gravidade em seu entorno. - url: /pt/case-studies/gw-discov - - title: Análise Esportiva - text: A análise de críquete está mudando o jogo ao melhorar o desempenho de jogadores e times através de modelagem estatística e análise preditiva. O NumPy possibilita muitas dessas análises. - img: /images/content_images/case_studies/sports.jpg - alttext: Bola de críquete em um campo verde - url: /pt/case-studies/cricket-analytics - - title: Estimação de poses usando deep learning - text: DeepLabCut usa o NumPy para acelerar estudos científicos que envolvem comportamento animal para entender melhor o controle motor em várias espécies e escalas de tempo. - img: /images/content_images/case_studies/deeplabcut.png - alttext: Análise de pose de um guepardo - url: /pt/case-studies/deeplabcut-dnn - - keyfeatures: - features: - - title: Arrays n-dimensionais poderosas - text: Rápidos e versáteis, os conceitos de vetorização, indexação e broadcasting do NumPy são, na prática, o padrão em computação com arrays. - - title: Ferramentas de computação numérica - text: O NumPy oferece um conjunto completo de funções matemáticas, geradores de números aleatórios, rotinas de álgebra linear, transformadas de Fourier, e mais. - - title: Interoperabilidade - text: O NumPy suporta um grande número de plataformas de hardware e computação, e pode ser combinada com bibliotecas de computação com arrays esparsas, distribuidas ou em GPUs. - - title: Alto desempenho - text: O núcleo do NumPy é feito de código otimizado em C. Experimente a flexibilidade do Python com a velocidade de código compilado. - - title: Fácil de usar - text: A sintaxe de alto nível do NumPy torna-o acessível e produtivo para programadores de qualquer nível de experiência e formação. - - title: Código aberto - text: Distribuido com uma [licença BSD](https://github.com/numpy/numpy/blob/main/LICENSE.txt) liberal, o NumPy é desenvolvido e mantido [publicamente no GitHub](https://github.com/numpy/numpy) por uma [comunidade](/pt/community) vibrante, responsiva, e diversa. - + - + title: A Primeira Imagem de um Buraco Negro + text: Como o NumPy, junto com outras bibliotecas como SciPy e Matplotlib que dependem do NumPy, permitiram ao Event Horizon Telescope gerar a primeira imagem de um buraco negro da história. + img: /images/content_images/case_studies/blackhole.png + alttext: Primeira imagem de um buraco negro. É um círculo laranja em um fundo preto. + url: /pt/case-studies/blackhole-image + - + title: Descoberta de Ondas Gravitacionais + text: Em 1916, Albert Einstein previu ondas gravitacionais; 100 anos depois, sua existência foi confirmada pelos cientistas do LIGO usando NumPy. + img: /images/content_images/case_studies/gravitional.png + alttext: Duas esferas orbitando a si mesmas. Elas deslocam a gravidade em seu entorno. + url: /pt/case-studies/gw-discov + - + title: Análise Esportiva + text: A análise de críquete está mudando o jogo ao melhorar o desempenho de jogadores e times através de modelagem estatística e análise preditiva. O NumPy possibilita muitas dessas análises. + img: /images/content_images/case_studies/sports.jpg + alttext: Bola de críquete em um campo verde + url: /pt/case-studies/cricket-analytics + - + title: Estimação de poses usando deep learning + text: DeepLabCut usa o NumPy para acelerar estudos científicos que envolvem comportamento animal para entender melhor o controle motor em várias espécies e escalas de tempo. + img: /images/content_images/case_studies/deeplabcut.png + alttext: Análise de pose de um guepardo + url: /pt/case-studies/deeplabcut-dnn tabs: title: ECOSSISTEMA section5: false - -navbar: - - title: Instalação - url: /pt/install - - title: Documentação - url: https://numpy.org/doc/stable - - title: Aprenda - url: /pt/learn - - title: Comunidade - url: /pt/community - - title: Sobre - url: /pt/about - - title: Contribuir - url: /pt/contribute - -footer: - logo: logo.svg - socialmediatitle: "" - socialmedia: - - link: https://github.com/numpy/numpy - icon: github - - link: https://www.youtube.com/channel/UCguIL9NZ7ybWK5WQ53qbHng - icon: youtube - - link: https://twitter.com/numpy_team - icon: twitter - quicklinks: - column1: - title: "" - links: - - text: Instalação - link: /pt/install - - text: Documentação - link: https://numpy.org/doc/stable - - text: Aprenda - link: /pt/learn - - text: Citando o Numpy - link: /pt/citing-numpy - - text: Roadmap - link: https://numpy.org/neps/roadmap.html - column2: - links: - - text: Sobre - link: /pt/about - - text: Comunidade - link: /pt/community - - text: User surveys - link: /pt/user-surveys - - text: Contribuir - link: /pt/contribute - - text: Código de Conduta - link: /pt/code-of-conduct - column3: - links: - - text: Ajuda - link: /pt/gethelp - - text: Termos de uso (EN) - link: /pt/terms - - text: Privacidade - link: /pt/privacy - - text: Kit de imprensa - link: /pt/press-kit - + navbar: + - + title: Instalação + url: /pt/install + - + title: Documentação + url: https://numpy.org/doc/stable + - + title: Aprenda + url: /pt/learn + - + title: Comunidade + url: /pt/community + - + title: Sobre + url: /pt/about + - + title: Notícias + url: /pt/news + - + title: Contribuir + url: /pt/contribute + footer: + logo: logo.svg + socialmediatitle: "" + socialmedia: + - + link: https://github.com/numpy/numpy + icon: github + - + link: https://www.youtube.com/channel/UCguIL9NZ7ybWK5WQ53qbHng + icon: youtube + - + link: https://twitter.com/numpy_team + icon: twitter + quicklinks: + column1: + title: "" + links: + - + text: Instalação + link: /pt/install + - + text: Documentação + link: https://numpy.org/doc/stable + - + text: Aprenda + link: /pt/learn + - + text: Citando o Numpy + link: /pt/citing-numpy + - + text: Roadmap + link: https://numpy.org/neps/roadmap.html + column2: + links: + - + text: Sobre + link: /pt/about + - + text: Comunidade + link: /pt/community + - + text: Pesquisas de usuário + link: /pt/user-surveys + - + text: Contribuir + link: /pt/contribute + - + text: Código de Conduta + link: /pt/code-of-conduct + column3: + links: + - + text: Ajuda + link: /pt/gethelp + - + text: Termos de uso (EN) + link: /pt/terms + - + text: Privacidade + link: /pt/privacy + - + text: Kit de imprensa + link: /pt/press-kit diff --git a/content/pt/contribute.md b/content/pt/contribute.md index 5bf71db883..65b82636b8 100644 --- a/content/pt/contribute.md +++ b/content/pt/contribute.md @@ -5,30 +5,18 @@ sidebar: false O projeto NumPy precisa de sua experiência e entusiasmo! Suas opções de não são limitadas à programação -- além de -- [Escrever código]({{< relref "contribute.md#writing-code" >}}) - -você pode: - -- [Revisar pull requests]({{< relref "contribute.md#reviewing-pull-requests" >}}) -- [Desenvolver tutoriais, apresentações e outros materiais educacionais]({{< relref "contribute.md#developing-educational-materials" >}}) -- [Fazer triagem em issues]({{< relref "contribute.md#issue-triaging" >}}) -- [Trabalhar no nosso site]({{< relref "contribute.md#website-development" >}}) -- [Contribuir com design gráfico]({{< relref "contribute.md#graphic-design" >}}) -- [Traduzir conteúdo do site]({{< relref "contribute.md#translating-website-content" >}}) -- [Trabalhar coordenando a comunidade]({{< relref "contribute.md#community-coordination-and-outreach" >}}) -- [Escrever propostas e ajudar com outras atividades para financiamento]({{< relref "contribute.md#fundraising" >}}) - Se você não sabe por onde começar ou como suas habilidades podem ajudar, _fale conosco!_ Você pode perguntar na nossa [lista de emails](https://mail.python.org/mailman/listinfo/numpy-discussion) ou [GitHub](http://github.com/numpy/numpy) (abrindo uma [issue](https://github.com/numpy/numpy/issues) ou comentando em uma issue relevante). Estes são os nossos canais de comunicação preferidos (projetos de código aberto são abertos por natureza!). No entanto, se você preferir discutir em privado, entre em contato com os coordenadores da comunidade em ou no [Slack](https://numpy-team.slack.com) (envie um e-mail para para obter um convite antes de entrar). -Nós também temos uma _reunião aberta da comunidade_ a cada duas semanas. Os detalhes são anunciados na nossa [lista de emails](https://mail.python.org/mailman/listinfo/numpy-discussion). Convidamos você a participar desta chamada se quiser. Se você nunca contribuiu para projetos de código aberto, recomendamos fortemente que você leita [esse guia](https://opensource.guide/how-to-contribute/). +Nós também temos uma _reunião aberta da comunidade_ a cada duas semanas. Os detalhes são anunciados na nossa [lista de emails](https://mail.python.org/mailman/listinfo/numpy-discussion). Convidamos você a participar. Se você nunca contribuiu para projetos de código aberto, recomendamos fortemente que você leita [esse guia](https://opensource.guide/how-to-contribute/). Nossa comunidade deseja tratar todos da mesma forma e valorizar todas as contribuições. Temos um [Código de Conduta](/pt/code-of-conduct) para promover um ambiente aberto e acolhedor. ### Escrevendo código -Para pessoas programadoras, este [guia](https://numpy.org/devdocs/dev/index.html#development-process-summary) explica como contribuir para a base de código. +Para pessoas programadoras, este [guia](https://numpy.org/devdocs/dev/index.html#development-process-summary) explica como contribuir para a base de código.
Confira também nosso [canal do YouTube](https://www.youtube.com/playlist?list=PLCK6zCrcN3GXBUUzDr9L4__LnXZVtaIzS) para obter informações adicionais. + ### Revisar pull requests O projeto tem mais de 250 pull requests abertos -- o que significa que muitas potenciais melhorias e muitos contribuidores de código aberto estão aguardando feedback. Se você é uma pessoa programadora que conhece o NumPy, você pode ajudar, mesmo que não tenha familiaridade com o código. Você pode: @@ -76,4 +64,3 @@ Através do contato com a comunidade podemos compartilhar nosso trabalho para ma ### Financiamento O NumPy foi um projeto totalmente voluntário por muitos anos, mas conforme sua importância cresceu, tornou-se clara a necessidade de apoio financeiro para garantir estabilidade e crescimento. [Esta palestra na SciPy'19](https://www.youtube.com/watch?v=dBTJD_FDVjU) explica quanta diferença esse suporte fez. Como todo o mundo das organizações sem fins lucrativos, nós estamos constantemente procurando bolsas, patrocinadores e outros tipos de apoio. Nós temos uma série de ideias e é claro que nós damos as boas-vindas a mais. Habilidade de buscar financiamento é uma habilidade rara aqui -- apreciaríamos a sua ajuda. - diff --git a/content/pt/gethelp.md b/content/pt/gethelp.md index bba586e7f2..f9f0e0322a 100644 --- a/content/pt/gethelp.md +++ b/content/pt/gethelp.md @@ -3,11 +3,9 @@ title: Obter ajuda sidebar: false --- -**Perguntas de usuários:** A melhor maneira de obter ajuda é postar sua pergunta em um site como [StackOverflow](http://stackoverflow.com/questions/tagged/numpy), com milhares de usuários disponíveis para responder. Outras alternativas incluem [IRC](https://webchat.freenode.net/?channels=%23numpy), [Gitter](https://gitter.im/numpy/numpy)e [Reddit](https://www.reddit.com/r/Numpy/). Gostaríamos de poder ficar de olho nestes sites, ou responder perguntas diretamente, mas o volume é imenso! - -**Issues sobre desenvolvimento:** Para assuntos relacionados ao desenvolvimento do NumPy (por exemplo, relatórios de bugs), veja a [Comunidade](/community). - +**Development issues:** For NumPy development-related matters (e.g., bug reports), please see [Community](/community). +**User questions:** The best way to get help is to post your question to a site like [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) or [Reddit](https://www.reddit.com/r/Numpy/). We wish we could keep an eye on these sites, or answer questions directly, but the volume is a little overwhelming! ### [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) @@ -20,15 +18,3 @@ Um fórum para fazer perguntas sobre a utilização da biblioteca, por exemplo: Outro fórum para perguntas de utilização. *** - -### [Gitter](https://gitter.im/numpy/numpy) - -Uma sala de bate-papo em tempo real onde usuários e membros da comunidade se ajudam uns aos outros. - -*** - -### [IRC](https://webchat.freenode.net/?channels=%23numpy) - -Outra sala de bate-papo em tempo real onde usuários e membros da comunidade se ajudam uns aos outros. - -*** diff --git a/content/pt/install.md b/content/pt/install.md index 0701460ee9..ff2f33845a 100644 --- a/content/pt/install.md +++ b/content/pt/install.md @@ -30,6 +30,7 @@ pip install numpy ``` Também ao usar o pip, é uma boa prática usar um ambiente virtual - veja em [Instalações Reprodutíveis](#reproducible-installs) abaixo por quê, e [esse guia](https://dev.to/bowmanjd/python-tools-for-managing-virtual-environments-3bko#howto) para detalhes sobre o uso de ambientes virtuais. + # Guia de instalação do Python e do NumPy @@ -65,7 +66,7 @@ Para usuários que preferem uma solução baseada em pip/PyPI, por preferência ## Gerenciamento de pacotes Python -Gerenciar pacotes é um problema desafiador e, como resultado, há muitas ferramentas. Para o desenvolvimento web e de propósito geral em Python, há uma [série de ferramentas](https://packaging.python.org/guides/tool-recommendations/) complementares com pip. Para computação de alto desempenho (HPC), vale a pena considerar o [Spack](https://github.com/spack/spack). Para a maioria dos usuários NumPy, porém, o [conda](https://conda.io/en/latest/) e o [pip](https://pip.pypa.io/en/stable/) são as duas ferramentas mais populares. +Gerenciar pacotes é um problema desafiador e, como resultado, há muitas ferramentas. Para o desenvolvimento web e de propósito geral em Python, há uma [série de ferramentas](https://packaging.python.org/guides/tool-recommendations/) complementares com pip. Para computação de alto desempenho (HPC), vale a pena considerar o [Spack](https://github.com/spack/spack). Para computação de alto desempenho (HPC), vale a pena considerar o [Spack](https://github.com/spack/spack). Para a maioria dos usuários NumPy, porém, o [conda](https://conda.io/en/latest/) e o [pip](https://pip.pypa.io/en/stable/) são as duas ferramentas mais populares. ### Pip & conda diff --git a/content/pt/learn.md b/content/pt/learn.md index 9738385ffd..e20462b9c1 100644 --- a/content/pt/learn.md +++ b/content/pt/learn.md @@ -5,9 +5,10 @@ sidebar: false Para a **documentação oficial do NumPy** visite [numpy.org/doc/stable](https://numpy.org/doc/stable). -Abaixo está uma coleção de recursos externos selecionados. Para contribuir, veja o [fim desta página](#add-to-this-list). *** +Abaixo está uma coleção de recursos educacionais, tanto para autoaprendizado como para ensinar outras pessoas, desenvolvidos pelos colaboradores do NumPy e selecionados pela comunidade. + ## Iniciantes Há uma tonelada de informações sobre o NumPy lá fora. Se você está começando, recomendamos fortemente estes: @@ -15,13 +16,12 @@ Há uma tonelada de informações sobre o NumPy lá fora. Se você está começa **Tutoriais** * [NumPy Quickstart Tutorial (Tutorial de Início Rápido)](https://numpy.org/devdocs/user/quickstart.html) +* [NumPy Tutorials](https://numpy.org/numpy-tutorials) Uma coleção de tutoriais e materiais educacionais no formato de Notebooks Jupyter desenvolvidos e mantidos pelo time de documentação do NumPy. Se você tiver interesse em adicionar o seu próprio conteúdo, verifique o repositório [numpy-tutorials no GitHub](https://github.com/numpy/numpy-tutorials). +* [NumPy Illustrated: The Visual Guide to NumPy *por Lev Maximov*](https://betterprogramming.pub/3b1d4976de1d?sk=57b908a77aa44075a49293fa1631dd9b) * [SciPy Lectures](https://scipy-lectures.org/) Além de incluir conteúdo sobre a NumPy, estas aulas oferecem uma introdução mais ampla ao ecossistema científico do Python. * [NumPy: the absolute basics for beginners ("o básico absoluto para inciantes")](https://numpy.org/devdocs/user/absolute_beginners.html) -* [Machine Learning Plus - Introduction to ndarray](https://www.machinelearningplus.com/python/numpy-tutorial-part1-array-python-examples/) -* [Edureka - Learn NumPy Arrays with Examples ](https://www.edureka.co/blog/python-numpy-tutorial/) -* [Dataquest - NumPy Tutorial: Data Analysis with Python](https://www.dataquest.io/blog/numpy-tutorial-python/) * [NumPy tutorial *por Nicolas Rougier*](https://github.com/rougier/numpy-tutorial) -* [Stanford CS231 *by Justin Johnson*](http://cs231n.github.io/python-numpy-tutorial/) +* [Stanford CS231 *por Justin Johnson*](http://cs231n.github.io/python-numpy-tutorial/) * [NumPy User Guide (Guia de Usuário NumPy)](https://numpy.org/devdocs) **Livros** @@ -47,9 +47,7 @@ Experimente esses recursos avançados para uma melhor compreensão dos conceitos * [100 NumPy Exercises](http://www.labri.fr/perso/nrougier/teaching/numpy.100/index.html) *por Nicolas P. Rougier* * [An Introduction to NumPy and Scipy](https://engineering.ucsb.edu/~shell/che210d/numpy.pdf) *por M. Scott Shell* * [Numpy Medkits](http://mentat.za.net/numpy/numpy_advanced_slides/) *por Stéfan van der Walt* -* [NumPy in Python (Advanced)](https://www.geeksforgeeks.org/numpy-python-set-2-advanced/) -* [Advanced Indexing](https://www.tutorialspoint.com/numpy/numpy_advanced_indexing.htm) -* [Machine Learning and Data Analytics with NumPy](https://www.machinelearningplus.com/python/numpy-tutorial-python-part2/) +* [NumPy Tutorials](https://numpy.org/numpy-tutorials) Uma coleção de tutoriais e materiais educacionais no formato de Notebooks Jupyter desenvolvidos e mantidos pelo time de documentação do NumPy. Se você tiver interesse em adicionar o seu próprio conteúdo, verifique o repositório [numpy-tutorials no GitHub](https://github.com/numpy/numpy-tutorials). **Livros** @@ -60,7 +58,6 @@ Experimente esses recursos avançados para uma melhor compreensão dos conceitos **Vídeos** * [Advanced NumPy - broadcasting rules, strides, and advanced indexing](https://www.youtube.com/watch?v=cYugp9IN1-Q) *por Juan Nunuz-Iglesias* -* [Advanced Indexing Operations in NumPy Arrays](https://www.youtube.com/watch?v=2WTDrSkQBng) *por Amuls Academy* *** @@ -77,8 +74,3 @@ Experimente esses recursos avançados para uma melhor compreensão dos conceitos ## Citando a NumPy Se a NumPy é importante na sua pesquisa, e você gostaria de dar reconhecimento ao projeto na sua publicação acadêmica, por favor veja [estas informações sobre citações](/pt/citing-numpy). - -## Contribua para esta lista - - -Para adicionar a essa coleção, envie uma recomendação [através de um pull request](https://github.com/numpy/numpy.org/blob/main/content/en/learn.md). Diga por que sua recomendação merece ser mencionada nesta página e também qual o público que mais se beneficiaria. diff --git a/content/pt/news.md b/content/pt/news.md index dd7a55e3c8..085a49d625 100644 --- a/content/pt/news.md +++ b/content/pt/news.md @@ -1,11 +1,172 @@ --- title: Notícias sidebar: false +newsHeader: "NumPy 2.0 release date: June 16" +date: 2023-09-16 --- +### NumPy 2.0 release date: June 16 + +_23 May, 2024_ -- We are excited to announce that NumPy 2.0 is planned to be released on June 16, 2024. This release has been over a year in the making, and is the first major release since 2006. Importantly, in addition to many new features and performance improvement, it contains **breaking changes** to the ABI as well as the Python and C APIs. It is likely that downstream packages and end user code needs to be adapted - if you can, please verify whether your code works with NumPy `2.0.0rc2`. **Please see the following for more details:** + +- The [NumPy 2.0 migration guide](https://numpy.org/devdocs/numpy_2_0_migration_guide.html) +- The [2.0.0 release notes](https://numpy.org/devdocs/release/2.0.0-notes.html) +- Announcement issue for status updates: [numpy#24300](https://github.com/numpy/numpy/issues/24300) + + +### Lançado o NumPy versão 1.26.0 +_19 de dez, 2023_ -- O NumFOCUS se juntou ao PyCharm durante sua campanha de final de ano para oferecer 30% de desconto em licenças de PyCharm para novos usuários. Todas as receitas do primeiro ano das compras do PyCharm a partir de agora até 23 de dezembro, 2023 irão diretamente para os programas NumFOCUS. + +Use a URL única que permitirá rastrear as compras https://lp.jetbrains.com/support-data-science/ ou um código de cupom ISUPPORTDATASCIENCE  + +### Lançado o NumPy versão 1.26.0 + +_16 de setembro de 2023_ -- [NumPy 1.26.0](https://numpy.org/doc/stable/release/1.26.0-notes.html) está disponível. Os destaques desta versão são: + +* Suport ao Python 3.12.0. +* Compatibilidade com Cython 3.0.0. +* Utilização do sistema Meson para compilação +* Suport a SIMD atualizado +* Melhorias para f2py, suporte a meson e bind(x) +* Suporte à versão mais recente da biblioteca Accelerate BLAS/LAPACK + +A versão 1.26.0 é uma continuação da série de versões 1.25.x que marcam a transição para o sistema de compilação Meson e oferecem suporte preliminar para o Cython 3.0.0. Um total de 20 pessoas contribuíram para este lançamento e 59 pull requests foram incorporadas. + +As versões do Python suportadas por esta versão são 3.9-3.12. + +### numpy.org agora está disponível em japonês e português + +_2 de agosto de 2023_ -- numpy.org agora está disponível em 2 idiomas adicionais: japonês e português. Isto não seria possível sem nossos voluntários dedicados: + +_Português:_ +* Melissa Weber Mendonça (melissawm) +* Ricardo Prins (ricardoprins) +* Getúlio Silva (getuliosilva) +* Julio Batista Silva (jbsilva) +* Alexandre de Siqueira (alexdesiqueira) +* Alexandre B A Villares (villares) +* Vini Salazar (vinisalazar) + +_Japonês:_ +* Atsushi Sakai (AtsushiSakai) +* KKunai +* Tom Kelly (TomKellyGenetics) +* Yuji Kanagawa (kngwyu) +* Tetsuo Koyama (tkoyama010) + +O trabalho na infraestrutura de tradução é apoiado com financiamento da CZI. + +Futuramente, nós adoraríamos traduzir o site para mais idiomas. Se você quiser ajudar, conecte-se com a Equipe de Traduções NumPy no Slack: https://join.slack.com/t/numpy-team/shared_invite/zt-1gokbq56s-bvEpo10Ef7aHbVtVFeZv2w. (Procure pelo canal de #translations.) Também estamos construindo uma Equipe de Traduções que vai trabalhar na localização da documentação e conteúdo educacional por todo o ecossistema de Python científico. Se isto despertou o seu interesse, junte-se a nós no servidor Discord Scientific Python: https://discord.gg/khWtqY6RKr. (Procure pelo canal #translation.) + +### Lançado o NumPy 1.25.0 + +_17 de junho, 2023_ -- [NumPy 1.25.0](https://numpy.org/doc/stable/release/1.25.0-notes.html) está disponível agora. Os destaques desta versão são: + +* Suporte para MUSL, agora existem rodas MUSL. +* Suporte para o compilador Fujitsu C/C++. +* Arrays de objetos agora são suportados em einsum. +* Suporte para a multiplicação da matriz inplace (`@=`). + +A versão 1.25.0 do NumPy continua o trabalho de melhorias no suporte e promoção de dtypes, na velocidade e execução, e na documentação. Também tem havido trabalho preparatório para a futura versão 2.0.0, resultando em um grande número de depreciações novas e expiradas. + +Um total de 148 pessoas contribuíram para este lançamento e 530 pull requests foram incorporadas. + +As versões do Python suportadas por esta versão são 3.9-3.11. + +### Promovendo uma cultura inclusiva: Chamada de participação + +_10 de maio de 2023_ -- Promovendo uma Cultura Inclusiva: Chamada de Participação + +Como podemos ser melhores quando se trata de diversidade e de inclusão? Leia o relatório e descubra como colaborar [aqui](https://contributor-experience.org/docs/posts/dei-report/). + +### Transição de liderança do time de documentação do NumPy + +_6 de janeiro de 2023_ –- Mukulika Pahari e Ross Barnowski são nomeados como lideres do time de documentação do NumPy, substituindo Melissa Mendonça. Agradecemos a Melissa por todas suas contribuições para a documentação oficial do NumPy e materiais educacionais, e Mukulika e Ross por aceitarem o desafio. + +### NumPy versão 1.24.0 + +_18 de dezembro de 2022_ -- [NumPy 1.24.0](https://numpy.org/doc/stable/release/1.24.0-notes.html) está agora disponível. Os destaques desta versão são: + +* Novas palavras-chave "dtype" e "casting" para funções que atuam com stacking. +* Novas funcionalidades e correções do F2PY. +* Muitas depreciações novas, confira. +* Muitas depreciações expiradas. + +A versão 1.24.0 do NumPy continua o trabalho de melhorias no suporte e promoção de dtypes, na velocidade e execução, e na documentação. Há um grande número de depreciações novas e expiradas devido a mudanças na promoção de dtypes e limpezas no código. É o trabalho de 177 contribuidores espalhados em 444 pull requests. As versões suportadas do Python são 3.8-3.11. + +### NumPy versão 1.23.0 + +_22 de junho de 2022_ -- O [NumPy 1.23.0](https://numpy.org/doc/stable/release/1.23.0-notes.html) está disponível. Os destaques desta versão são: + +* Implementação de `loadtxt` em C, melhorando muito seu desempenho. +* Exposição do DLPack ao nível de Python para facilitar a troca de dados. +* Mudanças na promoção e comparações de dtypes estruturados. +* Melhorias no f2py. + +A versão 1.23.0 do NumPy continua o trabalho de melhorias no suporte e promoção de dtypes, na velocidade de execução, na documentação e na expiração de depreciações. É o trabalho de 151 contribuidores espalhados em 494 pull requests. As versões do Python suportadas por esta versão 3.8-3.10. Python 3.11 será suportado quando chegar na etapa rc. + +### Pesquisa NumFOCUS DEI: chamada para participação + +_13 de abril de 2022_ -- O NumPy está trabalhando com a [NumFOCUS](http://numfocus.org/) em um [projeto de pesquisa](https://numfocus.org/diversity-inclusion-disc/a-pivotal-time-in-numfocuss-project-aimed-dei-efforts?eType=EmailBlastContent&eId=f41a86c3-60d4-4cf9-86cf-58eb49dc968c) financiado pela [Gordon & Betty Moore Foundation](https://www.moore.org/) para entender as barreiras à participação que contribuidores, especialmente aqueles de grupos historicamente subrepresentados, enfrentam na comunidade open source. A equipe da pesquisa gostaria de falar com novos colaboradores, desenvolvedores e mantenedores, e aqueles que contribuíram no passado sobre suas experiências contribuindo para o NumPy. + +**Quer compartilhar suas experiências?** + +Por favor, preencha este breve formulário: ["Participant Interest form"](https://numfocus.typeform.com/to/WBWVJSqe) que contém informações adicionais sobre os objetivos da pesquisa, privacidade e considerações de confidencialidade. Sua participação será valiosa para o crescimento e sustentabilidade de comunidades de software open source diversas e inclusivas. Os participantes aceitos participarão de uma entrevista de 30 minutos com um membro da equipe de pesquisa. + +### NumPy versão 1.22.0 + +_31 de dezembro de 2021_ -- [NumPy 1.22.0](https://numpy.org/doc/stable/release/1.22.0-notes.html) está agora disponível. Os destaques desta versão são: + +* Anotações de tipo do namespace principal estão praticamente completas. Ainda há trabalho a se fazer no upstream, mas a maior parte do trabalho está feita. Esta é provavelmente a melhoria mais visível para os usuários nesta versão. +* Uma versão preliminar da proposta do [array API Standard](https://data-apis.org/array-api/latest/) está disponível (veja [NEP 47](https://numpy.org/neps/nep-0047-array-api-standard.html)). Este é um passo na criação de uma coleção padrão de funções que podem ser compartilhadas entre bibliotecas como CuPy e JAX. +* NumPy agora tem um backend de DLPack. DLPack fornece um formato comum de compartilhamento para dados de arrays (tensores). +* Novos métodos para `quantile`, `percentile`, e funções relacionadas. Os novos métodos fornecem um conjunto completo dos métodos comumente encontrados na literatura. +* As funções universais foram refatoradas para implementar a maior parte da [NEP 43](https://numpy.org/neps/nep-0043-extensible-ufuncs.html). Isso também desbloqueia a capacidade de experimentar a futura API DType. +* Um novo alocador de memória configurável para uso pelos projetos downstream. + +NumPy 1.22.0 é uma versão importante com o trabalho de 153 contribuidores espalhados por mais de 609 pull requests. As versões do Python suportadas por esta versão são 3.8-3.10. + +### Promovendo uma cultura inclusiva no ecossistema científico de Python + +_31 de agosto de 2021_ -- Estamos felizes em anunciar que a Chan Zuckerberg Initiative [vai financiar](https://chanzuckerberg.com/newsroom/czi-awards-16-million-for-foundational-open-source-software-tools-essential-to-biomedicine/) um projeto para apoiar a integração, inclusão, e retenção de pessoas de grupos marginalizados historicamente em projetos científicos em Python, e para estruturalmente melhorar a dinâmica das comunidades para o NumPy, SciPy, Matplotlib, e Pandas. + +Como parte do programa [CZI's Essential Open Source Software for Science](https://chanzuckerberg.com/eoss/), esse [financiamento adicional para diversidade e inclusão](https://cziscience.medium.com/advancing-diversity-and-inclusion-in-scientific-open-source-eaabe6a5488b) vai apoiar a criação de posições de Contributor Experience Lead para identificar, documentar e implementar práticas para fomentar comunidades open source inclusivas. Este projeto será liderado por Melissa Mendonça (NumPy), com apoio adicional de Ralf Gommers (NumPy, SciPy), Hannah Aizenman e Thomas Caswell (Matplotlib), Matt Haberland (SciPy), e Joris Van den Bossche (Pandas). + +Esse é um projeto ambicioso que visa descobrir e implementar atividades que devem estruturalmente melhorar a dinâmica da comunidade de nossos projetos. Ao criar essas novas funções entre projetos, esperamos introduzir um novo modelo de colaboração às comunidades de Python científico, permitir que o trabalho de construção da comunidade no ecossistema seja feito de forma mais eficiente e com maiores resultados. Também esperamos desenvolver uma imagem mais clara do que funciona e o que não funciona em nossos projetos para engajar e reter novos colaboradores, especialmente de grupos historicamente sub-representados. Finalmente, planejamos produzir relatórios detalhados sobre as ações executadas, explicando como eles afetaram nossos projetos em termos de representação e interação com nossas comunidades. + +O projeto de dois anos deverá começar em novembro de 2021 e estamos animados para ver os resultados deste trabalho! [Você pode ler a proposta completa aqui](https://figshare.com/articles/online_resource/Advancing_an_inclusive_culture_in_the_scientific_Python_ecosystem/16548063). + +### Pesquisa NumPy 2021 + +_12 de julho de 2021_ -- Nós do NumPy acreditamos no poder da nossa comunidade. 1,236 usuários do NumPy de 75 países participaram da nossa primeira pesquisa ano passado. Os resultados da pesquisa nos ajudaram a compreender muito bem o que devemos fazer pelos 12 meses seguintes. + +Chegou a hora de fazer outra pesquisa e estamos contando com você novamente. Vai levar cerca de 15 minutos do seu tempo. Além de Inglês, o questionário de pesquisa está disponível em 8 idiomas adicionais: Bangla, Francês, Hindi, Japonês, Mandarim, Português, Russo e Espanhol. + +Siga o link para começar: https://berkeley.qualtrics.com/jfe/form/SV_aaOONjgcBXDSl4q. + + +### NumPy versão 1.19.0 + +_23 de junho de 2021_ -- O [NumPy 1.21.0](https://numpy.org/doc/stable/release/1.21.0-notes.html) está disponível. Os destaques desta versão são: + +- a continuação do trabalho com SIMD para suportar mais funções e plataformas, +- trabalho inicial na infraestrutura e conversão de novos dtypes, +- wheels universal2 para Python 3.8 e Python 3.9 no Mac, +- melhorias na documentação, +- melhorias nas anotações de tipos, +- novo bitgenerator `PCG64DXSM` para números aleatórios. + +Esta versão do NumPy é o resultado de 581 pull requests aceitos, a partir das contribuições de 175 pessoas. As versões do Python suportadas por esta versão são 3.7-3.9; o suporte para o Python 3.10 será adicionado após o lançamento do Python 3.10. + + +### Resultados da pesquisa NumPy 2020 + +_22 de junho de 2021_ -- Em 2020, o time de pesquisas NumPy, em parceria com estudantes e professores da Universidade de Michigan e da Universidade de Maryland, realizou a primeira pesquisa oficial sobre a comunidade NumPy. Encontre os resultados da pesquisa aqui: https://numpy.org/user-survey-2020/. + + ### NumPy versão 1.20.0 -_30 de janeiro de 2021_ -- O [NumPy 1.20.0](https://numpy.org/doc/stable/release/1.20.0-notes.html) está disponível. Este é o maior release do NumPy até agora, graças a mais de 180 contribuidores. As duas novidades mais emocionantes são: +_30 de janeiro de 2021_ -- O [NumPy 1.20.0](https://numpy.org/doc/stable/release/1.20.0-notes.html) está disponível. Este é o maior lançamento do NumPy até hoje, graças a mais de 180 colaboradores. As duas novidades mais emocionantes são: - Anotações de tipos para grandes partes do NumPy, e um novo submódulo `numpy.typing` contendo aliases `ArrayLike` e `DtypeLike` que usuários e bibliotecas downstream podem usar quando quiserem adicionar anotações de tipos em seu próprio código. - Otimizações de compilação SIMD multi-plataforma, com suporte para instruções x86 (SSE, AVX), ARM64 (Neon) e PowerPC (VSX). Isso rendeu melhorias significativas de desempenho para muitas funções (exemplos: [sen/cos](https://github.com/numpy/numpy/pull/17587), [einsum](https://github.com/numpy/numpy/pull/18194)). @@ -43,7 +204,7 @@ _24 de junho de 2020_ -- NumPy agora tem um novo logo: NumPy logo -O logo é uma versão moderna do antigo, com um design mais limpo. Obrigado a Isabela Presedo-Floyd por projetar o novo logo, bem como o Travis Vaught pelo o logo antigo que nos serviu bem durante mais de 15 anos. +O logotipo é uma versão moderna do antigo, com um design mais limpo. Obrigado à Isabela Presedo-Floyd por projetar o novo logotipo, bem como ao Travis Vaught pelo o logotipo antigo que nos serviu bem durante mais de 15 anos. ### NumPy versão 1.19.0 @@ -58,32 +219,61 @@ _11 de maio de 2020_ -- O NumPy foi aceito como uma das organizações mentoras ### NumPy versão 1.18.0 -_22 de dezembro de 2019_ -- O NumPy 1.18.0 está disponível. Após as principais mudanças em 1.17.0, esta é uma versão de consolidação. Esta é a última versão menor que irá suportar Python 3.5. Destaques dessa versão incluem a adição de uma infraestrutura básica para permitir o link com as bibliotecas BLAS e LAPACK em 64 bits durante a compilação, e uma nova C-API para `numpy.random`. +_22 de dezembro de 2019_ -- O NumPy 1.18.0 está disponível. Após as principais mudanças em 1.17.0, esta é uma versão de consolidação. É a última versão menor que suportará Python 3.5. Destaques dessa versão incluem a adição de uma infraestrutura básica para permitir o link com as bibliotecas BLAS e LAPACK em 64 bits durante a compilação, e uma nova C-API para `numpy.random`. Por favor, veja as [notas de lançamento](https://github.com/numpy/numpy/releases/tag/v1.18.0) para mais detalhes. ### O NumPy recebe financiamento da Chan Zuckerberg Initiative -_15 de novembro de 2019_ -- Estamos felizes em anunciar que o NumPy e a OpenBLAS, uma das dependências-chave da NumPy, receberam um auxílio conjunto de $195,000 da Chan Zuckerberg Initiative através do seu programa [Essential Open Source Software for Science](https://chanzuckerberg.com/eoss/) que apoia a manutenção, crescimento, desenvolvimento e envolvimento com a comunidade de ferramentas de software open source fundamentais para a ciência. +_15 de novembro de 2019_ -- Estamos felizes em anunciar que o NumPy e a OpenBLAS, uma das dependências-chave do NumPy, receberam um auxílio conjunto de $195,000 da Chan Zuckerberg Initiative através do seu programa [Essential Open Source Software for Science](https://chanzuckerberg.com/eoss/) que apoia a manutenção, crescimento, desenvolvimento e envolvimento da comunidade em ferramentas de código aberto fundamentais para a ciência. -Este auxílio será usado para aumentar os esforços de melhoria da documentação do NumPy, atualização do design do site, e desenvolvimento comunitário para servir melhor a nossa grande e rápida base de usuários, e garantir a sustentabilidade do projeto a longo prazo. Enquanto a equipe OpenBLAS se concentrará em tratar de um conjunto de questões técnicas fundamentais, em particular relacionadas a *thread-safety*, AVX-512, e *thread-local storage* (TLS), bem como melhorias algorítmicas na ReLAPACK (Recursive LAPACK) da qual a OpenBLAS depende. +Este auxílio será usado para aumentar os esforços de melhoria da documentação do NumPy, reformulação do site, desenvolvimento comunitário para melhor servir a nossa grande, e rapidamente crescente, base de usuários, assim como para garantir a sustentabilidade do projeto a longo prazo. Enquanto a equipe OpenBLAS se concentrará em tratar de um conjunto de questões técnicas fundamentais, em particular relacionadas a *thread-safety*, AVX-512, e *thread-local storage* (TLS), bem como melhorias algorítmicas na ReLAPACK (Recursive LAPACK) da qual a OpenBLAS depende. Mais detalhes sobre nossas propostas e resultados esperados podem ser encontrados na [proposta completa de concessão de auxílio](https://figshare.com/articles/Proposal_NumPy_OpenBLAS_for_Chan_Zuckerberg_Initiative_EOSS_2019_round_1/10302167). O trabalho está agendado para começar no dia 1 de dezembro de 2019 e continuar pelos próximos 12 meses. + + ## Lançamentos -Aqui está uma lista de versões do NumPy, com links para notas de lançamento. Todos os lançamentos de bugfix (apenas o `z` muda no formato `x.y.z` do número da versão) não tem novos recursos; versões menores (o `y` aumenta) contém novos recursos. - -- NumPy 1.18.4 ([notas de lançamento](https://github.com/numpy/numpy/releases/tag/v1.18.4)) -- _3 de maio de 2020_. -- NumPy 1.18.3 ([notas de lançamento](https://github.com/numpy/numpy/releases/tag/v1.18.3)) -- _19 de abril de 2020_. -- NumPy 1.18.2 ([notas de lançamento](https://github.com/numpy/numpy/releases/tag/v1.18.2)) -- _17 de março de 2020_. -- NumPy 1.18.1 ([notas de lançamento](https://github.com/numpy/numpy/releases/tag/v1.18.1)) -- _6 de janeiro de 2020_. -- NumPy 1.17.5 ([notas de lançamento](https://github.com/numpy/numpy/releases/tag/v1.17.5)) -- _1 de janeiro de 2020_. -- NumPy 1.18.0 ([notas de lançamento](https://github.com/numpy/numpy/releases/tag/v1.18.0)) -- _22 de dezembro de 2019_. -- NumPy 1.17.4 ([notas de lançamento](https://github.com/numpy/numpy/releases/tag/v1.17.4)) -- _11 de novembro de 2019_. -- NumPy 1.17.0 ([notas de lançamento](https://github.com/numpy/numpy/releases/tag/v1.17.0)) -- _26 de julho de 2019_. -- NumPy 1.16.0 ([notas de lançamento](https://github.com/numpy/numpy/releases/tag/v1.16.0)) -- _14 de janeiro de 2019_. -- NumPy 1.15.0 ([notas de lançamento](https://github.com/numpy/numpy/releases/tag/v1.15.0)) -- _23 de julho de 2018_. -- NumPy 1.14.0 ([notas de lançamento](https://github.com/numpy/numpy/releases/tag/v1.14.0)) -- _7 de janeiro de 2018_. +Aqui está uma lista de versões do NumPy, com links para notas de lançamento. Bugfix lança (apenas o `z` muda no `x.y.` número da versão) não tem novos recursos; versões menores (o `y` aumenta) sim. + +- NumPy 1.26.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.4)) -- _5 Feb 2024_. +- NumPy 1.26.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.3)) -- _2 Jan 2024_. +- NumPy 1.26.2 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.26.2)) -- _12 de novembro de 2023_. +- NumPy 1.26.1 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.26.1)) -- _14 de outubro de 2023_. +- NumPy 1.26.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.26.0)) -- _16 de setembro de 2023_. +- NumPy 1.25.2 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.25.2)) -- _31 de julho de 2023_. +- NumPy 1.25.1 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.25.1)) -- _8 de julho de 2023_. +- NumPy 1.24.4 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.24.4)) -- _26 de junho de 2023_. +- NumPy 1.25.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.25.0)) -- _17 de junho de 2023_. +- NumPy 1.24.3 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.24.3)) -- _22 de abril de 2023_. +- NumPy 1.24.2 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.24.2)) -- _5 de fevereiro de 2023_. +- NumPy 1.24.1 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.24.1)) -- _26 de dezembro de 2022_. +- NumPy 1.24.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.24.0)) -- _18 de dezembro de 2022_. +- NumPy 1.23.5 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.23.5)) -- _19 de novembro de 2022_. +- NumPy 1.23.4 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.23.4)) -- _12 de outubro de 2022_. +- NumPy 1.23.3 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.23.3)) -- _9 de setembro de 2022_. +- NumPy 1.23.2 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.23.2)) -- _14 de agosto de 2022_. +- NumPy 1.23.1 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.23.1)) -- _8 de julho de 2022_. +- NumPy 1.23.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.23.0)) -- _22 de junho de 2022_. +- NumPy 1.22.4 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.22.4)) -- _20 de maio de 2022_. +- NumPy 1.21.6 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.21.6)) -- _12 de abril de 2022_. +- NumPy 1.22.3 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.22.3)) -- _7 de março de 2022_. +- NumPy 1.22.2 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.22.2)) -- _3 de fevereiro de 2022_. +- NumPy 1.22.1 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.22.1)) -- _14 de janeiro de 2022_. +- NumPy 1.22.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.22.0)) -- _31 de dezembro de 2021_. +- NumPy 1.21.5 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.21.5)) -- _19 de dezembro de 2021_. +- NumPy 1.21.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.21.0)) -- _22 de junho de 2021_. +- NumPy 1.20.3 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.20.3)) -- _10 de maio de 2021_. +- NumPy 1.20.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.20.0)) -- _30 de janeiro de 2021_. +- NumPy 1.19.5 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.19.5)) -- _5 de janeiro de 2021_. +- NumPy 1.19.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.19.0)) -- _20 de junho de 2020_. +- NumPy 1.18.4 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.18.4)) -- _3 de maio de 2020_. +- NumPy 1.17.5 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.17.5)) -- _1 de janeiro de 2020_. +- NumPy 1.18.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.18.0)) -- _22 de dezembro de 2019_. +- NumPy 1.17.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.17.0)) -- _26 de julho de 2019_. +- NumPy 1.16.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.16.0)) -- _14 de janeiro de 2019_. +- NumPy 1.15.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.15.0)) -- _23 de julho de 2018_. +- NumPy 1.14.0 ([notas de versão](https://github.com/numpy/numpy/releases/tag/v1.14.0)) -- _7 de janeiro de 2018_. diff --git a/content/pt/privacy.md b/content/pt/privacy.md index be4b6613da..c95f1d5ec1 100644 --- a/content/pt/privacy.md +++ b/content/pt/privacy.md @@ -6,13 +6,3 @@ sidebar: false **numpy.org** é operado por [NumFOCUS, Inc.](https://numfocus.org), o patrocinador fiscal do projeto NumPy. Para a Política de Privacidade deste site, consulte https://numfocus.org/privacy-policy. Se você tiver alguma dúvida sobre a política ou as práticas de coleta de dados do NumFOCUS, uso e divulgação, entre em contato com a equipe do NumFOCUS em privacy@numfocus.org. - - - - - - - - - - diff --git a/content/pt/tabcontents.yaml b/content/pt/tabcontents.yaml index 9f4deb6121..b84da40386 100644 --- a/content/pt/tabcontents.yaml +++ b/content/pt/tabcontents.yaml @@ -1,189 +1,373 @@ -machinelearning: - paras: - - para1: NumPy forms the basis of powerful machine learning libraries like [scikit-learn](https://scikit-learn.org) and [SciPy](https://www.scipy.org). As machine learning grows, so does the list of libraries built on NumPy. [TensorFlow’s](https://www.tensorflow.org) deep learning capabilities have broad applications — among them speech and image recognition, text-based applications, time-series analysis, and video detection. [PyTorch](https://pytorch.org), another deep learning library, is popular among researchers in computer vision and natural language processing. [MXNet](https://github.com/apache/incubator-mxnet) is another AI package, providing blueprints and templates for deep learning. - para2: Statistical techniques called [ensemble](https://towardsdatascience.com/ensemble-methods-bagging-boosting-and-stacking-c9214a10a205) methods such as binning, bagging, stacking, and boosting are among the ML algorithms implemented by tools such as [XGBoost](https://github.com/dmlc/xgboost), [LightGBM](https://lightgbm.readthedocs.io/en/latest/), and [CatBoost](https://catboost.ai) — one of the fastest inference engines. [Yellowbrick](https://www.scikit-yb.org/en/latest/) and [Eli5](https://eli5.readthedocs.io/en/latest/) offer machine learning visualizations. - -arraylibraries: - intro: - - text: NumPy's API is the starting point when libraries are written to exploit innovative hardware, create specialized array types, or add capabilities beyond what NumPy provides. - - headers: - - text: Array Library - - text: Capabilities & Application areas - - libraries: - - title: Dask - text: Distributed arrays and advanced parallelism for analytics, enabling performance at scale. - img: /images/content_images/arlib/dask.png - alttext: Dask - url: https://dask.org/ - - title: CuPy - text: NumPy-compatible array library for GPU-accelerated computing with Python. - img: /images/content_images/arlib/cupy.png - alttext: CuPy - url: https://cupy.chainer.org - - title: JAX - text: "Composable transformations of NumPy programs differentiate: vectorize, just-in-time compilation to GPU/TPU." - img: /images/content_images/arlib/jax_logo_250px.png - alttext: JAX - url: https://github.com/google/jax - - title: Xarray - text: Labeled, indexed multi-dimensional arrays for advanced analytics and visualization - img: /images/content_images/arlib/xarray.png - alttext: xarray - url: https://xarray.pydata.org/en/stable/index.html - - title: Sparse - text: NumPy-compatible sparse array library that integrates with Dask and SciPy's sparse linear algebra. - img: /images/content_images/arlib/sparse.png - alttext: sparse - url: https://sparse.pydata.org/en/latest/ - - title: PyTorch - text: Deep learning framework that accelerates the path from research prototyping to production deployment. - img: /images/content_images/arlib/pytorch-logo-dark.svg - alttext: PyTorch - url: https://pytorch.org/ - - title: TensorFlow - text: An end-to-end platform for machine learning to easily build and deploy ML powered applications. - img: /images/content_images/arlib/tensorflow-logo.svg - alttext: TensorFlow - url: https://www.tensorflow.org - - title: MXNet - text: Deep learning framework suited for flexible research prototyping and production. - img: /images/content_images/arlib/mxnet_logo.png - alttext: MXNet - url: https://mxnet.apache.org/ - - title: Arrow - text: A cross-language development platform for columnar in-memory data and analytics. - img: /images/content_images/arlib/arrow.png - alttext: arrow - url: https://github.com/apache/arrow - - title: xtensor - text: Multi-dimensional arrays with broadcasting and lazy computing for numerical analysis. - img: /images/content_images/arlib/xtensor.png - alttext: xtensor - url: https://github.com/xtensor-stack/xtensor-python - - title: XND - text: Develop libraries for array computing, recreating NumPy's foundational concepts. - img: /images/content_images/arlib/xnd.png - alttext: xnd - url: https://xnd.io - - title: uarray - text: Python backend system that decouples API from implementation; unumpy provides a NumPy API. - img: /images/content_images/arlib/uarray.png - alttext: uarray - url: https://uarray.org/en/latest/ - - title: tensorly - text: Tensor learning, algebra and backends to seamlessly use NumPy, MXNet, PyTorch, TensorFlow or CuPy. - img: /images/content_images/arlib/tensorly.png - alttext: tensorly - url: http://tensorly.org/stable/home.html - -scientificdomains: - intro: - - text: Nearly every scientist working in Python draws on the power of NumPy. - - text: "NumPy brings the computational power of languages like C and Fortran to Python, a language much easier to learn and use. With this power comes simplicity: a solution in NumPy is often clear and elegant." - - librariesrow1: - - title: Quantum Computing - alttext: A computer chip. - img: /images/content_images/sc_dom_img/quantum_computing.svg - - title: Statistical Computing - alttext: A line graph with the line moving up. - img: /images/content_images/sc_dom_img/statistical_computing.svg - - title: Signal Processing - alttext: A bar chart with positive and negative values. - img: /images/content_images/sc_dom_img/signal_processing.svg - - title: Image Processing - alttext: An photograph of the mountains. - img: /images/content_images/sc_dom_img/image_processing.svg - - title: Graphs and Networks - alttext: A simple graph. - img: /images/content_images/sc_dom_img/sd6.svg - - title: Astronomy Processes - alttext: A telescope. - img: /images/content_images/sc_dom_img/astronomy_processes.svg - - title: Cognitive Psychology - alttext: A human head with gears. - img: /images/content_images/sc_dom_img/cognitive_psychology.svg - - librariesrow2: - - title: Bioinformatics - alttext: A strand of DNA. - img: /images/content_images/sc_dom_img/bioinformatics.svg - - title: Bayesian Inference - alttext: A graph with a bell-shaped curve. - img: /images/content_images/sc_dom_img/bayesian_inference.svg - - title: Mathematical Analysis - alttext: Four mathematical symbols. - img: /images/content_images/sc_dom_img/mathematical_analysis.svg - - title: Chemistry - alttext: A test tube. - img: /images/content_images/sc_dom_img/chemistry.svg - - title: Geoscience - alttext: The Earth. - img: /images/content_images/sc_dom_img/geoscience.svg - - title: Geographic Processing - alttext: A map. - img: /images/content_images/sc_dom_img/GIS.svg - - title: Architecture & Engineering - alttext: A microprocessor development board. - img: /images/content_images/sc_dom_img/robotics.svg - -datascience: - - intro: "NumPy lies at the core of a rich ecosystem of data science libraries. A typical exploratory data science workflow might look like:" - - image1: - - img: /images/content_images/ds-landscape.png - alttext: Diagram of Python Libraries. The five catagories are 'Extract, Transform, Load', 'Data Exploration', 'Data Modeling', 'Data Evaluation' and 'Data Presentation'. - - image2: - - img: /images/content_images/data-science.png - alttext: Diagram of three overlapping circle. The circles labeled 'Mathematics', 'Computer Science' and 'Domain Expertise'. In the middle of the diagram, which has the three circles overlapping it, is an area labeled 'Data Science'. - - examples: - - text: "Extract, Transform, Load: [Pandas](https://pandas.pydata.org),[ Intake](https://intake.readthedocs.io),[PyJanitor](https://pyjanitor.readthedocs.io/)" - - text: "Exploratory analysis: [Jupyter](https://jupyter.org),[Seaborn](https://seaborn.pydata.org),[ Matplotlib](https://matplotlib.org),[ Altair](https://altair-viz.github.io)" - - text: "Model and evaluate: [scikit-learn](https://scikit-learn.org),[ statsmodels](https://www.statsmodels.org/stable/index.html),[ PyMC3](https://docs.pymc.io),[ spaCy](https://spacy.io)" - - text: "Report in a dashboard: [Dash](https://plotly.com/dash),[ Panel](https://panel.holoviz.org),[ Voila](https://github.com/voila-dashboards/voila)" - - content: - - text: For high data volumes, [Dask](https://dask.org) and[Ray](https://ray.io/) are designed to scale. Stabledeployments rely on data versioning ([DVC](https://dvc.org)),experiment tracking ([MLFlow](https://mlflow.org)), andworkflow automation ([Airflow](https://airflow.apache.org) and[Prefect](https://www.prefect.io)). - -visualization: - images: - - url: https://www.fusioncharts.com/blog/best-python-data-visualization-libraries - img: /images/content_images/v_matplotlib.png - alttext: A streamplot made in matplotlib - - url: https://github.com/yhat/ggpy - img: /images/content_images/v_ggpy.png - alttext: A scatter-plot graph made in ggpy - - url: https://www.journaldev.com/19692/python-plotly-tutorial - img: /images/content_images/v_plotly.png - alttext: A box-plot made in plotly - - url: https://altair-viz.github.io/gallery/streamgraph.html - img: /images/content_images/v_altair.png - alttext: A streamgraph made in altair - - url: https://seaborn.pydata.org - img: /images/content_images/v_seaborn.png - alttext: A pairplot of two types of graph, a plot-graph and a frequency graph made in seaborn" - - url: https://docs.pyvista.org/examples/index.html - img: /images/content_images/v_pyvista.png - alttext: A 3D volume rendering made in PyVista. - - url: https://napari.org - img: /images/content_images/v_napari.png - alttext: A multi-dimensionan image made in napari. - - url: https://vispy.org/gallery/index.html - img: /images/content_images/v_vispy.png - alttext: A Voronoi diagram made in vispy. - - content: - - text: NumPy is an essential component in the burgeoning - [Python visualization landscape](https://pyviz.org/overviews/index.html), - which includes [Matplotlib](https://matplotlib.org), - [Seaborn](https://seaborn.pydata.org), [Plotly](https://plot.ly), - [Altair](https://altair-viz.github.io), [Bokeh](https://docs.bokeh.org/en/latest/), - [Holoviz](https://holoviz.org), [Vispy](http://vispy.org), [Napari](https://github.com/napari/napari), - and [PyVista](https://github.com/pyvista/pyvista), to name a few. - - text: NumPy's accelerated processing of large arrays allows researchers to visualize - datasets far larger than native Python could handle. \ No newline at end of file +params: + machinelearning: + paras: + - + para1: O NumPy forma a base de bibliotecas de aprendizagem de máquina poderosas como [scikit-learn](https://scikit-learn.org) e [SciPy](https://www.scipy.org). À medida que a disciplina de aprendizagem de máquina cresce, a lista de bibliotecas construidas a partir do NumPy também cresce. As funcionalidades de deep learning do [TensorFlow](https://www.tensorflow.org) tem diversas aplicações — entre elas, reconhecimento de imagem e de fala, aplicações baseadas em texto, análise de séries temporais, e detecção de vídeo. O [PyTorch](https://pytorch.org), outra biblioteca de deep learning, é popular entre pesquisadores em visão computacional e processamento de linguagem natural. O [MXNet](https://github.com/apache/incubator-mxnet) é outro pacote de IA, que fornece templates e protótipos para deep learning. + para2: Técnicas estatísticas chamadas métodos de [ensemble](https://towardsdatascience.com/ensemble-methods-bagging-boosting-and-stacking-c9214a10a205) tais como binning, bagging, stacking, e boosting estão entre os algoritmos de ML implementados por ferramentas tais como [XGBoost](https://github.com/dmlc/xgboost), [LightGBM](https://lightgbm.readthedocs.io/en/latest/), e [CatBoost](https://catboost.ai) — um dos motores de inferência mais rápidos. [Yellowbrick](https://www.scikit-yb.org/en/latest/) e [Eli5](https://eli5.readthedocs.io/en/latest/) oferecem visualizações para aprendizagem de máquina. + arraylibraries: + intro: + - + text: A API do NumPy é o ponto de partida quando bibliotecas são escritas para explorar hardware inovador, criar tipos de arrays especializados, ou adicionar capacidades além do que o NumPy fornece. + headers: + - + text: Biblioteca de Arrays + - + text: Recursos e áreas de aplicação + libraries: + - + title: Dask + text: Arrays distribuídas e paralelismo avançado para análise, permitindo desempenho em escala. + img: /images/content_images/arlib/dask.png + alttext: Dask + url: https://dask.org/ + - + title: CuPy + text: Biblioteca de matriz compatível com NumPy para computação acelerada pela GPU com Python. + img: /images/content_images/arlib/cupy.png + alttext: CuPy + url: https://cupy.chainer.org + - + title: JAX + text: "Composable transformations of NumPy programs differentiate: vectorize, just-in-time compilation to GPU/TPU." + img: /images/content_images/arlib/jax_logo_250px.png + alttext: JAX + url: https://github.com/google/jax + - + title: Xarray + text: Arrays multidimensionais rotuladas e indexadas para análise e visualização avançadas + img: /images/content_images/arlib/xarray.png + alttext: xarray + url: https://xarray.pydata.org/en/stable/index.html + - + title: Sparse + text: Biblioteca de arrays compatíveis com o NumPy que pode ser integrada com Dask e álgebra linear esparsa da SciPy. + img: /images/content_images/arlib/sparse.png + alttext: sparse + url: https://sparse.pydata.org/en/latest/ + - + title: PyTorch + text: Framework de deep learning que acelera o caminho entre prototipação de pesquisa e colocação em produção. + img: /images/content_images/arlib/pytorch-logo-dark.svg + alttext: PyTorch + url: https://pytorch.org/ + - + title: TensorFlow + text: Uma plataforma completa para aprendizagem de máquina que permite construir e colocar em produção aplicações usando ML facilmente. + img: /images/content_images/arlib/tensorflow-logo.svg + alttext: TensorFlow + url: https://www.tensorflow.org + - + title: MXNet + text: Framework de deep learning voltado para flexibilizar prototipação em pesquisa e produção. + img: /images/content_images/arlib/mxnet_logo.png + alttext: MXNet + url: https://mxnet.apache.org/ + - + title: Arrow + text: Uma plataforma de desenvolvimento multi-linguagens para dados e análise para dados armazenados em colunas na memória. + img: /images/content_images/arlib/arrow.png + alttext: arrow + url: https://github.com/apache/arrow + - + title: xtensor + text: Arrays multidimensionais com broadcasting e avaliação preguiçosa (lazy computing) para análise numérica. + img: /images/content_images/arlib/xtensor.png + alttext: xtensor + url: https://github.com/xtensor-stack/xtensor-python + - + title: Awkward Array + text: Python backend system that decouples API from implementation; unumpy provides a NumPy API. + img: /images/content_images/arlib/awkward.svg + alttext: awkward + url: https://awkward-array.org/ + - + title: uarray + text: Sistema de backend Python que dissocia a API da implementação; unumpy fornece uma API NumPy. + img: /images/content_images/arlib/uarray.png + alttext: uarray + url: https://uarray.org/en/latest/ + scientificdomains: + intro: + - + text: Quase todos os cientistas que trabalham em Python se baseiam na potência do NumPy. + - + text: "NumPy traz o poder computacional de linguagens como C e Fortran para Python, uma linguagem muito mais fácil de aprender e usar. Com esse poder vem a simplicidade: uma solução no NumPy é frequentemente clara e elegante." + libraries: + - + title: Computação quântica + alttext: Um chip de computador. + img: /images/content_images/sc_dom_img/quantum_computing.svg + links: + - + url: http://qutip.org + label: QuTiP + - + url: https://pyquil-docs.rigetti.com/en/stable + label: PyQuil + - + url: https://qiskit.org + label: Qiskit + - + url: https://pennylane.ai + label: PennyLane + - + title: Computação estatística + alttext: Um gráfico com uma linha em movimento para cima. + img: /images/content_images/sc_dom_img/statistical_computing.svg + links: + - + url: https://pandas.pydata.org/ + label: Pandas + - + url: https://github.com/statsmodels/statsmodels + label: statsmodels + - + url: https://xarray.pydata.org/en/stable/ + label: Xarray + - + url: https://github.com/mwaskom/seaborn + label: Seaborn + - + title: Processamento de sinais + alttext: Um gráfico de barras com valores positivos e negativos. + img: /images/content_images/sc_dom_img/signal_processing.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://pywavelets.readthedocs.io/ + label: PyWavelets + - + url: https://python-control.org/ + label: python-control + - + url: https://hyperspy.org/ + label: HyperSpy + - + title: Processamento de imagens + alttext: Uma fotografia das montanhas. + img: /images/content_images/sc_dom_img/image_processing.svg + links: + - + url: https://scikit-image.org/ + label: Scikit-image + - + url: https://opencv.org/ + label: OpenCV + - + url: https://mahotas.rtfd.io/ + label: Mahotas + - + title: Graphs and Networks + alttext: A simple graph. + img: /images/content_images/sc_dom_img/sd6.svg + links: + - + url: https://networkx.org/ + label: NetworkX + - + url: https://graph-tool.skewed.de/ + label: graph-tool + - + url: https://igraph.org/python/ + label: igraph + - + url: https://pygsp.rtfd.io/ + label: PyGSP + - + title: Astronomy + alttext: A telescope. + img: /images/content_images/sc_dom_img/astronomy_processes.svg + links: + - + url: https://www.astropy.org/ + label: AstroPy + - + url: https://github.com/sunpy/sunpy + label: SunPy + - + url: https://github.com/spacepy/spacepy + label: SpacePy + - + title: Psicologia Cognitiva + alttext: Uma cabeça humana com engrenagens. + img: /images/content_images/sc_dom_img/cognitive_psychology.svg + links: + - + url: https://www.psychopy.org/ + label: PsychoPy + - + title: Bioinformática + alttext: Uma fita de DNA. + img: /images/content_images/sc_dom_img/bioinformatics.svg + links: + - + url: https://biopython.org/ + label: BioPython + - + url: http://scikit-bio.org/ + label: Scikit-Bio + - + url: https://github.com/openvax/pyensembl + label: PyEnsembl + - + url: http://etetoolkit.org/ + label: ETE + - + title: Inferência Bayesiana + alttext: Um gráfico com uma curva em forma de sino. + img: /images/content_images/sc_dom_img/bayesian_inference.svg + links: + - + url: https://pystan.readthedocs.io/en/latest/ + label: PyStan + - + url: https://docs.pymc.io/ + label: PyMC3 + - + url: https://arviz-devs.github.io/arviz/ + label: ArviZ + - + url: https://emcee.readthedocs.io/ + label: emcee + - + title: Análise Matemática + alttext: Quatro símbolos matemáticos. + img: /images/content_images/sc_dom_img/mathematical_analysis.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://www.sympy.org/ + label: SymPy + - + url: https://github.com/cvxgrp/cvxpy + label: cvxpy + - + url: https://fenicsproject.org/ + label: FEniCS + - + title: Chemistry + alttext: Um tubo de ensaio. + img: /images/content_images/sc_dom_img/chemistry.svg + links: + - + url: https://cantera.org/ + label: Cantera + - + url: https://www.mdanalysis.org/ + label: MDAnalysis + - + url: https://github.com/rdkit/rdkit + label: RDKit + - + url: https://www.pybamm.org/ + label: PyBaMM + - + title: Geoscience + alttext: A Terra. + img: /images/content_images/sc_dom_img/geoscience.svg + links: + - + url: https://pangeo.io/ + label: Pangeo + - + url: https://simpeg.xyz/ + label: Simpeg + - + url: https://github.com/obspy/obspy/wiki + label: ObsPy + - + url: https://www.fatiando.org/ + label: Fatiando a Terra + - + title: Processamento Geográfico + alttext: Um mapa. + img: /images/content_images/sc_dom_img/GIS.svg + links: + - + url: https://shapely.readthedocs.io/ + label: Shapely + - + url: https://geopandas.org/ + label: GeoPandas + - + url: https://python-visualization.github.io/folium + label: Folium + - + title: Arquitetura e Engenharia + alttext: Uma placa de desenvolvimento de microprocessador. + img: /images/content_images/sc_dom_img/robotics.svg + links: + - + url: https://compas.dev/ + label: COMPAS + - + url: https://cityenergyanalyst.com/ + label: City Energy Analyst + - + url: https://nortikin.github.io/sverchok/ + label: Sverchok + datascience: + intro: "NumPy está no centro de um rico ecossistema de bibliotecas de ciência de dados. Um fluxo de trabalho típico de ciência de dados exploratório pode parecer assim:" + image1: + - + img: /images/content_images/ds-landscape.png + alttext: Diagrama de bibliotecas Python. As cinco categorias são 'Extrair, Transformar, Carregar', 'Exploração de Dados', 'Modelo de Dados', 'Avaliação de Dados' e 'Apresentação de Dados'. + image2: + - + img: /images/content_images/data-science.png + alttext: Diagram of three overlapping circle. The circles labeled 'Mathematics', 'Computer Science' and 'Domain Expertise'. In the middle of the diagram, which has the three circles overlapping it, is an area labeled 'Data Science'. + examples: + - + text: "Extract, Transform, Load: [Pandas](https://pandas.pydata.org),[ Intake](https://intake.readthedocs.io),[PyJanitor](https://pyjanitor.readthedocs.io/)" + - + text: "Exploratory analysis: [Jupyter](https://jupyter.org),[Seaborn](https://seaborn.pydata.org),[ Matplotlib](https://matplotlib.org),[ Altair](https://altair-viz.github.io)" + - + text: "Model and evaluate: [scikit-learn](https://scikit-learn.org),[ statsmodels](https://www.statsmodels.org/stable/index.html),[ PyMC3](https://docs.pymc.io),[ spaCy](https://spacy.io)" + - + text: "Report in a dashboard: [Dash](https://plotly.com/dash),[ Panel](https://panel.holoviz.org),[ Voila](https://github.com/voila-dashboards/voila)" + content: + - + text: For high data volumes, [Dask](https://dask.org) and[Ray](https://ray.io/) are designed to scale. Stabledeployments rely on data versioning ([DVC](https://dvc.org)),experiment tracking ([MLFlow](https://mlflow.org)), andworkflow automation ([Airflow](https://airflow.apache.org) and[Prefect](https://www.prefect.io)). + visualization: + images: + - + url: https://www.fusioncharts.com/blog/best-python-data-visualization-libraries + img: /images/content_images/v_matplotlib.png + alttext: Um streamplot feito em matplotlib + - + url: https://github.com/yhat/ggpy + img: /images/content_images/v_ggpy.png + alttext: Um gráfico scatter-plot feito em ggpy + - + url: https://www.journaldev.com/19692/python-plotly-tutorial + img: /images/content_images/v_plotly.png + alttext: Um box-plot feito no plotly + - + url: https://altair-viz.github.io/gallery/streamgraph.html + img: /images/content_images/v_altair.png + alttext: Um gráfico streamgraph feito em altair + - + url: https://seaborn.pydata.org + img: /images/content_images/v_seaborn.png + alttext: A plot duplo com dois tipos de gráficos, um plot-graph e um gráfico de frequência feitos no seaborn + - + url: https://docs.pyvista.org/ + img: /images/content_images/v_pyvista.png + alttext: Uma renderização de volume 3D feita no PyVista. + - + url: https://napari.org + img: /images/content_images/v_napari.png + alttext: Uma imagem multidimensional, feita em napari. + - + url: https://vispy.org/gallery/index.html + img: /images/content_images/v_vispy.png + alttext: Diagrama de Voronoi feito com vispy. + content: + - + text: NumPy é um componente essencial no crescente [campo de visualização em Python](https://pyviz.org/overviews/index.html), que inclui [Matplotlib](https://matplotlib.org), [Seaborn](https://seaborn.pydata.org), [Plotly](https://plot.ly), [Altair](https://altair-viz.github.io), [Bokeh](https://docs.bokeh.org/en/latest/), [Holoviz](https://holoviz.org), [Vispy](http://vispy.org), [Napari](https://github.com/napari/napari), e [PyVista](https://github.com/pyvista/pyvista), para citar alguns. + - + text: O processamento de grandes arrays acelerado pela NumPy permite que os pesquisadores visualizem conjuntos de dados muito maiores do que o Python nativo poderia permitir. diff --git a/content/pt/teams.md b/content/pt/teams.md index 365aedcf19..fcbe93bcfb 100644 --- a/content/pt/teams.md +++ b/content/pt/teams.md @@ -1,11 +1,9 @@ --- -title: NumPy Teams +title: Times NumPy sidebar: false --- -We are an international team on a mission to support scientific and research -communities worldwide by building quality, open-source software. -[Join us]({{< relref "/contribute" >}})! +Somos uma equipe internacional com a missão de apoiar comunidades científicas e de pesquisa em todo o mundo construindo software de código aberto de qualidade. [Junte-se a nós]({{< relref "/contribute" >}})! {{< include-html "static/gallery/maintainers.html" >}} @@ -19,6 +17,6 @@ communities worldwide by building quality, open-source software. {{< include-html "static/gallery/emeritus-maintainers.html" >}} -# Governance +# Governança -For the list of people on the Steering Council, please see [here](https://numpy.org/devdocs/dev/governance/people.html). +Para a lista de pessoas no Conselho Diretor, veja [aqui](https://numpy.org/devdocs/dev/governance/people.html). diff --git a/content/pt/user-survey-2020.md b/content/pt/user-survey-2020.md new file mode 100644 index 0000000000..bbf3cb6a17 --- /dev/null +++ b/content/pt/user-survey-2020.md @@ -0,0 +1,18 @@ +--- +title: PESQUISA SOBRE A COMUNIDADE NUMPY 2020 +sidebar: false +--- + +Em 2020, o time de pesquisas do NumPy realizou a primeira pesquisa oficial sobre a comunidade NumPy, em parceria com alunos e docentes de um Mestrado em metodologia de pesquisa realizado conjuntamente pela Universidade de Michigan e pela Universidade da Maryland. Mais de 1200 usuários de 75 países participaram para nos ajudar a mapear uma paisagem da comunidade NumPy e expressaram seus pensamentos sobre o futuro do projeto. + +{{< figure src="/surveys/NumPy_usersurvey_2020_report_cover.png" class="fig-left" alt="Página de capa do relatório da pesquisa de usuários do NumPy 2020, chamado 'NumPy Community Survey 2020 - results'" width="250">}} +src = '/surveys/NumPy_usersurvey_2020_report_cover. ng' alt = 'Página de rosto do relatório da pesquisa de usuário NumPy de 2020, intitulado "NumPy Community Survey 2020 - resultados"' width = '250' +{{< /figure >}} + +**[Faça o download do relatório](/surveys/NumPy_usersurvey_2020_report.pdf)** para ver os detalhes sobre os resultados encontrados. + + +Para os destaques, confira **[este infográfico](https://github.com/numpy/numpy-surveys/blob/master/images/2020NumPysurveyresults_community_infographic.pdf)**. + +Quer saber mais? Visite **https://numpy.org/user-survey-2020-details/**. + diff --git a/content/pt/user-surveys.md b/content/pt/user-surveys.md new file mode 100644 index 0000000000..4f60686926 --- /dev/null +++ b/content/pt/user-surveys.md @@ -0,0 +1,10 @@ +--- +title: PESQUISA DE USUÁRIOS NUMPY +sidebar: false +--- + +**2020** O time de pesquisas da NumPy, em parceria com estudantes e professores da Universidade de Michigan e da Universidade de Maryland, conduziram a primeira pesquisa oficial sobre a comunidade NumPy. Você pode encontrar os resultados da pesquisa [aqui (em inglês)](https://numpy.org/user-survey-2020/). + +**2021** Os dados coletados estão em análise. + +Se você tem dúvidas ou sugestões sobre as pesquisas já realizadas ou futuras, por favor crie uma issue [aqui](https://github.com/numpy/numpy-surveys/issues). diff --git a/content/ru/404.md b/content/ru/404.md new file mode 100644 index 0000000000..da192c53c0 --- /dev/null +++ b/content/ru/404.md @@ -0,0 +1,8 @@ +--- +title: 404 +sidebar: false +--- + +Oops! You've reached a dead end. + +If you think something should be here, you can [open an issue](https://github.com/numpy/numpy.org/issues) on GitHub. diff --git a/content/ru/_index.md b/content/ru/_index.md new file mode 100644 index 0000000000..be88f9e642 --- /dev/null +++ b/content/ru/_index.md @@ -0,0 +1,49 @@ +--- +title: null +--- + +{{< grid columns="1 2 2 3" >}} + +[[item]] +type = 'card' +title = 'Powerful N-dimensional arrays' +body = ''' +Fast and versatile, the NumPy vectorization, indexing, and broadcasting concepts are the de-facto standards of array computing today. +''' + +[[item]] +type = 'card' +title = 'Numerical computing tools' +body = ''' +NumPy offers comprehensive mathematical functions, random number generators, linear algebra routines, Fourier transforms, and more. +''' + +[[item]] +type = 'card' +title = 'Open source' +body = ''' +Distributed under a liberal [BSD license](https://github.com/numpy/numpy/blob/main/LICENSE.txt), NumPy is developed and maintained [publicly on GitHub](https://github.com/numpy/numpy) by a vibrant, responsive, and diverse [community](/community). +''' + +[[item]] +type = 'card' +title = 'Interoperable' +body = ''' +NumPy supports a wide range of hardware and computing platforms, and plays well with distributed, GPU, and sparse array libraries. +''' + +[[item]] +type = 'card' +title = 'Performant' +body = ''' +The core of NumPy is well-optimized C code. Enjoy the flexibility of Python with the speed of compiled code. +''' + +[[item]] +type = 'card' +title = 'Easy to use' +body = ''' +NumPy's high level syntax makes it accessible and productive for programmers from any background or experience level. +''' + +{{< /grid>}} diff --git a/content/ru/about.md b/content/ru/about.md new file mode 100644 index 0000000000..243f7083ae --- /dev/null +++ b/content/ru/about.md @@ -0,0 +1,90 @@ +--- +title: About Us +sidebar: false +--- + +NumPy is an open source project that enables numerical computing with Python. It was created in 2005 building on the early work of the Numeric and Numarray libraries. NumPy will always be 100% open source software and free for all to use. It is released under the liberal terms of the [modified BSD license](https://github.com/numpy/numpy/blob/main/LICENSE.txt). + +NumPy is developed in the open on GitHub, through the consensus of the NumPy and wider scientific Python community. For more information on our governance approach, please see our [Governance Document](https://www.numpy.org/devdocs/dev/governance/index.html). + + +## Steering Council + +The NumPy Steering Council is the project's governing body. Its role is to ensure, through working with and serving the broader NumPy community, the long-term sustainability of the project, both as a software package and community. The NumPy Steering Council currently consists of the following members (in alphabetical order, by last name): + +- Sebastian Berg +- Ralf Gommers +- Charles Harris +- Stephan Hoyer +- Inessa Pawson +- Matti Picus +- Stéfan van der Walt +- Melissa Weber Mendonça +- Eric Wieser + +Emeritus: + +- Alex Griffing (2015-2017) +- Allan Haldane (2015-2021) +- Marten van Kerkwijk (2017-2019) +- Travis Oliphant (project founder, 2005-2012) +- Nathaniel Smith (2012-2021) +- Julian Taylor (2013-2021) +- Jaime Fernández del Río (2014-2021) +- Pauli Virtanen (2008-2021) + +To contact the NumPy Steering Council, please email numpy-team@googlegroups.com. + +## Teams + +The NumPy project leadership is actively working on diversifying contribution pathways to the project.
NumPy currently has the following teams: + +- development +- documentation +- triage +- website +- survey +- translations +- sprint mentors +- optimization +- funding and grants + +See the [Team](/teams) page for more info. + +## NumFOCUS Subcommittee + +- Charles Harris +- Ralf Gommers +- Inessa Pawson +- Sebastian Berg +- External member: Thomas Caswell + +## Sponsors + +NumPy receives direct funding from the following sources: +{{< sponsors >}} + + +## Institutional Partners + +Institutional Partners are organizations that support the project by employing people that contribute to NumPy as part of their job. Current Institutional Partners include: + +- UC Berkeley (Stéfan van der Walt) +- Quansight (Nathan Goldbaum, Ralf Gommers, Matti Picus, Melissa Weber Mendonça) +- NVIDIA (Sebastian Berg) + +{{< partners >}} + + +## Donate + +If you have found NumPy useful in your work, research, or company, please consider a donation to the project commensurate with your resources. Any amount helps! All donations will be used strictly to fund the development of NumPy’s open source software, documentation, and community. + +NumPy is a Sponsored Project of NumFOCUS, a 501(c)(3) nonprofit charity in the United States. NumFOCUS provides NumPy with fiscal, legal, and administrative support to help ensure the health and sustainability of the project. Visit [numfocus.org](https://numfocus.org) for more information. + +Donations to NumPy are managed by [NumFOCUS](https://numfocus.org). For donors in the United States, your gift is tax-deductible to the extent provided by law. As with any donation, you should consult with your tax advisor about your particular tax situation. + +NumPy's Steering Council will make the decisions on how to best use any funds received. Technical and infrastructure priorities are documented on the [NumPy Roadmap](https://www.numpy.org/neps/index.html#roadmap). + +{{}} + diff --git a/content/ru/arraycomputing.md b/content/ru/arraycomputing.md new file mode 100644 index 0000000000..abd29d11c1 --- /dev/null +++ b/content/ru/arraycomputing.md @@ -0,0 +1,21 @@ +--- +title: Array Computing +sidebar: false +--- + +*Array computing is the foundation of statistical, mathematical, scientific computing in various contemporary data science and analytics applications such as data visualization, digital signal processing, image processing, bioinformatics, machine learning, AI, and several others.* + +Large scale data manipulation and transformation depends on efficient, high-performance array computing. The language of choice for data analytics, machine learning, and productive numerical computing is **Python.** + +**Num**erical **Py**thon or NumPy is its de-facto standard Python programming language library that supports large, multi-dimensional arrays and matrices, and comes with a vast collection of high-level mathematical functions to operate on these arrays. + +Since the launch of NumPy in 2006, Pandas appeared on the landscape in 2008, and it was not until a couple of years ago that several array computing libraries showed up in succession, crowding the array computing landscape. Many of these newer libraries mimic NumPy-like features and capabilities, and pack newer algorithms and features geared towards machine learning and artificial intelligence applications. + +arraycl + +**Array computing** is based on **arrays** data structures. *Arrays* are used to organize vast amounts of data such that a related set of values can be easily sorted, searched, mathematically manipulated, and transformed easily and quickly. + +Array computing is *unique* as it involves operating on the data array *at once*. What this means is that any array operation applies to an entire set of values in one shot. This vectorized approach provides speed and simplicity by enabling programmers to code and operate on aggregates of data, without having to use loops of individual scalar operations. diff --git a/content/ru/case-studies/blackhole-image.md b/content/ru/case-studies/blackhole-image.md new file mode 100644 index 0000000000..22334ebed0 --- /dev/null +++ b/content/ru/case-studies/blackhole-image.md @@ -0,0 +1,80 @@ +--- +title: "Case Study: First Image of a Black Hole" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/blackhole.jpg' title = 'Black Hole M87' alt = 'black hole image' attribution = '(Image Credits: Event Horizon Telescope Collaboration)' attributionlink = 'https://www.jpl.nasa.gov/images/universe/20190410/blackhole20190410.jpg' +{{< /figure >}} + +{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="Katie Bouman, *Assistant Professor, Computing & Mathematical Sciences, Caltech*" +> }} Imaging the M87 Black Hole is like trying to see something that is by definition impossible to see. +> +> {{< /blockquote >}} + +## A telescope the size of the earth + +The [Event Horizon telescope (EHT)](https://eventhorizontelescope.org) is an array of eight ground-based radio telescopes forming a computational telescope the size of the earth, studing the universe with unprecedented sensitivity and resolution. The huge virtual telescope, which uses a technique called very-long-baseline interferometry (VLBI), has an angular resolution of [20 micro-arcseconds][resolution] — enough to read a newspaper in New York from a sidewalk café in Paris! + +### Key Goals and Results + +* **A New View of the Universe:** The groundwork for the EHT's groundbreaking image had been laid 100 years earlier when [Sir Arthur Eddington][eddington] yielded the first observational support of Einstein's theory of general relativity. + +* **The Black Hole:** EHT was trained on a supermassive black hole approximately 55 million light-years from Earth, lying at the center of the galaxy Messier 87 (M87) in the Virgo galaxy cluster. Its mass is 6.5 billion times the Sun's. It had been studied for [over 100 years](https://www.jpl.nasa.gov/news/news.php?feature=7385), but never before had a black hole been visually observed. + +* **Comparing Observations to Theory:** From Einstein’s general theory of relativity, scientists expected to find a shadow-like region caused by gravitational bending and capture of light. Scientists could use it to measure the black hole's enormous mass. + +### The Challenges + +* **Computational scale** + + EHT poses massive data-processing challenges, including rapid atmospheric phase fluctuations, large recording bandwidth, and telescopes that are widely dissimilar and geographically dispersed. + +* **Too much information** + + Each day EHT generates over 350 terabytes of observations, stored on helium-filled hard drives. Reducing the volume and complexity of this much data is enormously difficult. + +* **Into the unknown** + + When the goal is to see something never before seen, how can scientists be confident the image is correct? + +{{< figure >}} +src = '/images/content_images/cs/dataprocessbh.png' title = 'EHT Data Processing Pipeline' alt = 'data pipeline' align = 'center' attribution = '(Diagram Credits: The Astrophysical Journal, Event Horizon Telescope Collaboration)' attributionlink = 'https://iopscience.iop.org/article/10.3847/2041-8213/ab0c57' +{{< /figure >}} + +## NumPy’s Role + +What if there's a problem with the data? Or perhaps an algorithm relies too heavily on a particular assumption. Will the image change drastically if a single parameter is changed? + +The EHT collaboration met these challenges by having independent teams evaluate the data, using both established and cutting-edge image reconstruction techniques. When results proved consistent, they were combined to yield the first-of-a-kind image of the black hole. + +Their work illustrates the role the scientific Python ecosystem plays in advancing science through collaborative data analysis. + +{{< figure >}} +src = '/images/content_images/cs/bh_numpy_role.png' alt = 'role of numpy' title = 'The role of NumPy in Black Hole imaging' +{{< /figure >}} + +For example, the [`eht-imaging`][ehtim] Python package provides tools for simulating and performing image reconstruction on VLBI data. NumPy is at the core of array data processing used in this package, as illustrated by the partial software dependency chart below. + +{{< figure >}} +src = '/images/content_images/cs/ehtim_numpy.png' alt = 'ehtim dependency map highlighting numpy' title = 'Software dependency chart of ehtim package highlighting NumPy' +{{< /figure >}} + +Besides NumPy, many other packages, such as [SciPy](https://www.scipy.org) and [Pandas](https://pandas.io), are part of the data processing pipeline for imaging the black hole. The standard astronomical file formats and time/coordinate transformations were handled by [Astropy][astropy], while [Matplotlib][mpl] was used in visualizing data throughout the analysis pipeline, including the generation of the final image of the black hole. + +## Summary + +The efficient and adaptable n-dimensional array that is NumPy's central feature enabled researchers to manipulate large numerical datasets, providing a foundation for the first-ever image of a black hole. A landmark moment in science, it gives stunning visual evidence of Einstein’s theory. The achievement encompasses not only technological breakthroughs but also international collaboration among over 200 scientists and some of the world's best radio observatories. Innovative algorithms and data processing techniques, improving upon existing astronomical models, helped unfold a mystery of the universe. + +{{< figure >}} +src = '/images/content_images/cs/numpy_bh_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} + +[resolution]: https://eventhorizontelescope.org/press-release-april-10-2019-astronomers-capture-first-image-black-hole + +[eddington]: https://en.wikipedia.org/wiki/Eddington_experiment + +[ehtim]: https://github.com/achael/eht-imaging + +[astropy]: https://www.astropy.org/ +[mpl]: https://matplotlib.org/ diff --git a/content/ru/case-studies/cricket-analytics.md b/content/ru/case-studies/cricket-analytics.md new file mode 100644 index 0000000000..d0be271f50 --- /dev/null +++ b/content/ru/case-studies/cricket-analytics.md @@ -0,0 +1,72 @@ +--- +title: "Case Study: Cricket Analytics, the game changer!" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/ipl-stadium.png' title = 'IPLT20, the biggest Cricket Festival in India' alt = 'Indian Premier League Cricket cup and stadium' attribution = '(Image credits: IPLT20 (cup and logo) & Akash Yadav (stadium))' attributionlink = 'https://unsplash.com/@aksh1802' +{{< /figure >}} + +{{< blockquote cite="https://www.scoopwhoop.com/sports/ms-dhoni/" by="M S Dhoni, *International Cricket Player, ex-captain, Indian Team, plays for Chennai Super Kings in IPL*" +> }} You don't play for the crowd, you play for the country. +> +> {{< /blockquote >}} + +## About Cricket + +It would be an understatement to state that Indians love cricket. The game is played in just about every nook and cranny of India, rural or urban, popular with the young and the old alike, connecting billions in India unlike any other sport. Cricket enjoys lots of media attention. There is a significant amount of [money](https://www.statista.com/topics/4543/indian-premier-league-ipl/) and fame at stake. Over the last several years, technology has literally been a game changer. Audiences are spoilt for choice with streaming media, tournaments, affordable access to mobile based live cricket watching, and more. + +The Indian Premier League (IPL) is a professional Twenty20 cricket league, founded in 2008. It is one of the most attended cricketing events in the world, valued at [$6.7 billion](https://en.wikipedia.org/wiki/Indian_Premier_League) in 2019. + +Cricket is a game of numbers - the runs scored by a batsman, the wickets taken by a bowler, the matches won by a cricket team, the number of times a batsman responds in a certain way to a kind of bowling attack, etc. The capability to dig into cricketing numbers for both improving performance and studying the business opportunities, overall market, and economics of cricket via powerful analytics tools, powered by numerical computing software such as NumPy, is a big deal. Cricket analytics provides interesting insights into the game and predictive intelligence regarding game outcomes. + +Today, there are rich and almost infinite troves of cricket game records and statistics available, e.g., [ESPN cricinfo](https://stats.espncricinfo.com/ci/engine/stats/index.html) and [cricsheet](https://cricsheet.org). These and several such cricket databases have been used for [cricket analysis](https://www.researchgate.net/publication/336886516_Data_visualization_and_toss_related_analysis_of_IPL_teams_and_batsmen_performances) using the latest machine learning and predictive modelling algorithms. Media and entertainment platforms along with professional sports bodies associated with the game use technology and analytics for determining key metrics for improving match winning chances: + +* batting performance moving average, +* score forecasting, +* gaining insights into fitness and performance of a player against different opposition, +* player contribution to wins and losses for making strategic decisions on team composition + +{{< figure >}} +src = '/images/content_images/cs/cricket-pitch.png' title = 'Cricket Pitch, the focal point in the field' alt = 'A cricket pitch with bowler and batsmen' align = 'center' attribution = '(Image credit: Debarghya Das)' attributionlink = 'http://debarghyadas.com/files/IPLpaper.pdf' +{{< /figure >}} + +### Key Data Analytics Objectives + +* Sports data analytics are used not only in cricket but many [other sports](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx) for improving the overall team performance and maximizing winning chances. +* Real-time data analytics can help in gaining insights even during the game for changing tactics by the team and by associated businesses for economic benefits and growth. +* Besides historical analysis, predictive models are harnessed to determine the possible match outcomes that require significant number crunching and data science know-how, visualization tools and capability to include newer observations in the analysis. + +{{< figure >}} +src = '/images/content_images/cs/player-pose-estimator.png' alt = 'pose estimator' title = 'Cricket Pose Estimator' attribution = '(Image credit: connect.vin)' attributionlink = 'https://connect.vin/2019/05/ai-for-cricket-batsman-pose-analysis/' +{{< /figure >}} + +### The Challenges + +* **Data Cleaning and preprocessing** + + IPL has expanded cricket beyond the classic test match format to a much larger scale. The number of matches played every season across various formats has increased and so has the data, the algorithms, newer sports data analysis technologies and simulation models. Cricket data analysis requires field mapping, player tracking, ball tracking, player shot analysis, and several other aspects involved in how the ball is delivered, its angle, spin, velocity, and trajectory. All these factors together have increased the complexity of data cleaning and preprocessing. + +* **Dynamic Modeling** + + In cricket, just like any other sport, there can be a large number of variables related to tracking various numbers of players on the field, their attributes, the ball, and several possibilities of potential actions. The complexity of data analytics and modeling is directly proportional to the kind of predictive questions that are put forth during analysis and are highly dependent on data representation and the model. Things get even more challenging in terms of computation, data comparisons when dynamic cricket play predictions are sought such as what would have happened if the batsman had hit the ball at a different angle or velocity. + +* **Predictive Analytics Complexity** + + Much of the decision making in cricket is based on questions such as "how often does a batsman play a certain kind of shot if the ball delivery is of a particular type", or "how does a bowler change his line and length if the batsman responds to his delivery in a certain way". This kind of predictive analytics query requires highly granular dataset availability and the capability to synthesize data and create generative models that are highly accurate. + +## NumPy’s Role in Cricket Analytics + +Sports Analytics is a thriving field. Many researchers and companies [use NumPy](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx) and other PyData packages like Scikit-learn, SciPy, Matplotlib, and Jupyter, besides using the latest machine learning and AI techniques. NumPy has been used for various kinds of cricket related sporting analytics such as: + +* **Statistical Analysis:** NumPy's numerical capabilities help estimate the statistical significance of observational data or match events in the context of various player and game tactics, estimating the game outcome by comparison with a generative or static model. [Causal analysis](https://amplitude.com/blog/2017/01/19/causation-correlation) and [big data approaches](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4996805/) are used for tactical analysis. + +* **Data Visualization:** Data graphing and [visualization](https://towardsdatascience.com/advanced-sports-visualization-with-pandas-matplotlib-and-seaborn-9c16df80a81b) provide useful insights into relationship between various datasets. + +## Summary + +Sports Analytics is a game changer when it comes to how professional games are played, especially how strategic decision making happens, which until recently was primarily done based on “gut feeling" or adherence to past traditions. NumPy forms a solid foundation for a large set of Python packages which provide higher level functions related to data analytics, machine learning, and AI algorithms. These packages are widely deployed to gain real-time insights that help in decision making for game-changing outcomes, both on field as well as to draw inferences and drive business around the game of cricket. Finding out the hidden parameters, patterns, and attributes that lead to the outcome of a cricket match helps the stakeholders to take notice of game insights that are otherwise hidden in numbers and statistics. + +{{< figure >}} +src = '/images/content_images/cs/numpy_ca_benefits.png' alt = 'Diagram showing benefits of using NumPy for cricket analytics' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} diff --git a/content/ru/case-studies/deeplabcut-dnn.md b/content/ru/case-studies/deeplabcut-dnn.md new file mode 100644 index 0000000000..257905de91 --- /dev/null +++ b/content/ru/case-studies/deeplabcut-dnn.md @@ -0,0 +1,102 @@ +--- +title: "Case Study: DeepLabCut 3D Pose Estimation" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/mice-hand.gif' title = 'Analyzing mice hand-movement using DeepLapCut' alt = 'micehandanim' attribution = '(Source: www.deeplabcut.org )' attributionlink = 'http://www.mousemotorlab.org/deeplabcut' +{{< /figure >}} + +{{< blockquote cite="https://news.harvard.edu/gazette/story/newsplus/harvard-researchers-awarded-czi-open-source-award/" by="Alexander Mathis, *Assistant Professor, École polytechnique fédérale de Lausanne* ([EPFL](https://www.epfl.ch/en/))" +> }} Open Source Software is accelerating Biomedicine. DeepLabCut enables automated video analysis of animal behavior using Deep Learning. +> +> {{< /blockquote >}} + +## About DeepLabCut + +[DeepLabCut](https://github.com/DeepLabCut/DeepLabCut) is an open source toolbox that empowers researchers at hundreds of institutions worldwide to track behaviour of laboratory animals, with very little training data, at human-level accuracy. With DeepLabCut technology, scientists can delve deeper into the scientific understanding of motor control and behavior across animal species and timescales. + +Several areas of research, including neuroscience, medicine, and biomechanics, use data from tracking animal movement. DeepLabCut helps in understanding what humans and other animals are doing by parsing actions that have been recorded on film. Using automation for laborious tasks of tagging and monitoring, along with deep neural network based data analysis, DeepLabCut makes scientific studies involving observing animals, such as primates, mice, fish, flies etc., much faster and more accurate. + +{{< figure >}} +src = '/images/content_images/cs/race-horse.gif' title = 'Colored dots track the positions of a racehorse’s body part' alt = 'horserideranim' attribution = '(Source: Mackenzie Mathis)' +{{< /figure >}} + +DeepLabCut's non-invasive behavioral tracking of animals by extracting the poses of animals is crucial for scientific pursuits in domains such as biomechanics, genetics, ethology & neuroscience. Measuring animal poses non-invasively from video - without markers - in dynamically changing backgrounds is computationally challenging, both technically as well as in terms of resource needs and training data required. + +DeepLabCut allows researchers to estimate the pose of the subject, efficiently enabling them to quantify the behavior through a Python based software toolkit. With DeepLabCut, researchers can identify distinct frames from videos, digitally label specific body parts in a few dozen frames with a tailored GUI, and then the deep learning based pose estimation architectures in DeepLabCut learn how to pick out those same features in the rest of the video and in other similar videos of animals. It works across species of animals, from common laboratory animals such as flies and mice to more unusual animals like [cheetahs][cheetah-movement]. + +DeepLabCut uses a principle called [transfer learning](https://arxiv.org/pdf/1909.11229), which greatly reduces the amount of training data required and speeds up the convergence of the training period. Depending on the needs, users can pick different network architectures that provide faster inference (e.g. MobileNetV2), which can also be combined with real-time experimental feedback. DeepLabCut originally used the feature detectors from a top-performing human pose estimation architecture, called [DeeperCut](https://arxiv.org/abs/1605.03170), which inspired the name. The package now has been significantly changed to include additional architectures, augmentation methods, and a full front-end user experience. Furthermore, to support large-scale biological experiments DeepLabCut provides active learning capabilities so that users can increase the training set over time to cover edge cases and make their pose estimation algorithm robust within the specific context. + +Recently, the [DeepLabCut model zoo](http://www.mousemotorlab.org/dlc-modelzoo) was introduced, which provides pre-trained models for various species and experimental conditions from facial analysis in primates to dog posture. This can be run for instance in the cloud without any labeling of new data, or neural network training, and no programming experience is necessary. + +### Key Goals and Results + +* **Automation of animal pose analysis for scientific studies:** + + The primary objective of DeepLabCut technology is to measure and track posture of animals in a diverse settings. This data can be used, for example, in neuroscience studies to understand how the brain controls movement, or to elucidate how animals socially interact. Researchers have observed a [tenfold performance boost](https://www.biorxiv.org/content/10.1101/457242v1) with DeepLabCut. Poses can be inferred offline at up to 1200 frames per second (FPS). + +* **Creation of an easy-to-use Python toolkit for pose estimation:** + + DeepLabCut wanted to share their animal pose-estimation technology in the form of an easy to use tool that can be adopted by researchers easily. So they have created a complete, easy-to-use Python toolbox with project management features as well. These enable not only automation of pose-estimation but also managing the project end-to-end by helping the DeepLabCut Toolkit user right from the dataset collection stage to creating shareable and reusable analysis pipelines. + + Their [toolkit][DLCToolkit] is now available as open source. + + A typical DeepLabCut Workflow includes: + + - creation and refining of training sets via active learning + - creation of tailored neural networks for specific animals and scenarios + - code for large-scale inference on videos + - draw inferences using integrated visualization tools + +{{< figure >}} +src = '/images/content_images/cs/deeplabcut-toolkit-steps.png' title = 'Pose estimation steps with DeepLabCut' alt = 'dlcsteps' align = 'center' attribution = '(Source: DeepLabCut)' attributionlink = 'https://twitter.com/DeepLabCut/status/1198046918284210176/photo/1' +{{< /figure >}} + +### The Challenges + +* **Speed** + + Fast processing of animal behavior videos in order to measure their behavior and at the same time make scientific experiments more efficient, accurate. Extracting detailed animal poses for laboratory experiments, without markers, in dynamically changing backgrounds, can be challenging, both technically as well as in terms of resource needs and training data required. Coming up with a tool that is easy to use without the need for skills such as computer vision expertise that enables scientists to do research in more real-world contexts, is a non-trivial problem to solve. + +* **Combinatorics** + + Combinatorics involves assembly and integration of movement of multiple limbs into individual animal behavior. Assembling keypoints and their connections into individual animal movements and linking them across time is a complex process that requires heavy-duty numerical analysis, especially in case of multi-animal movement tracking in experiment videos. + +* **Data Processing** + + Last but not the least, array manipulation - processing large stacks of arrays corresponding to various images, target tensors and keypoints is fairly challenging. + +{{< figure >}} +src = '/images/content_images/cs/pose-estimation.png' title = 'Pose estimation variety and complexity' alt = 'challengesfig' align = 'center' attribution = '(Source: Mackenzie Mathis)' attributionlink = 'https://www.biorxiv.org/content/10.1101/476531v1.full.pdf' +{{< /figure >}} + +## NumPy's Role in meeting Pose Estimation Challenges + +NumPy addresses DeepLabCut technology's core need of numerical computations at high speed for behavioural analytics. Besides NumPy, DeepLabCut employs various Python software that utilize NumPy at their core, such as [SciPy](https://www.scipy.org), [Pandas](https://pandas.pydata.org), [matplotlib](https://matplotlib.org), [Tensorpack](https://github.com/tensorpack/tensorpack), [imgaug](https://github.com/aleju/imgaug), [scikit-learn](https://scikit-learn.org/stable/), [scikit-image](https://scikit-image.org) and [Tensorflow](https://www.tensorflow.org). + +The following features of NumPy played a key role in addressing the image processing, combinatorics requirements and need for fast computation in DeepLabCut pose estimation algorithms: + +* Vectorization +* Masked Array Operations +* Linear Algebra +* Random Sampling +* Reshaping of large arrays + +DeepLabCut utilizes NumPy’s array capabilities throughout the workflow offered by the toolkit. In particular, NumPy is used for sampling distinct frames for human annotation labeling, and for writing, editing and processing annotation data. Within TensorFlow the neural network is trained by DeepLabCut technology over thousands of iterations to predict the ground truth annotations from frames. For this purpose, target densities (scoremaps) are created to cast pose estimation as a image-to-image translation problem. To make the neural networks robust, data augmentation is employed, which requires the calculation of target scoremaps subject to various geometric and image processing steps. To make training fast, NumPy’s vectorization capabilities are leveraged. For inference, the most likely predictions from target scoremaps need to extracted and one needs to efficiently “link predictions to assemble individual animals”. + +{{< figure >}} +src = '/images/content_images/cs/deeplabcut-workflow.png' title = 'DeepLabCut Workflow' alt = 'workflow' attribution = '(Source: Mackenzie Mathis)' attributionlink = 'https://www.researchgate.net/figure/DeepLabCut-work-flow-The-diagram-delineates-the-work-flow-as-well-as-the-directory-and_fig1_329185962' +{{< /figure >}} + +## Summary + +Observing and efficiently describing behavior is a core tenant of modern ethology, neuroscience, medicine, and technology. [DeepLabCut](http://orga.cvss.cc/wp-content/uploads/2019/05/NathMathis2019.pdf) allows researchers to estimate the pose of the subject, efficiently enabling them to quantify the behavior. With only a small set of training images, the DeepLabCut Python toolbox allows training a neural network to within human level labeling accuracy, thus expanding its application to not only behavior analysis in the laboratory, but to potentially also in sports, gait analysis, medicine and rehabilitation studies. Complex combinatorics, data processing challenges faced by DeepLabCut algorithms are addressed through the use of NumPy's array manipulation capabilities. + +{{< figure >}} +src = '/images/content_images/cs/numpy_dlc_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} + +[cheetah-movement]: https://www.technologynetworks.com/neuroscience/articles/interview-a-deeper-cut-into-behavior-with-mackenzie-mathis-327618 + +[DLCToolkit]: https://github.com/DeepLabCut/DeepLabCut diff --git a/content/ru/case-studies/gw-discov.md b/content/ru/case-studies/gw-discov.md new file mode 100644 index 0000000000..08f54b7547 --- /dev/null +++ b/content/ru/case-studies/gw-discov.md @@ -0,0 +1,77 @@ +--- +title: "Case Study: Discovery of Gravitational Waves" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/gw_sxs_image.png' title = 'Gravitational Waves' alt = 'binary coalesce black hole generating gravitational waves' attribution = '(Image Credits: The Simulating eXtreme Spacetimes (SXS) Project at LIGO)' attributionlink = 'https://youtu.be/Zt8Z_uzG71o' +{{< /figure >}} + +{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="David Shoemaker, *LIGO Scientific Collaboration*" >}} The scientific Python ecosystem is critical infrastructure for the research done at LIGO. +{{< /blockquote >}} + +## About [Gravitational Waves](https://www.nationalgeographic.com/news/2017/10/what-are-gravitational-waves-ligo-astronomy-science/) and [LIGO](https://www.ligo.caltech.edu) + +Gravitational waves are ripples in the fabric of space and time, generated by cataclysmic events in the universe such as collision and merging of two black holes or coalescing binary stars or supernovae. Observing GW can not only help in studying gravity but also in understanding some of the obscure phenomena in the distant universe and its impact. + +The [Laser Interferometer Gravitational-Wave Observatory (LIGO)](https://www.ligo.caltech.edu) was designed to open the field of gravitational-wave astrophysics through the direct detection of gravitational waves predicted by Einstein’s General Theory of Relativity. It comprises two widely separated interferometers within the United States — one in Hanford, Washington and the other in Livingston, Louisiana — operated in unison to detect gravitational waves. Each of them has multi-kilometer-scale gravitational wave detectors that use laser interferometry. The LIGO Scientific Collaboration (LSC), is a group of more than 1000 scientists from universities around the United States and in 14 other countries supported by more than 90 universities and research institutes; approximately 250 students actively contributing to the collaboration. The new LIGO discovery is the first observation of gravitational waves themselves, made by measuring the tiny disturbances the waves make to space and time as they pass through the earth. It has opened up new astrophysical frontiers that explore the warped side of the universe—objects and phenomena that are made from warped spacetime. + + +### Key Objectives + +* Though its [mission](https://www.ligo.caltech.edu/page/what-is-ligo) is to detect gravitational waves from some of the most violent and energetic processes in the Universe, the data LIGO collects may have far-reaching effects on many areas of physics including gravitation, relativity, astrophysics, cosmology, particle physics, and nuclear physics. +* Crunch observed data via numerical relativity computations that involves complex maths in order to discern signal from noise, filter out relevant signal and statistically estimate significance of observed data +* Data visualization so that the binary / numerical results can be comprehended. + + + +### The Challenges + +* **Computation** + + Gravitational Waves are hard to detect as they produce a very small effect and have tiny interaction with matter. Processing and analyzing all of LIGO's data requires a vast computing infrastructure.After taking care of noise, which is billions of times of the signal, there is still very complex relativity equations and huge amounts of data which present a computational challenge: [O(10^7) CPU hrs needed for binary merger analyses](https://youtu.be/7mcHknWWzNI) spread on 6 dedicated LIGO clusters + +* **Data Deluge** + + As observational devices become more sensitive and reliable, the challenges posed by data deluge and finding a needle in a haystack rise multi-fold. LIGO generates terabytes of data every day! Making sense of this data requires an enormous effort for each and every detection. For example, the signals being collected by LIGO must be matched by supercomputers against hundreds of thousands of templates of possible gravitational-wave signatures. + +* **Visualization** + + Once the obstacles related to understanding Einstein’s equations well enough to solve them using supercomputers are taken care of, the next big challenge was making data comprehensible to the human brain. Simulation modeling as well as signal detection requires effective visualization techniques. Visualization also plays a role in lending more credibility to numerical relativity in the eyes of pure science aficionados, who did not give enough importance to numerical relativity until imaging and simulations made it easier to comprehend results for a larger audience. Speed of complex computations and rendering, re-rendering images and simulations using latest experimental inputs and insights can be a time consuming activity that challenges researchers in this domain. + +{{< figure >}} +src = '/images/content_images/cs/gw_strain_amplitude.png' alt = 'gravitational waves strain amplitude' title = 'Estimated gravitational-wave strain amplitude from GW150914' attribution = '(Graph Credits: Observation of Gravitational Waves from a Binary Black Hole Merger, ResearchGate Publication)' attributionlink = 'https://www.researchgate.net/publication/293886905_Observation_of_Gravitational_Waves_from_a_Binary_Black_Hole_Merger' +{{< /figure >}} + +## NumPy’s Role in the Detection of Gravitational Waves + +Gravitational waves emitted from the merger cannot be computed using any technique except brute force numerical relativity using supercomputers. The amount of data LIGO collects is as incomprehensibly large as gravitational wave signals are small. + +NumPy, the standard numerical analysis package for Python, was utilized by the software used for various tasks performed during the GW detection project at LIGO. NumPy helped in solving complex maths and data manipulation at high speed. Here are some examples: + +* [Signal Processing](https://www.uv.es/virgogroup/Denoising_ROF.html): Glitch detection, [Noise identification and Data Characterization](https://ep2016.europython.eu/media/conference/slides/pyhton-in-gravitational-waves-research-communities.pdf) (NumPy, scikit-learn, scipy, matplotlib, pandas, pyCharm) +* Data retrieval: Deciding which data can be analyzed, figuring out whether it contains a signal - needle in a haystack +* Statistical analysis: estimate the statistical significance of observational data, estimating the signal parameters (e.g. masses of stars, spin velocity, and distance) by comparison with a model. +* Visualization of data + - Time series + - Spectrograms +* Compute Correlations +* Key [Software](https://github.com/lscsoft) developed in GW data analysis such as [GwPy](https://gwpy.github.io/docs/stable/overview.html) and [PyCBC](https://pycbc.org) uses NumPy and AstroPy under the hood for providing object based interfaces to utilities, tools, and methods for studying data from gravitational-wave detectors. + +{{< figure >}} +src = '/images/content_images/cs/gwpy-numpy-dep-graph.png' alt = 'gwpy-numpy depgraph' title = 'Dependency graph showing how GwPy package depends on NumPy' +{{< /figure >}} + +---- + +{{< figure >}} +src = '/images/content_images/cs/PyCBC-numpy-dep-graph.png' alt = 'PyCBC-numpy depgraph' title = 'Dependency graph showing how PyCBC package depends on NumPy' +{{< /figure >}} + +## Summary + +GW detection has enabled researchers to discover entirely unexpected phenomena while providing new insight into many of the most profound astrophysical phenomena known. Number crunching and data visualization is a crucial step that helps scientists gain insights into data gathered from the scientific observations and understand the results. The computations are complex and cannot be comprehended by humans unless it is visualized using computer simulations that are fed with the real observed data and analysis. NumPy along with other Python packages such as matplotlib, pandas, and scikit-learn is [enabling researchers](https://www.gw-openscience.org/events/GW150914/) to answer complex questions and discover new horizons in our understanding of the universe. + +{{< figure >}} +src = '/images/content_images/cs/numpy_gw_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} diff --git a/content/ru/citing-numpy.md b/content/ru/citing-numpy.md new file mode 100644 index 0000000000..5bb5d791b4 --- /dev/null +++ b/content/ru/citing-numpy.md @@ -0,0 +1,35 @@ +--- +title: Citing NumPy +sidebar: false +--- + +If NumPy has been significant in your research, and you would like to acknowledge the project in your academic publication, we suggest citing the following paper: + +* Harris, C.R., Millman, K.J., van der Walt, S.J. et al. _Array programming with NumPy_. Nature 585, 357–362 (2020). DOI: [10.1038/s41586-020-2649-2](https://doi.org/10.1038/s41586-020-2649-2). ([Publisher link](https://www.nature.com/articles/s41586-020-2649-2)). + +_In BibTeX format:_ + + ``` +@Article{ harris2020array, + title = {Array programming with {NumPy}}, + author = {Charles R. Harris and K. Jarrod Millman and St{\'{e}}fan J. + van der Walt and Ralf Gommers and Pauli Virtanen and David + Cournapeau and Eric Wieser and Julian Taylor and Sebastian + Berg and Nathaniel J. Smith and Robert Kern and Matti Picus + and Stephan Hoyer and Marten H. van Kerkwijk and Matthew + Brett and Allan Haldane and Jaime Fern{\'{a}}ndez del + R{\'{i}}o and Mark Wiebe and Pearu Peterson and Pierre + G{\'{e}}rard-Marchant and Kevin Sheppard and Tyler Reddy and + Warren Weckesser and Hameer Abbasi and Christoph Gohlke and + Travis E. Oliphant}, + year = {2020}, + month = sep, + journal = {Nature}, + volume = {585}, + number = {7825}, + pages = {357--362}, + doi = {10.1038/s41586-020-2649-2}, + publisher = {Springer Science and Business Media {LLC}}, + url = {https://doi.org/10.1038/s41586-020-2649-2} +} +``` diff --git a/content/ru/code-of-conduct.md b/content/ru/code-of-conduct.md new file mode 100644 index 0000000000..5ee1f4bcbe --- /dev/null +++ b/content/ru/code-of-conduct.md @@ -0,0 +1,83 @@ +--- +title: NumPy Code of Conduct +sidebar: false +aliases: + - /conduct.html +--- + +### Introduction + +This Code of Conduct applies to all spaces managed by the NumPy project, including all public and private mailing lists, issue trackers, wikis, blogs, Twitter, and any other communication channel used by our community. The NumPy project does not organise in-person events, however events related to our community should have a code of conduct similar in spirit to this one. + +This Code of Conduct should be honored by everyone who participates in the NumPy community formally or informally, or claims any affiliation with the project, in any project-related activities and especially when representing the project, in any role. + +This code is not exhaustive or complete. It serves to distill our common understanding of a collaborative, shared environment and goals. Please try to follow this code in spirit as much as in letter, to create a friendly and productive environment that enriches the surrounding community. + +### Specific Guidelines + +We strive to: + +1. Be open. We invite anyone to participate in our community. We prefer to use public methods of communication for project-related messages, unless discussing something sensitive. This applies to messages for help or project-related support, too; not only is a public support request much more likely to result in an answer to a question, it also ensures that any inadvertent mistakes in answering are more easily detected and corrected. +2. Be empathetic, welcoming, friendly, and patient. We work together to resolve conflict, and assume good intentions. We may all experience some frustration from time to time, but we do not allow frustration to turn into a personal attack. A community where people feel uncomfortable or threatened is not a productive one. +3. Be collaborative. Our work will be used by other people, and in turn we will depend on the work of others. When we make something for the benefit of the project, we are willing to explain to others how it works, so that they can build on the work to make it even better. Any decision we make will affect users and colleagues, and we take those consequences seriously when making decisions. +4. Be inquisitive. Nobody knows everything! Asking questions early avoids many problems later, so we encourage questions, although we may direct them to the appropriate forum. We will try hard to be responsive and helpful. +5. Be careful in the words that we choose. We are careful and respectful in our communication, and we take responsibility for our own speech. Be kind to others. Do not insult or put down other participants. We will not accept harassment or other exclusionary behaviour, such as: + * Violent threats or language directed against another person. + * Sexist, racist, or otherwise discriminatory jokes and language. + * Posting sexually explicit or violent material. + * Posting (or threatening to post) other people’s personally identifying information (“doxing”). + * Sharing private content, such as emails sent privately or non-publicly, or unlogged forums such as IRC channel history, without the sender’s consent. + * Personal insults, especially those using racist or sexist terms. + * Unwelcome sexual attention. + * Excessive profanity. Please avoid swearwords; people differ greatly in their sensitivity to swearing. + * Repeated harassment of others. In general, if someone asks you to stop, then stop. + * Advocating for, or encouraging, any of the above behaviour. + +### Diversity Statement + +The NumPy project welcomes and encourages participation by everyone. We are committed to being a community that everyone enjoys being part of. Although we may not always be able to accommodate each individual’s preferences, we try our best to treat everyone kindly. + +No matter how you identify yourself or how others perceive you: we welcome you. Though no list can hope to be comprehensive, we explicitly honour diversity in: age, culture, ethnicity, genotype, gender identity or expression, language, national origin, neurotype, phenotype, political beliefs, profession, race, religion, sexual orientation, socioeconomic status, subculture and technical ability, to the extent that these do not conflict with this code of conduct. + +Though we welcome people fluent in all languages, NumPy development is conducted in English. + +Standards for behaviour in the NumPy community are detailed in the Code of Conduct above. Participants in our community should uphold these standards in all their interactions and help others to do so as well (see next section). + +### Reporting Guidelines + +We know that it is painfully common for internet communication to start at or devolve into obvious and flagrant abuse. We also recognize that sometimes people may have a bad day, or be unaware of some of the guidelines in this Code of Conduct. Please keep this in mind when deciding on how to respond to a breach of this Code. + +For clearly intentional breaches, report those to the Code of Conduct Committee (see below). For possibly unintentional breaches, you may reply to the person and point out this code of conduct (either in public or in private, whatever is most appropriate). If you would prefer not to do that, please feel free to report to the Code of Conduct Committee directly, or ask the Committee for advice, in confidence. + +You can report issues to the NumPy Code of Conduct Committee at numpy-conduct@googlegroups.com. + +Currently, the Committee consists of: + +* Stefan van der Walt +* Melissa Weber Mendonça +* Rohit Goswami + +If your report involves any members of the Committee, or if they feel they have a conflict of interest in handling it, then they will recuse themselves from considering your report. Alternatively, if for any reason you feel uncomfortable making a report to the Committee, then you can also contact senior NumFOCUS staff at [conduct@numfocus.org](https://numfocus.org/code-of-conduct#persons-responsible). + +### Incident reporting resolution & Code of Conduct enforcement + +_This section summarizes the most important points, more details can be found in_ [NumPy Code of Conduct - How to follow up on a report](report-handling-manual). + +We will investigate and respond to all complaints. The NumPy Code of Conduct Committee and the NumPy Steering Committee (if involved) will protect the identity of the reporter, and treat the content of complaints as confidential (unless the reporter agrees otherwise). + +In case of severe and obvious breaches, e.g. personal threat or violent, sexist or racist language, we will immediately disconnect the originator from NumPy communication channels; please see the manual for details. + +In cases not involving clear severe and obvious breaches of this Code of Conduct the process for acting on any received Code of Conduct violation report will be: + +1. acknowledge report is received, +2. reasonable discussion/feedback, +3. mediation (if feedback didn’t help, and only if both reporter and reportee agree to this), +4. enforcement via transparent decision (see [Resolutions](report-handling-manual/#resolutions)) by the Code of Conduct Committee. + +The Committee will respond to any report as soon as possible, and at most within 72 hours. + +### Endnotes + +We are thankful to the groups behind the following documents, from which we drew content and inspiration: + +- [The SciPy Code of Conduct](https://docs.scipy.org/doc/scipy/dev/conduct/code_of_conduct.html) diff --git a/content/ru/community.md b/content/ru/community.md new file mode 100644 index 0000000000..5034fba239 --- /dev/null +++ b/content/ru/community.md @@ -0,0 +1,66 @@ +--- +title: Community +sidebar: false +--- + +NumPy is a community-driven open source project developed by a diverse group of [contributors](/teams/). The NumPy leadership has made a strong commitment to creating an open, inclusive, and positive community. Please read the [NumPy Code of Conduct](/code-of-conduct) for guidance on how to interact with others in a way that makes the community thrive. + +We offer several communication channels to learn, share your knowledge and connect with others within the NumPy community. + + +## Participate online + +The following are ways to engage directly with the NumPy project and community. _Please note that we encourage users and community members to support each other for usage questions - see [Get Help](/gethelp)._ + + +### [NumPy mailing list](https://mail.python.org/mailman/listinfo/numpy-discussion) + +This list is the main forum for longer-form discussions, like adding new features to NumPy, making changes to the NumPy Roadmap, and all kinds of project-wide decision making. Announcements about NumPy, such as for releases, developer meetings, sprints or conference talks are also made on this list. + +On this list please use bottom posting, reply to the list (rather than to another sender), and don't reply to digests. A searchable archive of this list is available [here](https://mail.python.org/archives/list/numpy-discussion@python.org/). + +*** + +### [GitHub issue tracker](https://github.com/numpy/numpy/issues) + +- For bug reports (e.g. "`np.arange(3).shape` returns `(5,)`, when it should return `(3,)`"); +- documentation issues (e.g. "I found this section unclear"); +- and feature requests (e.g. "I would like to have a new interpolation method in `np.percentile`"). + +_Please note that GitHub is not the right place to report a security vulnerability. If you think you have found a security vulnerability in NumPy, please report it [here](https://tidelift.com/docs/security)._ + +*** + +### [Slack](https://numpy-team.slack.com) + +A real-time chat room to ask questions about _contributing_ to NumPy. This is a private space, specifically meant for people who are hesitant to bring up their questions or ideas on a large public mailing list or GitHub. Please see [here](https://numpy.org/devdocs/dev/index.html#contributing-to-numpy) for more details and how to get an invite. + + +## Study Groups and Meetups + +If you would like to find a local meetup or study group to learn more about NumPy and the wider ecosystem of Python packages for data science and scientific computing, we recommend exploring the [PyData meetups](https://www.meetup.com/pro/pydata/) (150+ meetups, 100,000+ members). + +NumPy also organizes in-person sprints for its team and interested contributors occasionally. These are typically planned several months in advance and will be announced on the [mailing list](https://mail.python.org/mailman/listinfo/numpy-discussion) and [Twitter](https://twitter.com/numpy_team). + + +## Conferences + +The NumPy project doesn't organize its own conferences. The conferences that have traditionally been most popular with NumPy maintainers, contributors and users are the SciPy and PyData conference series: + +- [SciPy US](https://conference.scipy.org) +- [EuroSciPy](https://www.euroscipy.org) +- [SciPy Latin America](https://www.scipyla.org) +- [SciPy India](https://scipy.in) +- [SciPy Japan](https://conference.scipy.org) +- [PyData conferences](https://pydata.org/event-schedule/) (15-20 events a year spread over many countries) + +Many of these conferences include tutorial days that cover NumPy and/or sprints where you can learn how to contribute to NumPy or related open source projects. + + +## Join the NumPy community + +To thrive, the NumPy project needs your expertise and enthusiasm. Not a coder? Not a problem! There are many ways to contribute to NumPy. + +If you are interested in becoming a NumPy contributor (yay!) we recommend checking out our [Contribute](/contribute) page. + +Also, feel free to stop by and say hi at one of our community meetings. To keep track of them, check out our events calendar [here](https://scientific-python.org/calendars/). diff --git a/content/ru/config.yaml b/content/ru/config.yaml new file mode 100644 index 0000000000..4aaf75b2e6 --- /dev/null +++ b/content/ru/config.yaml @@ -0,0 +1,140 @@ +languageName: English +params: + description: Why NumPy? Powerful n-dimensional arrays. Numerical computing tools. Interoperable. Performant. Open source. + navbarlogo: + image: logo.svg + text: NumPy + link: / + hero: + #Main hero title + title: NumPy + #Hero subtitle (optional) + subtitle: The fundamental package for scientific computing with Python + #Button text + buttontext: "Latest release: NumPy 1.26. View all releases" + #Where the main hero button links to + buttonlink: "/news/#releases" + #Hero image (from static/images/___) + image: logo.svg + shell: + title: placeholder + intro: + - + title: Try NumPy + text: Use the interactive shell to try NumPy in the browser + docslink: Don't forget to check out the docs. + casestudies: + title: CASE STUDIES + features: + - + title: First Image of a Black Hole + text: How NumPy, together with libraries like SciPy and Matplotlib that depend on NumPy, enabled the Event Horizon Telescope to produce the first ever image of a black hole + img: /images/content_images/case_studies/blackhole.png + alttext: First image of a black hole. It is an orange circle in a black background. + url: /case-studies/blackhole-image + - + title: Detection of Gravitational Waves + text: In 1916, Albert Einstein predicted gravitational waves; 100 years later their existence was confirmed by LIGO scientists using NumPy. + img: /images/content_images/case_studies/gravitional.png + alttext: Two orbs orbiting each other. They are displacing gravity around them. + url: /case-studies/gw-discov + - + title: Sports Analytics + text: Cricket Analytics is changing the game by improving player and team performance through statistical modelling and predictive analytics. NumPy enables many of these analyses. + img: /images/content_images/case_studies/sports.jpg + alttext: Cricket ball on green field. + url: /case-studies/cricket-analytics + - + title: Pose Estimation using deep learning + text: DeepLabCut uses NumPy for accelerating scientific studies that involve observing animal behavior for better understanding of motor control, across species and timescales. + img: /images/content_images/case_studies/deeplabcut.png + alttext: Cheetah pose analysis + url: /case-studies/deeplabcut-dnn + tabs: + title: ECOSYSTEM + section5: false + navbar: + - + title: Install + url: /install + - + title: Documentation + url: https://numpy.org/doc/stable + - + title: Learn + url: /learn + - + title: Community + url: /community + - + title: About Us + url: /about + - + title: News + url: /news + - + title: Contribute + url: /contribute + footer: + logo: logo.svg + socialmediatitle: "" + socialmedia: + - + link: https://github.com/numpy/numpy + icon: github + - + link: https://www.youtube.com/channel/UCguIL9NZ7ybWK5WQ53qbHng + icon: youtube + - + link: https://twitter.com/numpy_team + icon: twitter + quicklinks: + column1: + title: "" + links: + - + text: Install + link: /install + - + text: Documentation + link: https://numpy.org/doc/stable + - + text: Learn + link: /learn + - + text: Citing Numpy + link: /citing-numpy + - + text: Roadmap + link: https://numpy.org/neps/roadmap.html + column2: + links: + - + text: About us + link: /about + - + text: Community + link: /community + - + text: User surveys + link: /user-surveys + - + text: Contribute + link: /contribute + - + text: Code of conduct + link: /code-of-conduct + column3: + links: + - + text: Get help + link: /gethelp + - + text: Terms of use + link: /terms + - + text: Privacy + link: /privacy + - + text: Press kit + link: /press-kit diff --git a/content/ru/contribute.md b/content/ru/contribute.md new file mode 100644 index 0000000000..6efff53624 --- /dev/null +++ b/content/ru/contribute.md @@ -0,0 +1,66 @@ +--- +title: Contribute to NumPy +sidebar: false +--- + +The NumPy project welcomes your expertise and enthusiasm! Your choices aren't limited to programming, as you can see below there are many areas where we need **your** help. + +If you're unsure where to start or how your skills fit in, _reach out!_ You can ask on the [mailing list](https://mail.python.org/mailman/listinfo/numpy-discussion) or [GitHub](http://github.com/numpy/numpy) (open an [issue](https://github.com/numpy/numpy/issues) or comment on a relevant issue). + +Those are our preferred channels (open source is open by nature), but if you prefer to talk privately, contact our community coordinators at or on [Slack](https://numpy-team.slack.com) (write for an invite). + +We also have a biweekly _community call_, details of which are announced on the [mailing list](https://mail.python.org/mailman/listinfo/numpy-discussion). You are very welcome to join. If you are new to contributing to open source, we also highly recommend reading [this guide](https://opensource.guide/how-to-contribute/). + +Our community aspires to treat everyone equally and to value all contributions. We have a [Code of Conduct](/code-of-conduct) to foster an open and welcoming environment. + +### Writing code + +Programmers, this [guide](https://numpy.org/devdocs/dev/index.html#development-process-summary) explains how to contribute to the NumPy codebase.
Check out also our [YouTube channel](https://www.youtube.com/playlist?list=PLCK6zCrcN3GXBUUzDr9L4__LnXZVtaIzS) for additional advice. + + +### Reviewing pull requests +The project has more than 250 open pull requests -- meaning many potential improvements and many open-source contributors waiting for feedback. If you're a developer who knows NumPy, you can help even if you're not familiar with the codebase. You can: +* summarize a long-running discussion +* triage documentation PRs +* test proposed changes + + +### Developing educational materials + +NumPy's [User Guide](https://numpy.org/devdocs) is undergoing rehabilitation. We're in need of new tutorials, how-to's, and deep-dive explanations, and the site needs restructuring. Opportunities aren't limited to writers. We'd also welcome worked examples, notebooks, and videos. [NEP 44 — Restructuring the NumPyDocumentation](https://numpy.org/neps/nep-0044-restructuring-numpy-docs.html) lays out our ideas -- and you may have others. + + +### Issue triaging + +The [NumPy issue tracker](https://github.com/numpy/numpy/issues) has a _lot_ of open issues. Some are no longer valid, some should be prioritized, and some would make good issues for new contributors. You can: + +* check if older bugs are still present +* find duplicate issues and link related ones +* add good self-contained reproducers to issues +* label issues correctly (this requires triage rights -- just ask) + +Please just dive in. + + +### Website development + +We've just revamped our website, but we're far from done. If you love web development, these [issues](https://github.com/numpy/numpy.org/issues?q=is%3Aissue+is%3Aopen+label%3Adesign) list some of our unmet needs -- and feel free to share your own ideas. + + +### Graphic design + +We can barely begin to list the contributions a graphic designer can make here. Our docs are parched for illustration; our growing website craves images -- opportunities abound. + + +### Translating website content + +We plan multiple translations of [numpy.org](https://numpy.org) to make NumPy accessible to users in their native language. Volunteer translators are at the heart of this effort. See [here](https://numpy.org/neps/nep-0028-website-redesign.html#translation-multilingual-i18n) for background; comment on [this GitHub issue](https://github.com/numpy/numpy.org/issues/55) to sign up. + + +### Community coordination and outreach + +Through community contact we share our work more widely and learn where we're falling short. We're eager to get more people involved in efforts like our [Twitter](https://twitter.com/numpy_team) account, organizing NumPy [code sprints](https://scisprints.github.io/), a newsletter, and perhaps a blog. + +### Fundraising + +NumPy was all-volunteer for many years, but as its importance grew it became clear that to ensure stability and growth we'd need financial support. [This SciPy'19 talk](https://www.youtube.com/watch?v=dBTJD_FDVjU) explains how much difference that support has made. Like all the nonprofit world, we're constantly searching for grants, sponsorships, and other kinds of support. We have a number of ideas and of course we welcome more. Fundraising is a scarce skill here -- we'd appreciate your help. diff --git a/content/ru/gethelp.md b/content/ru/gethelp.md new file mode 100644 index 0000000000..da3bff67a5 --- /dev/null +++ b/content/ru/gethelp.md @@ -0,0 +1,20 @@ +--- +title: Get Help +sidebar: false +--- + +**Development issues:** For NumPy development-related matters (e.g., bug reports), please see [Community](/community). + +**User questions:** The best way to get help is to post your question to a site like [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) or [Reddit](https://www.reddit.com/r/Numpy/). We wish we could keep an eye on these sites, or answer questions directly, but the volume is a little overwhelming! + +### [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) + +A forum for asking usage questions, e.g. "How do I do X in NumPy?”. Please [use the `#numpy` tag](https://stackoverflow.com/help/tagging) + +*** + +### [Reddit](https://www.reddit.com/r/Numpy/) + +Another forum for usage questions. + +*** diff --git a/content/ru/history.md b/content/ru/history.md new file mode 100644 index 0000000000..aa669d375b --- /dev/null +++ b/content/ru/history.md @@ -0,0 +1,21 @@ +--- +title: History of NumPy +sidebar: false +--- + +NumPy is a foundational Python library that provides array data structures and related fast numerical routines. When started, the library had little funding, and was written mainly by graduate students—many of them without computer science education, and often without a blessing of their advisors. To even imagine that a small group of “rogue” student programmers could upend the already well-established ecosystem of research software—backed by millions in funding and many hundreds of highly qualified engineers — was preposterous. Yet, the philosophical motivations behind a fully open tool stack, in combination with the excited, friendly community with a singular focus, have proven auspicious in the long run. Nowadays, NumPy is relied upon by scientists, engineers, and many other professionals around the world. For example, the published scripts used in the analysis of gravitational waves import NumPy, and the M87 black hole imaging project directly cites NumPy. + +For the in-depth account on milestones in the development of NumPy and related libraries please see [arxiv.org](https://arxiv.org/abs/1907.10121). + +If you’d like to obtain a copy of the original Numeric and Numarray libraries, follow the links below: + +[Download Page for *Numeric*](https://sourceforge.net/projects/numpy/files/Old%20Numeric/)* + +[Download Page for *Numarray*](https://sourceforge.net/projects/numpy/files/Old%20Numarray/)* + +*Please note that these older array packages are no longer maintained, and users are strongly advised to use NumPy for any array-related purposes or refactor any pre-existing code to utilize the NumPy library. + +### Historic Documentation + +[Download *`Numeric'* Manual](static/numeric-manual.pdf) + diff --git a/content/ru/install.md b/content/ru/install.md new file mode 100644 index 0000000000..f64a17841d --- /dev/null +++ b/content/ru/install.md @@ -0,0 +1,126 @@ +--- +title: Installing NumPy +sidebar: false +--- + +The only prerequisite for installing NumPy is Python itself. If you don't have Python yet and want the simplest way to get started, we recommend you use the [Anaconda Distribution](https://www.anaconda.com/download) - it includes Python, NumPy, and many other commonly used packages for scientific computing and data science. + +NumPy can be installed with `conda`, with `pip`, with a package manager on macOS and Linux, or [from source](https://numpy.org/devdocs/building). For more detailed instructions, consult our [Python and NumPy installation guide](#python-numpy-install-guide) below. + +**CONDA** + +If you use `conda`, you can install NumPy from the `defaults` or `conda-forge` channels: + +```bash +# Best practice, use an environment rather than install in the base env +conda create -n my-env +conda activate my-env +# If you want to install from conda-forge +conda config --env --add channels conda-forge +# The actual install command +conda install numpy +``` + +**PIP** + +If you use `pip`, you can install NumPy with: + +```bash +pip install numpy +``` +Also when using pip, it's good practice to use a virtual environment - see [Reproducible Installs](#reproducible-installs) below for why, and [this guide](https://dev.to/bowmanjd/python-tools-for-managing-virtual-environments-3bko#howto) for details on using virtual environments. + + + + +# Python and NumPy installation guide + +Installing and managing packages in Python is complicated, there are a number of alternative solutions for most tasks. This guide tries to give the reader a sense of the best (or most popular) solutions, and give clear recommendations. It focuses on users of Python, NumPy, and the PyData (or numerical computing) stack on common operating systems and hardware. + +## Recommendations + +We'll start with recommendations based on the user's experience level and operating system of interest. If you're in between "beginning" and "advanced", please go with "beginning" if you want to keep things simple, and with "advanced" if you want to work according to best practices that go a longer way in the future. + +### Beginning users + +On all of Windows, macOS, and Linux: + +- Install [Anaconda](https://www.anaconda.com/download) (it installs all packages you need and all other tools mentioned below). +- For writing and executing code, use notebooks in [JupyterLab](https://jupyterlab.readthedocs.io/en/stable/index.html) for exploratory and interactive computing, and [Spyder](https://www.spyder-ide.org/) or [Visual Studio Code](https://code.visualstudio.com/) for writing scripts and packages. +- Use [Anaconda Navigator](https://docs.anaconda.com/anaconda/navigator/) to manage your packages and start JupyterLab, Spyder, or Visual Studio Code. + + +### Advanced users + +#### Conda + +- Install [Miniforge](https://github.com/conda-forge/miniforge). +- Keep the `base` conda environment minimal, and use one or more [conda environments](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html) to install the package you need for the task or project you're working on. + +#### Alternative if you prefer pip/PyPI + +For users who know, from personal preference or reading about the main differences between conda and pip below, they prefer a pip/PyPI-based solution, we recommend: +- Install Python from [python.org](https://www.python.org/downloads/), [Homebrew](https://brew.sh/), or your Linux package manager. +- Use [Poetry](https://python-poetry.org/) as the most well-maintained tool that provides a dependency resolver and environment management capabilities in a similar fashion as conda does. + + +## Python package management + +Managing packages is a challenging problem, and, as a result, there are lots of tools. For web and general purpose Python development there's a whole [host of tools](https://packaging.python.org/guides/tool-recommendations/) complementary with pip. For high-performance computing (HPC), [Spack](https://github.com/spack/spack) is worth considering. For most NumPy users though, [conda](https://conda.io/en/latest/) and [pip](https://pip.pypa.io/en/stable/) are the two most popular tools. + + +### Pip & conda + +The two main tools that install Python packages are `pip` and `conda`. Their functionality partially overlaps (e.g. both can install `numpy`), however, they can also work together. We'll discuss the major differences between pip and conda here - this is important to understand if you want to manage packages effectively. + +The first difference is that conda is cross-language and it can install Python, while pip is installed for a particular Python on your system and installs other packages to that same Python install only. This also means conda can install non-Python libraries and tools you may need (e.g. compilers, CUDA, HDF5), while pip can't. + +The second difference is that pip installs from the Python Packaging Index (PyPI), while conda installs from its own channels (typically "defaults" or "conda-forge"). PyPI is the largest collection of packages by far, however, all popular packages are available for conda as well. + +The third difference is that conda is an integrated solution for managing packages, dependencies and environments, while with pip you may need another tool (there are many!) for dealing with environments or complex dependencies. + + + +### Reproducible installs + +As libraries get updated, results from running your code can change, or your code can break completely. It's important to be able to reconstruct the set of packages and versions you're using. Best practice is to: + +1. use a different environment per project you're working on, +2. record package names and versions using your package installer; each has its own metadata format for this: + - Conda: [conda environments and environment.yml](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html) + - Pip: [virtual environments](https://docs.python.org/3/tutorial/venv.html) and [requirements.txt](https://pip.readthedocs.io/en/latest/user_guide/#requirements-files) + - Poetry: [virtual environments and pyproject.toml](https://python-poetry.org/docs/basic-usage/) + + + +## NumPy packages & accelerated linear algebra libraries + +NumPy doesn't depend on any other Python packages, however, it does depend on an accelerated linear algebra library - typically [Intel MKL](https://software.intel.com/en-us/mkl) or [OpenBLAS](https://www.openblas.net/). Users don't have to worry about installing those (they're automatically included in all NumPy install methods). Power users may still want to know the details, because the used BLAS can affect performance, behavior and size on disk: + +- The NumPy wheels on PyPI, which is what pip installs, are built with OpenBLAS. The OpenBLAS libraries are included in the wheel. This makes the wheel larger, and if a user installs (for example) SciPy as well, they will now have two copies of OpenBLAS on disk. + +- In the conda defaults channel, NumPy is built against Intel MKL. MKL is a separate package that will be installed in the users' environment when they install NumPy. + +- In the conda-forge channel, NumPy is built against a dummy "BLAS" package. When a user installs NumPy from conda-forge, that BLAS package then gets installed together with the actual library - this defaults to OpenBLAS, but it can also be MKL (from the defaults channel), or even [BLIS](https://github.com/flame/blis) or reference BLAS. + +- The MKL package is a lot larger than OpenBLAS, it's about 700 MB on disk while OpenBLAS is about 30 MB. + +- MKL is typically a little faster and more robust than OpenBLAS. + +Besides install sizes, performance and robustness, there are two more things to consider: + +- Intel MKL is not open source. For normal use this is not a problem, but if a user needs to redistribute an application built with NumPy, this could be an issue. +- Both MKL and OpenBLAS will use multi-threading for function calls like `np.dot`, with the number of threads being determined by both a build-time option and an environment variable. Often all CPU cores will be used. This is sometimes unexpected for users; NumPy itself doesn't auto-parallelize any function calls. It typically yields better performance, but can also be harmful - for example when using another level of parallelization with Dask, scikit-learn or multiprocessing. + + +## Troubleshooting + +If your installation fails with the message below, see [Troubleshooting ImportError](https://numpy.org/doc/stable/user/troubleshooting-importerror.html). + +``` +IMPORTANT: PLEASE READ THIS FOR ADVICE ON HOW TO SOLVE THIS ISSUE! + +Importing the numpy c-extensions failed. This error can happen for +different reasons, often due to issues with your setup. +``` + diff --git a/content/ru/learn.md b/content/ru/learn.md new file mode 100644 index 0000000000..4f9fa53ae3 --- /dev/null +++ b/content/ru/learn.md @@ -0,0 +1,76 @@ +--- +title: Learn +sidebar: false +--- + +For the **official NumPy documentation** visit [numpy.org/doc/stable](https://numpy.org/doc/stable). + +*** + +Below is a curated collection of educational resources, both for self-learning and teaching others, developed by NumPy contributors and vetted by the community. + +## Beginners + +There's a ton of information about NumPy out there. If you are just starting, we'd strongly recommend the following: + + **Tutorials** + +* [NumPy Quickstart Tutorial](https://numpy.org/devdocs/user/quickstart.html) +* [NumPy Tutorials](https://numpy.org/numpy-tutorials) A collection of tutorials and educational materials in the format of Jupyter Notebooks developed and maintained by the NumPy Documentation team. To submit your own content, visit the [numpy-tutorials repository on GitHub](https://github.com/numpy/numpy-tutorials). +* [NumPy Illustrated: The Visual Guide to NumPy *by Lev Maximov*](https://betterprogramming.pub/3b1d4976de1d?sk=57b908a77aa44075a49293fa1631dd9b) +* [Scientific Python Lectures](https://lectures.scientific-python.org/) Besides covering NumPy, these lectures offer a broader introduction to the scientific Python ecosystem. +* [NumPy: the absolute basics for beginners](https://numpy.org/devdocs/user/absolute_beginners.html) +* [NumPy tutorial *by Nicolas Rougier*](https://github.com/rougier/numpy-tutorial) +* [Stanford CS231 *by Justin Johnson*](http://cs231n.github.io/python-numpy-tutorial/) +* [NumPy User Guide](https://numpy.org/devdocs) + + **Books** + +* [Guide to NumPy *by Travis E. Oliphant*](https://web.mit.edu/dvp/Public/numpybook.pdf) This is a free version 1 from 2006. For the latest copy (2015) see [here](https://www.barnesandnoble.com/w/guide-to-numpy-travis-e-oliphant-phd/1144670472). +* [From Python to NumPy *by Nicolas P. Rougier*](https://www.labri.fr/perso/nrougier/from-python-to-numpy/) +* [Elegant SciPy](https://www.amazon.com/Elegant-SciPy-Art-Scientific-Python/dp/1491922877) *by Juan Nunez-Iglesias, Stefan van der Walt, and Harriet Dashnow* + +You may also want to check out the [Goodreads list](https://www.goodreads.com/shelf/show/python-scipy) on the subject of "Python+SciPy." Most books there are about the "SciPy ecosystem," which has NumPy at its core. + + **Videos** + +* [Introduction to Numerical Computing with NumPy](http://youtu.be/ZB7BZMhfPgk) *by Alex Chabot-Leclerc* + +*** + +## Advanced + +Try these advanced resources for a better understanding of NumPy concepts like advanced indexing, splitting, stacking, linear algebra, and more. + + **Tutorials** + +* [100 NumPy Exercises](http://www.labri.fr/perso/nrougier/teaching/numpy.100/index.html) *by Nicolas P. Rougier* +* [An Introduction to NumPy and Scipy](https://engineering.ucsb.edu/~shell/che210d/numpy.pdf) *by M. Scott Shell* +* [Numpy Medkits](http://mentat.za.net/numpy/numpy_advanced_slides/) *by Stéfan van der Walt* +* [NumPy Tutorials](https://numpy.org/numpy-tutorials) A collection of tutorials and educational materials in the format of Jupyter Notebooks developed and maintained by the NumPy Documentation team. To submit your own content, visit the [numpy-tutorials repository on GitHub](https://github.com/numpy/numpy-tutorials). + + **Books** + +* [Python Data Science Handbook](https://www.amazon.com/Python-Data-Science-Handbook-Essential/dp/1098121228) *by Jake Vanderplas* +* [Python for Data Analysis](https://www.amazon.com/Python-Data-Analysis-Wrangling-Jupyter/dp/109810403X) *by Wes McKinney* +* [Numerical Python: Scientific Computing and Data Science Applications with Numpy, SciPy, and Matplotlib](https://www.amazon.com/Numerical-Python-Scientific-Applications-Matplotlib/dp/1484242459) *by Robert Johansson* + + **Videos** + +* [Advanced NumPy - broadcasting rules, strides, and advanced indexing](https://www.youtube.com/watch?v=cYugp9IN1-Q) *by Juan Nunez-Iglesias* + +*** + +## NumPy Talks + +* [The Future of NumPy Indexing](https://www.youtube.com/watch?v=o0EacbIbf58) *by Jaime Fernández* (2016) +* [Evolution of Array Computing in Python](https://www.youtube.com/watch?v=HVLPJnvInzM&t=10s) *by Ralf Gommers* (2019) +* [NumPy: what has changed and what is going to change?](https://www.youtube.com/watch?v=YFLVQFjRmPY) *by Matti Picus* (2019) +* [Inside NumPy](https://www.youtube.com/watch?v=dBTJD_FDVjU) *by Ralf Gommers, Sebastian Berg, Matti Picus, Tyler Reddy, Stefan van der Walt, Charles Harris* (2019) +* [Brief Review of Array Computing in Python](https://www.youtube.com/watch?v=f176j2g2eNc) *by Travis Oliphant* (2019) + +*** + +## Citing NumPy + +If NumPy has been significant in your research, and you would like to acknowledge the project in your academic publication, please see [this citation information](/citing-numpy). diff --git a/content/ru/news.md b/content/ru/news.md new file mode 100644 index 0000000000..793619c0d1 --- /dev/null +++ b/content/ru/news.md @@ -0,0 +1,279 @@ +--- +title: News +sidebar: false +newsHeader: "NumPy 2.0 release date: June 16" +date: 2024-05-23 +--- + +### NumPy 2.0 release date: June 16 + +_23 May, 2024_ -- We are excited to announce that NumPy 2.0 is planned to be released on June 16, 2024. This release has been over a year in the making, and is the first major release since 2006. Importantly, in addition to many new features and performance improvement, it contains **breaking changes** to the ABI as well as the Python and C APIs. It is likely that downstream packages and end user code needs to be adapted - if you can, please verify whether your code works with NumPy `2.0.0rc2`. **Please see the following for more details:** + +- The [NumPy 2.0 migration guide](https://numpy.org/devdocs/numpy_2_0_migration_guide.html) +- The [2.0.0 release notes](https://numpy.org/devdocs/release/2.0.0-notes.html) +- Announcement issue for status updates: [numpy#24300](https://github.com/numpy/numpy/issues/24300) + + +### NumFOCUS end of the year fundraiser +_Dec 19, 2023_ -- NumFOCUS has teamed up with PyCharm during their EOY campaign to offer a 30% discount on first-time PyCharm licenses. All year-one revenue from PyCharm purchases from now until December 23rd, 2023 will go directly to the NumFOCUS programs. + +Use unique URL that will allow to track purchases https://lp.jetbrains.com/support-data-science/ or a coupon code ISUPPORTDATASCIENCE  + +### NumPy 1.26.0 released + +_Sep 16, 2023_ -- [NumPy 1.26.0](https://numpy.org/doc/stable/release/1.26.0-notes.html) is now available. The highlights of the release are: + +* Python 3.12.0 support. +* Cython 3.0.0 compatibility. +* Use of the Meson build system +* Updated SIMD support +* f2py fixes, meson and bind(x) support +* Support for the updated Accelerate BLAS/LAPACK library + +The NumPy 1.26.0 release is a continuation of the 1.25.x series that marks the transition to the Meson build system and provision of support for Cython 3.0.0. A total of 20 people contributed to this release and 59 pull requests were merged. + +The Python versions supported by this release are 3.9-3.12. + +### numpy.org is now available in Japanese and Portuguese + +_Aug 2, 2023_ -- numpy.org is now available in 2 additional languages: Japanese and Portuguese. This wouldn’t be possible without our dedicated volunteers: + +_Portuguese:_ +* Melissa Weber Mendonça (melissawm) +* Ricardo Prins (ricardoprins) +* Getúlio Silva (getuliosilva) +* Julio Batista Silva (jbsilva) +* Alexandre de Siqueira (alexdesiqueira) +* Alexandre B A Villares (villares) +* Vini Salazar (vinisalazar) + +_Japanese:_ +* Atsushi Sakai (AtsushiSakai) +* KKunai +* Tom Kelly (TomKellyGenetics) +* Yuji Kanagawa (kngwyu) +* Tetsuo Koyama (tkoyama010) + +The work on the translation infrastructure is supported with funding from CZI. + +Looking ahead, we’d love to translate the website into more languages. If you’d like to help, please connect with the NumPy Translations Team on Slack: https://join.slack.com/t/numpy-team/shared_invite/zt-1gokbq56s-bvEpo10Ef7aHbVtVFeZv2w. (Look for the #translations channel.) We are also building a Translations Team who will be working on localizing documentation and educational content across the Scientific Python ecosystem. If this piqued your interest, join us on the Scientific Python Discord: https://discord.gg/khWtqY6RKr. (Look for the #translation channel.) + +### NumPy 1.25.0 released + +_Jun 17, 2023_ -- [NumPy 1.25.0](https://numpy.org/doc/stable/release/1.25.0-notes.html) is now available. The highlights of the release are: + +* Support for MUSL, there are now MUSL wheels. +* Support for the Fujitsu C/C++ compiler. +* Object arrays are now supported in einsum. +* Support for the inplace matrix multiplication (`@=`). + +The NumPy 1.25.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase the execution speed, and clarify the documentation. There has also been preparatory work for the future NumPy 2.0.0, resulting in a large number of new and expired deprecations. + +A total of 148 people contributed to this release and 530 pull requests were merged. + +The Python versions supported by this release are 3.9-3.11. + +### Fostering an Inclusive Culture: Call for Participation + +_May 10, 2023_ -- Fostering an Inclusive Culture: Call for Participation + +How can we be better when it comes to diversity and inclusion? Read the report and find out how to get involved [here](https://contributor-experience.org/docs/posts/dei-report/). + +### NumPy documentation team leadership transition + +_Jan 6, 2023_ –- Mukulika Pahari and Ross Barnowski are appointed as the new NumPy documentation team leads replacing Melissa Mendonça. We thank Melissa for all her contributions to the NumPy official documentation and educational materials, and Mukulika and Ross for stepping up. + +### NumPy 1.24.0 released + +_Dec 18, 2022_ -- [NumPy 1.24.0](https://numpy.org/doc/stable/release/1.24.0-notes.html) is now available. The highlights of the release are: + +* New "dtype" and "casting" keywords for stacking functions. +* New F2PY features and fixes. +* Many new deprecations, check them out. +* Many expired deprecations, + +The NumPy 1.24.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase execution speed, and clarify the documentation. There are a large number of new and expired deprecations due to changes in dtype promotion and cleanups. It is the work of 177 contributors spread over 444 pull requests. The supported Python versions are 3.8-3.11. + +### Numpy 1.23.0 released + +_Jun 22, 2022_ -- [NumPy 1.23.0](https://numpy.org/doc/stable/release/1.23.0-notes.html) is now available. The highlights of the release are: + +* Implementation of `loadtxt` in C, greatly improving its performance. +* Exposure of DLPack at the Python level for easy data exchange. +* Changes to the promotion and comparisons of structured dtypes. +* Improvements to f2py. + +The NumPy 1.23.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase the execution speed, clarify the documentation, and expire old deprecations. It is the work of 151 contributors spread over 494 pull requests. The Python versions supported by this release 3.8-3.10. Python 3.11 will be supported when it reaches the rc stage. + +### NumFOCUS DEI research study: call for participation + +_Apr 13, 2022_ -- NumPy is working with [NumFOCUS](http://numfocus.org/) on a [research project](https://numfocus.org/diversity-inclusion-disc/a-pivotal-time-in-numfocuss-project-aimed-dei-efforts?eType=EmailBlastContent&eId=f41a86c3-60d4-4cf9-86cf-58eb49dc968c) funded by the [Gordon & Betty Moore Foundation](https://www.moore.org/) to understand the barriers to participation that contributors, particularly those from historically underrepresented groups, face in the open-source software community. The research team would like to talk to new contributors, project developers and maintainers, and those who have contributed in the past about their experiences joining and contributing to NumPy. + +**Interested in sharing your experiences?** + +Please complete this brief [“Participant Interest” form](https://numfocus.typeform.com/to/WBWVJSqe) which contains additional information on the research goals, privacy, and confidentiality considerations. Your participation will be valuable to the growth and sustainability of diverse and inclusive open-source software communities. Accepted participants will participate in a 30-minute interview with a research team member. + +### Numpy 1.22.0 release + +_Dec 31, 2021_ -- [NumPy 1.22.0](https://numpy.org/doc/stable/release/1.22.0-notes.html) is now available. The highlights of the release are: + +* Type annotations of the main namespace are essentially complete. Upstream is a moving target, so there will likely be further improvements, but the major work is done. This is probably the most user visible enhancement in this release. +* A preliminary version of the proposed [array API Standard](https://data-apis.org/array-api/latest/) is provided (see [NEP 47](https://numpy.org/neps/nep-0047-array-api-standard.html)). This is a step in creating a standard collection of functions that can be used across libraries such as CuPy and JAX. +* NumPy now has a DLPack backend. DLPack provides a common interchange format for array (tensor) data. +* New methods for `quantile`, `percentile`, and related functions. The new methods provide a complete set of the methods commonly found in the literature. +* The universal functions have been refactored to implement most of [NEP 43](https://numpy.org/neps/nep-0043-extensible-ufuncs.html). This also unlocks the ability to experiment with the future DType API. +* A new configurable memory allocator for use by downstream projects. + +NumPy 1.22.0 is a big release featuring the work of 153 contributors spread over 609 pull requests. The Python versions supported by this release are 3.8-3.10. + +### Advancing an inclusive culture in the scientific Python ecosystem + +_August 31, 2021_ -- We are happy to announce the Chan Zuckerberg Initiative has [awarded a grant](https://chanzuckerberg.com/newsroom/czi-awards-16-million-for-foundational-open-source-software-tools-essential-to-biomedicine/) to support the onboarding, inclusion, and retention of people from historically marginalized groups on scientific Python projects, and to structurally improve the community dynamics for NumPy, SciPy, Matplotlib, and Pandas. + +As a part of [CZI's Essential Open Source Software for Science program](https://chanzuckerberg.com/eoss/), this [Diversity & Inclusion supplemental grant](https://cziscience.medium.com/advancing-diversity-and-inclusion-in-scientific-open-source-eaabe6a5488b) will support the creation of dedicated Contributor Experience Lead positions to identify, document, and implement practices to foster inclusive open-source communities. This project will be led by Melissa Mendonça (NumPy), with additional mentorship and guidance provided by Ralf Gommers (NumPy, SciPy), Hannah Aizenman and Thomas Caswell (Matplotlib), Matt Haberland (SciPy), and Joris Van den Bossche (Pandas). + +This is an ambitious project aiming to discover and implement activities that should structurally improve the community dynamics of our projects. By establishing these new cross-project roles, we hope to introduce a new collaboration model to the Scientific Python communities, allowing community-building work within the ecosystem to be done more efficiently and with greater outcomes. We also expect to develop a clearer picture of what works and what doesn't in our projects to engage and retain new contributors, especially from historically underrepresented groups. Finally, we plan on producing detailed reports on the actions executed, explaining how they have impacted our projects in terms of representation and interaction with our communities. + +The two-year project is expected to start by November 2021, and we are excited to see the results from this work! [You can read the full proposal here](https://figshare.com/articles/online_resource/Advancing_an_inclusive_culture_in_the_scientific_Python_ecosystem/16548063). + +### 2021 NumPy survey + +_July 12, 2021_ -- At NumPy, we believe in the power of our community. 1,236 NumPy users from 75 countries participated in our inaugural survey last year. The survey findings gave us a very good understanding of what we should focus on for the next 12 months. + +It’s time for another survey, and we are counting on you once again. It will take about 15 minutes of your time. Besides English, the survey questionnaire is available in 8 additional languages: Bangla, French, Hindi, Japanese, Mandarin, Portuguese, Russian, and Spanish. + +Follow the link to get started: https://berkeley.qualtrics.com/jfe/form/SV_aaOONjgcBXDSl4q. + + +### Numpy 1.21.0 release + +_Jun 23, 2021_ -- [NumPy 1.21.0](https://numpy.org/doc/stable/release/1.21.0-notes.html) is now available. The highlights of the release are: + +- continued SIMD work covering more functions and platforms, +- initial work on the new dtype infrastructure and casting, +- universal2 wheels for Python 3.8 and Python 3.9 on Mac, +- improved documentation, +- improved annotations, +- new `PCG64DXSM` bitgenerator for random numbers. + +This NumPy release is the result of 581 merged pull requests contributed by 175 people. The Python versions supported for this release are 3.7-3.9, support for Python 3.10 will be added after Python 3.10 is released. + + +### 2020 NumPy survey results + +_Jun 22, 2021_ -- In 2020, the NumPy survey team in partnership with students and faculty from the University of Michigan and the University of Maryland conducted the first official NumPy community survey. Find the survey results here: https://numpy.org/user-survey-2020/. + + +### Numpy 1.20.0 release + +_Jan 30, 2021_ -- [NumPy 1.20.0](https://numpy.org/doc/stable/release/1.20.0-notes.html) is now available. This is the largest NumPy release to date, thanks to 180+ contributors. The two most exciting new features are: +- Type annotations for large parts of NumPy, and a new `numpy.typing` submodule containing `ArrayLike` and `DtypeLike` aliases that users and downstream libraries can use when adding type annotations in their own code. +- Multi-platform SIMD compiler optimizations, with support for x86 (SSE, AVX), ARM64 (Neon), and PowerPC (VSX) instructions. This yielded significant performance improvements for many functions (examples: [sin/cos](https://github.com/numpy/numpy/pull/17587), [einsum](https://github.com/numpy/numpy/pull/18194)). + +### Diversity in the NumPy project + +_Sep 20, 2020_ -- We wrote a [statement on the state of, and discussion on social media around, diversity and inclusion in the NumPy project](/diversity_sep2020). + + +### First official NumPy paper published in Nature! + +_Sep 16, 2020_ -- We are pleased to announce the publication of [the first official paper on NumPy](https://www.nature.com/articles/s41586-020-2649-2) as a review article in Nature. This comes 14 years after the release of NumPy 1.0. The paper covers applications and fundamental concepts of array programming, the rich scientific Python ecosystem built on top of NumPy, and the recently added array protocols to facilitate interoperability with external array and tensor libraries like CuPy, Dask, and JAX. + + +### Python 3.9 is coming, when will NumPy release binary wheels? + +_Sept 14, 2020_ -- Python 3.9 will be released in a few weeks. If you are an early adopter of Python versions, you may be dissapointed to find that NumPy (and other binary packages like SciPy) will not have binary wheels ready on the day of the release. It is a major effort to adapt the build infrastructure to a new Python version and it typically takes a few weeks for the packages to appear on PyPI and conda-forge. In preparation for this event, please make sure to +- update your `pip` to version 20.1 at least to support `manylinux2010` and `manylinux2014` +- use [`--only-binary=numpy`](https://pip.pypa.io/en/stable/reference/pip_install/#cmdoption-only-binary) or `--only-binary=:all:` to prevent `pip` from trying to build from source. + + +### Numpy 1.19.2 release + +_Sep 10, 2020_ -- [NumPy 1.19.2](https://numpy.org/devdocs/release/1.19.2-notes.html) is now available. This latest release in the 1.19 series fixes several bugs, prepares for the [upcoming Cython 3.x release](http://docs.cython.org/en/latest/src/changes.html) and pins setuptools to keep distutils working while upstream modifications are ongoing. The aarch64 wheels are built with the latest manylinux2014 release that fixes the problem of differing page sizes used by different linux distros. + +### The inaugural NumPy survey is live! + +_Jul 2, 2020_ -- This survey is meant to guide and set priorities for decision-making about the development of NumPy as software and as a community. The survey is available in 8 additional languages besides English: Bangla, Hindi, Japanese, Mandarin, Portuguese, Russian, Spanish and French. + +Please help us make NumPy better and take the survey [here](https://umdsurvey.umd.edu/jfe/form/SV_8bJrXjbhXf7saAl). + + +### NumPy has a new logo! + +_Jun 24, 2020_ -- NumPy now has a new logo: + +NumPy logo + +The logo is a modern take on the old one, with a cleaner design. Thanks to Isabela Presedo-Floyd for designing the new logo, as well as to Travis Vaught for the old logo that served us well for 15+ years. + + +### NumPy 1.19.0 release + +_Jun 20, 2020_ -- NumPy 1.19.0 is now available. This is the first release without Python 2 support, hence it was a "clean-up release". The minimum supported Python version is now Python 3.6. An important new feature is that the random number generation infrastructure that was introduced in NumPy 1.17.0 is now accessible from Cython. + + +### Season of Docs acceptance + +_May 11, 2020_ -- NumPy has been accepted as one of the mentor organizations for the Google Season of Docs program. We are excited about the opportunity to work with a technical writer to improve NumPy's documentation once again! For more details, please see [the official Season of Docs site](https://developers.google.com/season-of-docs/) and our [ideas page](https://github.com/numpy/numpy/wiki/Google-Season-of-Docs-2020-Project-Ideas). + + +### NumPy 1.18.0 release + +_Dec 22, 2019_ -- NumPy 1.18.0 is now available. After the major changes in 1.17.0, this is a consolidation release. It is the last minor release that will support Python 3.5. Highlights of the release includes the addition of basic infrastructure for linking with 64-bit BLAS and LAPACK libraries, and a new C-API for `numpy.random`. + +Please see the [release notes](https://github.com/numpy/numpy/releases/tag/v1.18.0) for more details. + + +### NumPy receives a grant from the Chan Zuckerberg Initiative + +_Nov 15, 2019_ -- We are pleased to announce that NumPy and OpenBLAS, one of NumPy's key dependencies, have received a joint grant for $195,000 from the Chan Zuckerberg Initiative through their [Essential Open Source Software for Science program](https://chanzuckerberg.com/eoss/) that supports software maintenance, growth, development, and community engagement for open source tools critical to science. + +This grant will be used to ramp up the efforts in improving NumPy documentation, website redesign, and community development to better serve our large and rapidly growing user base, and ensure the long-term sustainability of the project. While the OpenBLAS team will focus on addressing sets of key technical issues, in particular thread-safety, AVX-512, and thread-local storage (TLS) issues, as well as algorithmic improvements in ReLAPACK (Recursive LAPACK) on which OpenBLAS depends. + +More details on our proposed initiatives and deliverables can be found in the [full grant proposal](https://figshare.com/articles/Proposal_NumPy_OpenBLAS_for_Chan_Zuckerberg_Initiative_EOSS_2019_round_1/10302167). The work is scheduled to start on Dec 1st, 2019 and continue for the next 12 months. + + + + +## Releases + +Here is a list of NumPy releases, with links to release notes. Bugfix releases (only the `z` changes in the `x.y.z` version number) have no new features; minor releases (the `y` increases) do. + +- NumPy 1.26.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.4)) -- _5 Feb 2024_. +- NumPy 1.26.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.3)) -- _2 Jan 2024_. +- NumPy 1.26.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.2)) -- _12 Nov 2023_. +- NumPy 1.26.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.1)) -- _14 Oct 2023_. +- NumPy 1.26.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.0)) -- _16 Sep 2023_. +- NumPy 1.25.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.25.2)) -- _31 Jul 2023_. +- NumPy 1.25.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.25.1)) -- _8 Jul 2023_. +- NumPy 1.24.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.4)) -- _26 Jun 2023_. +- NumPy 1.25.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.25.0)) -- _17 Jun 2023_. +- NumPy 1.24.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.3)) -- _22 Apr 2023_. +- NumPy 1.24.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.2)) -- _5 Feb 2023_. +- NumPy 1.24.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.1)) -- _26 Dec 2022_. +- NumPy 1.24.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.0)) -- _18 Dec 2022_. +- NumPy 1.23.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.5)) -- _19 Nov 2022_. +- NumPy 1.23.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.4)) -- _12 Oct 2022_. +- NumPy 1.23.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.3)) -- _9 Sep 2022_. +- NumPy 1.23.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.2)) -- _14 Aug 2022_. +- NumPy 1.23.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.1)) -- _8 Jul 2022_. +- NumPy 1.23.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.0)) -- _22 Jun 2022_. +- NumPy 1.22.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.4)) -- _20 May 2022_. +- NumPy 1.21.6 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.21.6)) -- _12 Apr 2022_. +- NumPy 1.22.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.3)) -- _7 Mar 2022_. +- NumPy 1.22.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.2)) -- _3 Feb 2022_. +- NumPy 1.22.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.1)) -- _14 Jan 2022_. +- NumPy 1.22.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.0)) -- _31 Dec 2021_. +- NumPy 1.21.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.21.5)) -- _19 Dec 2021_. +- NumPy 1.21.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.21.0)) -- _22 Jun 2021_. +- NumPy 1.20.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.20.3)) -- _10 May 2021_. +- NumPy 1.20.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.20.0)) -- _30 Jan 2021_. +- NumPy 1.19.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.19.5)) -- _5 Jan 2021_. +- NumPy 1.19.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.19.0)) -- _20 Jun 2020_. +- NumPy 1.18.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.18.4)) -- _3 May 2020_. +- NumPy 1.17.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.17.5)) -- _1 Jan 2020_. +- NumPy 1.18.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.18.0)) -- _22 Dec 2019_. +- NumPy 1.17.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.17.0)) -- _26 Jul 2019_. +- NumPy 1.16.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.16.0)) -- _14 Jan 2019_. +- NumPy 1.15.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.15.0)) -- _23 Jul 2018_. +- NumPy 1.14.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.14.0)) -- _7 Jan 2018_. diff --git a/content/ru/press-kit.md b/content/ru/press-kit.md new file mode 100644 index 0000000000..2c8970bb29 --- /dev/null +++ b/content/ru/press-kit.md @@ -0,0 +1,8 @@ +--- +title: Press kit +sidebar: false +--- + +We would like to make it easy for you to include the NumPy project identity in your next academic paper, course materials, or presentation. + +You will find several high-resolution versions of the NumPy logo [here](https://github.com/numpy/numpy/tree/main/branding/logo). Note that by using the numpy.org resources, you accept the [NumPy Code of Conduct](/code-of-conduct). diff --git a/content/ru/privacy.md b/content/ru/privacy.md new file mode 100644 index 0000000000..6064e4c4f1 --- /dev/null +++ b/content/ru/privacy.md @@ -0,0 +1,8 @@ +--- +title: Privacy Policy +sidebar: false +--- + +**numpy.org** is operated by [NumFOCUS, Inc.](https://numfocus.org), the fiscal sponsor of the NumPy project. For the Privacy Policy of this website please refer to https://numfocus.org/privacy-policy. + +If you have any questions about the policy or NumFOCUS’s data collection, use, and disclosure practices, please contact the NumFOCUS staff at privacy@numfocus.org. diff --git a/content/ru/report-handling-manual.md b/content/ru/report-handling-manual.md new file mode 100644 index 0000000000..5586668cba --- /dev/null +++ b/content/ru/report-handling-manual.md @@ -0,0 +1,95 @@ +--- +title: NumPy Code of Conduct - How to follow up on a report +sidebar: false +--- + +This is the manual followed by NumPy’s Code of Conduct Committee. It’s used when we respond to an issue to make sure we’re consistent and fair. + +Enforcing the [Code of Conduct](/code-of-conduct) impacts our community today and for the future. It’s an action that we do not take lightly. When reviewing enforcement measures, the Code of Conduct Committee will keep the following values and guidelines in mind: + +* Act in a personal manner rather than impersonal. The Committee can engage the parties to understand the situation while respecting the privacy and any necessary confidentiality of reporters. However, sometimes it is necessary to communicate with one or more individuals directly: the Committee’s goal is to improve the health of our community rather than only produce a formal decision. +* Emphasize empathy for individuals rather than judging behavior, avoiding binary labels of “good” and “bad/evil”. Overt, clear-cut aggression and harassment exist, and we will address them firmly. But many scenarios that can prove challenging to resolve are those where normal disagreements devolve into unhelpful or harmful behavior from multiple parties. Understanding the full context and finding a path that re-engages all is hard, but ultimately the most productive for our community. +* We understand that email is a difficult medium and can be isolating. Receiving criticism over email, without personal contact, can be particularly painful. This makes it especially important to keep an atmosphere of open-minded respect for the views of others. It also means that we must be transparent in our actions, and that we will do everything in our power to make sure that all our members are treated fairly and with sympathy. +* Discrimination can be subtle and it can be unconscious. It can show itself as unfairness and hostility in otherwise ordinary interactions. We know that this does occur, and we will take care to look out for it. We would very much like to hear from you if you feel you have been treated unfairly, and we will use these procedures to make sure that your complaint is heard and addressed. +* Help increase engagement in good discussion practice: try to identify where discussion may have broken down, and provide actionable information, pointers, and resources that can lead to positive change on these points. +* Be mindful of the needs of new members: provide them with explicit support and consideration, with the aim of increasing participation from underrepresented groups in particular. +* Individuals come from different cultural backgrounds and native languages. Try to identify any honest misunderstandings caused by a non-native speaker and help them understand the issue and what they can change to avoid causing offence. Complex discussion in a foreign language can be very intimidating, and we want to grow our diversity also across nationalities and cultures. + + +## Mediation + +Voluntary informal mediation is a tool at our disposal. In contexts such as when two or more parties have all escalated to the point of inappropriate behavior (something sadly common in human conflict), it may be useful to facilitate a mediation process. This is only an example: the Committee can consider mediation in any case, mindful that the process is meant to be strictly voluntary and no party can be pressured to participate. If the Committee suggests mediation, it should: + +* Find a candidate who can serve as a mediator. +* Obtain the agreement of the reporter(s). The reporter(s) have complete freedom to decline the mediation idea or to propose an alternate mediator. +* Obtain the agreement of the reported person(s). +* Settle on the mediator: while parties can propose a different mediator than the suggested candidate, only if a common agreement is reached on all terms can the process move forward. +* Establish a timeline for mediation to complete, ideally within two weeks. + +The mediator will engage with all the parties and seek a resolution that is satisfactory to all. Upon completion, the mediator will provide a report (vetted by all parties to the process) to the Committee, with recommendations on further steps. The Committee will then evaluate these results (whether a satisfactory resolution was achieved or not) and decide on any additional action deemed necessary. + + +## How the Committee will respond to reports + +When the Committee (or a Committee member) receives a report, they will first determine whether the report is about a clear and severe breach (as defined below). If so, immediate action needs to be taken in addition to the regular report handling process. + + +## Clear and severe breach actions + +We know that it is painfully common for internet communication to start at or devolve into obvious and flagrant abuse. We will deal quickly with clear and severe breaches like personal threats, violent, sexist or racist language. + +When a member of the Code of Conduct Committee becomes aware of a clear and severe breach, they will do the following: + +* Immediately disconnect the originator from all NumPy communication channels. +* Reply to the reporter that their report has been received and that the originator has been disconnected. +* In every case, the moderator should make a reasonable effort to contact the originator, and tell them specifically how their language or actions qualify as a “clear and severe breach”. The moderator should also say that, if the originator believes this is unfair or they want to be reconnected to NumPy, they have the right to ask for a review, as below, by the Code of Conduct Committee. The moderator should copy this explanation to the Code of Conduct Committee. +* The Code of Conduct Committee will formally review and sign off on all cases where this mechanism has been applied to make sure it is not being used to control ordinary heated disagreement. + + +## Report handling + +When a report is sent to the Committee they will immediately reply to the reporter to confirm receipt. This reply must be sent within 72 hours, and the group should strive to respond much quicker than that. + +If a report doesn’t contain enough information, the Committee will obtain all relevant data before acting. The Committee is empowered to act on the Steering Council’s behalf in contacting any individuals involved to get a more complete account of events. + +The Committee will then review the incident and determine, to the best of their ability: + +* What happened. +* Whether this event constitutes a Code of Conduct violation. +* Who are the responsible party(ies). +* Whether this is an ongoing situation, and there is a threat to anyone’s physical safety. + +This information will be collected in writing, and whenever possible the group’s deliberations will be recorded and retained (i.e. chat transcripts, email discussions, recorded conference calls, summaries of voice conversations, etc). + +It is important to retain an archive of all activities of this Committee to ensure consistency in behavior and provide institutional memory for the project. To assist in this, the default channel of discussion for this Committee will be a private mailing list accessible to current and future members of the Committee as well as members of the Steering Council upon justified request. If the Committee finds the need to use off-list communications (e.g. phone calls for early/rapid response), it should in all cases summarize these back to the list so there’s a good record of the process. + +The Code of Conduct Committee should aim to have a resolution agreed upon within two weeks. In the event that a resolution can’t be determined in that time, the Committee will respond to the reporter(s) with an update and projected timeline for resolution. + + +## Resolutions + +The Committee must agree on a resolution by consensus. If the group cannot reach consensus and deadlocks for over a week, the group will turn the matter over to the Steering Council for resolution. + +Possible responses may include: + +* Taking no further action: + - if we determine no violations have occurred; + - if the matter has been resolved publicly while the Committee was considering responses. +* Coordinating voluntary mediation: if all involved parties agree, the Committee may facilitate a mediation process as detailed above. +* Remind publicly, and point out that some behavior/actions/language have been judged inappropriate and why in the current context, or can but hurtful to some people, requesting the community to self-adjust. +* A private reprimand from the Committee to the individual(s) involved. In this case, the group chair will deliver that reprimand to the individual(s) over email, cc’ing the group. +* A public reprimand. In this case, the Committee chair will deliver that reprimand in the same venue that the violation occurred, within the limits of practicality. E.g., the original mailing list for an email violation, but for a chat room discussion where the person/context may be gone, they can be reached by other means. The group may choose to publish this message elsewhere for documentation purposes. +* A request for a public or private apology, assuming the reporter agrees to this idea: they may at their discretion refuse further contact with the violator. The chair will deliver this request. The Committee may, if it chooses, attach “strings” to this request: for example, the group may ask a violator to apologize in order to retain one’s membership on a mailing list. +* A “mutually agreed upon hiatus” where the Committee asks the individual to temporarily refrain from community participation. If the individual chooses not to take a temporary break voluntarily, the Committee may issue a “mandatory cooling off period”. +* A permanent or temporary ban from some or all NumPy spaces (mailing lists, gitter.im, etc.). The group will maintain records of all such bans so that they may be reviewed in the future or otherwise maintained. + +Once a resolution is agreed upon, but before it is enacted, the Committee will contact the original reporter and any other affected parties and explain the proposed resolution. The Committee will ask if this resolution is acceptable, and must note feedback for the record. + +Finally, the Committee will make a report to the NumPy Steering Council (as well as the NumPy core team in the event of an ongoing resolution, such as a ban). + +The Committee will never publicly discuss the issue; all public statements will be made by the chair of the Code of Conduct Committee or the NumPy Steering Council. + + +## Conflicts of Interest + +In the event of any conflict of interest, a Committee member must immediately notify the other members, and recuse themselves if necessary. diff --git a/content/ru/tabcontents.yaml b/content/ru/tabcontents.yaml new file mode 100644 index 0000000000..d74cba9bce --- /dev/null +++ b/content/ru/tabcontents.yaml @@ -0,0 +1,373 @@ +params: + machinelearning: + paras: + - + para1: NumPy forms the basis of powerful machine learning libraries like [scikit-learn](https://scikit-learn.org) and [SciPy](https://www.scipy.org). As machine learning grows, so does the list of libraries built on NumPy. [TensorFlow’s](https://www.tensorflow.org) deep learning capabilities have broad applications — among them speech and image recognition, text-based applications, time-series analysis, and video detection. [PyTorch](https://pytorch.org), another deep learning library, is popular among researchers in computer vision and natural language processing. + para2: Statistical techniques called [ensemble](https://towardsdatascience.com/ensemble-methods-bagging-boosting-and-stacking-c9214a10a205) methods such as binning, bagging, stacking, and boosting are among the ML algorithms implemented by tools such as [XGBoost](https://xgboost.readthedocs.io/), [LightGBM](https://lightgbm.readthedocs.io/en/latest/), and [CatBoost](https://catboost.ai) — one of the fastest inference engines. [Yellowbrick](https://www.scikit-yb.org/en/latest/) and [Eli5](https://eli5.readthedocs.io/en/latest/) offer machine learning visualizations. + arraylibraries: + intro: + - + text: NumPy's API is the starting point when libraries are written to exploit innovative hardware, create specialized array types, or add capabilities beyond what NumPy provides. + headers: + - + text: Array Library + - + text: Capabilities & Application areas + libraries: + - + title: Dask + text: Distributed arrays and advanced parallelism for analytics, enabling performance at scale. + img: /images/content_images/arlib/dask.png + alttext: Dask + url: https://dask.org/ + - + title: CuPy + text: NumPy-compatible array library for GPU-accelerated computing with Python. + img: /images/content_images/arlib/cupy.png + alttext: CuPy + url: https://cupy.dev + - + title: JAX + text: "Composable transformations of NumPy programs: differentiate, vectorize, just-in-time compilation to GPU/TPU." + img: /images/content_images/arlib/jax_logo_250px.png + alttext: JAX + url: https://jax.readthedocs.io/ + - + title: Xarray + text: Labeled, indexed multi-dimensional arrays for advanced analytics and visualization. + img: /images/content_images/arlib/xarray.png + alttext: xarray + url: https://xarray.pydata.org/en/stable/index.html + - + title: Sparse + text: NumPy-compatible sparse array library that integrates with Dask and SciPy's sparse linear algebra. + img: /images/content_images/arlib/sparse.png + alttext: sparse + url: https://sparse.pydata.org/en/latest/ + - + title: PyTorch + text: Deep learning framework that accelerates the path from research prototyping to production deployment. + img: /images/content_images/arlib/pytorch-logo-dark.svg + alttext: PyTorch + url: https://pytorch.org/ + - + title: TensorFlow + text: An end-to-end platform for machine learning to easily build and deploy ML powered applications. + img: /images/content_images/arlib/tensorflow-logo.svg + alttext: TensorFlow + url: https://www.tensorflow.org + - + title: Arrow + text: A cross-language development platform for columnar in-memory data and analytics. + img: /images/content_images/arlib/arrow.png + alttext: arrow + url: https://arrow.apache.org/ + - + title: xtensor + text: Multi-dimensional arrays with broadcasting and lazy computing for numerical analysis. + img: /images/content_images/arlib/xtensor.png + alttext: xtensor + url: https://github.com/xtensor-stack/xtensor-python + - + title: Awkward Array + text: Manipulate JSON-like data with NumPy-like idioms. + img: /images/content_images/arlib/awkward.svg + alttext: awkward + url: https://awkward-array.org/ + - + title: uarray + text: Python backend system that decouples API from implementation; unumpy provides a NumPy API. + img: /images/content_images/arlib/uarray.png + alttext: uarray + url: https://uarray.org/en/latest/ + - + title: tensorly + text: Tensor learning, algebra and backends to seamlessly use NumPy, PyTorch, TensorFlow or CuPy. + img: /images/content_images/arlib/tensorly.png + alttext: tensorly + url: http://tensorly.org/stable/home.html + scientificdomains: + intro: + - + text: Nearly every scientist working in Python draws on the power of NumPy. + - + text: "NumPy brings the computational power of languages like C and Fortran to Python, a language much easier to learn and use. With this power comes simplicity: a solution in NumPy is often clear and elegant." + libraries: + - + title: Quantum Computing + alttext: A computer chip. + img: /images/content_images/sc_dom_img/quantum_computing.svg + links: + - + url: http://qutip.org + label: QuTiP + - + url: https://pyquil-docs.rigetti.com/en/stable + label: PyQuil + - + url: https://qiskit.org + label: Qiskit + - + url: https://pennylane.ai + label: PennyLane + - + title: Statistical Computing + alttext: A line graph with the line moving up. + img: /images/content_images/sc_dom_img/statistical_computing.svg + links: + - + url: https://pandas.pydata.org/ + label: Pandas + - + url: https://www.statsmodels.org/ + label: statsmodels + - + url: https://xarray.pydata.org/en/stable/ + label: Xarray + - + url: https://seaborn.pydata.org/ + label: Seaborn + - + title: Signal Processing + alttext: A bar chart with positive and negative values. + img: /images/content_images/sc_dom_img/signal_processing.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://pywavelets.readthedocs.io/ + label: PyWavelets + - + url: https://python-control.org/ + label: python-control + - + url: https://hyperspy.org/ + label: HyperSpy + - + title: Image Processing + alttext: An photograph of the mountains. + img: /images/content_images/sc_dom_img/image_processing.svg + links: + - + url: https://scikit-image.org/ + label: Scikit-image + - + url: https://opencv.org/ + label: OpenCV + - + url: https://mahotas.rtfd.io/ + label: Mahotas + - + title: Graphs and Networks + alttext: A simple graph. + img: /images/content_images/sc_dom_img/sd6.svg + links: + - + url: https://networkx.org/ + label: NetworkX + - + url: https://graph-tool.skewed.de/ + label: graph-tool + - + url: https://igraph.org/python/ + label: igraph + - + url: https://pygsp.rtfd.io/ + label: PyGSP + - + title: Astronomy + alttext: A telescope. + img: /images/content_images/sc_dom_img/astronomy_processes.svg + links: + - + url: https://www.astropy.org/ + label: AstroPy + - + url: https://sunpy.org/ + label: SunPy + - + url: https://spacepy.github.io/ + label: SpacePy + - + title: Cognitive Psychology + alttext: A human head with gears. + img: /images/content_images/sc_dom_img/cognitive_psychology.svg + links: + - + url: https://www.psychopy.org/ + label: PsychoPy + - + title: Bioinformatics + alttext: A strand of DNA. + img: /images/content_images/sc_dom_img/bioinformatics.svg + links: + - + url: https://biopython.org/ + label: BioPython + - + url: http://scikit-bio.org/ + label: Scikit-Bio + - + url: https://github.com/openvax/pyensembl + label: PyEnsembl + - + url: http://etetoolkit.org/ + label: ETE + - + title: Bayesian Inference + alttext: A graph with a bell-shaped curve. + img: /images/content_images/sc_dom_img/bayesian_inference.svg + links: + - + url: https://pystan.readthedocs.io/en/latest/ + label: PyStan + - + url: https://docs.pymc.io/ + label: PyMC3 + - + url: https://arviz-devs.github.io/arviz/ + label: ArviZ + - + url: https://emcee.readthedocs.io/ + label: emcee + - + title: Mathematical Analysis + alttext: Four mathematical symbols. + img: /images/content_images/sc_dom_img/mathematical_analysis.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://www.sympy.org/ + label: SymPy + - + url: https://www.cvxpy.org/ + label: cvxpy + - + url: https://fenicsproject.org/ + label: FEniCS + - + title: Chemistry + alttext: A test tube. + img: /images/content_images/sc_dom_img/chemistry.svg + links: + - + url: https://cantera.org/ + label: Cantera + - + url: https://www.mdanalysis.org/ + label: MDAnalysis + - + url: https://github.com/rdkit/rdkit + label: RDKit + - + url: https://www.pybamm.org/ + label: PyBaMM + - + title: Geoscience + alttext: The Earth. + img: /images/content_images/sc_dom_img/geoscience.svg + links: + - + url: https://pangeo.io/ + label: Pangeo + - + url: https://simpeg.xyz/ + label: Simpeg + - + url: https://github.com/obspy/obspy/wiki + label: ObsPy + - + url: https://www.fatiando.org/ + label: Fatiando a Terra + - + title: Geographic Processing + alttext: A map. + img: /images/content_images/sc_dom_img/GIS.svg + links: + - + url: https://shapely.readthedocs.io/ + label: Shapely + - + url: https://geopandas.org/ + label: GeoPandas + - + url: https://python-visualization.github.io/folium + label: Folium + - + title: Architecture & Engineering + alttext: A microprocessor development board. + img: /images/content_images/sc_dom_img/robotics.svg + links: + - + url: https://compas.dev/ + label: COMPAS + - + url: https://cityenergyanalyst.com/ + label: City Energy Analyst + - + url: https://nortikin.github.io/sverchok/ + label: Sverchok + datascience: + intro: "NumPy lies at the core of a rich ecosystem of data science libraries. A typical exploratory data science workflow might look like:" + image1: + - + img: /images/content_images/ds-landscape.png + alttext: Diagram of Python Libraries. The five catagories are 'Extract, Transform, Load', 'Data Exploration', 'Data Modeling', 'Data Evaluation' and 'Data Presentation'. + image2: + - + img: /images/content_images/data-science.png + alttext: Diagram of three overlapping circles. The circles are labeled 'Mathematics', 'Computer Science' and 'Domain Expertise'. In the middle of the diagram, which has the three circles overlapping it, is an area labeled 'Data Science'. + examples: + - + text: "Extract, Transform, Load: [Pandas](https://pandas.pydata.org), [Intake](https://intake.readthedocs.io), [PyJanitor](https://pyjanitor-devs.github.io/pyjanitor/)" + - + text: "Exploratory analysis: [Jupyter](https://jupyter.org), [Seaborn](https://seaborn.pydata.org), [Matplotlib](https://matplotlib.org), [Altair](https://altair-viz.github.io)" + - + text: "Model and evaluate: [scikit-learn](https://scikit-learn.org), [statsmodels](https://www.statsmodels.org/stable/index.html), [PyMC3](https://docs.pymc.io), [spaCy](https://spacy.io)" + - + text: "Report in a dashboard: [Dash](https://plotly.com/dash), [Panel](https://panel.holoviz.org), [Voila](https://voila.readthedocs.io/)" + content: + - + text: For high data volumes, [Dask](https://dask.org) and [Ray](https://ray.io/) are designed to scale. Stable deployments rely on data versioning ([DVC](https://dvc.org)), experiment tracking ([MLFlow](https://mlflow.org)), and workflow automation ([Airflow](https://airflow.apache.org), [Dagster](https://dagster.io) and [Prefect](https://www.prefect.io)). + visualization: + images: + - + url: https://www.fusioncharts.com/blog/best-python-data-visualization-libraries + img: /images/content_images/v_matplotlib.png + alttext: A streamplot made in matplotlib + - + url: https://github.com/yhat/ggpy + img: /images/content_images/v_ggpy.png + alttext: A scatter-plot graph made in ggpy + - + url: https://www.journaldev.com/19692/python-plotly-tutorial + img: /images/content_images/v_plotly.png + alttext: A box-plot made in plotly + - + url: https://altair-viz.github.io/gallery/streamgraph.html + img: /images/content_images/v_altair.png + alttext: A streamgraph made in altair + - + url: https://seaborn.pydata.org + img: /images/content_images/v_seaborn.png + alttext: A pairplot of two types of graph, a plot-graph and a frequency graph made in seaborn" + - + url: https://docs.pyvista.org/ + img: /images/content_images/v_pyvista.png + alttext: A 3D volume rendering made in PyVista. + - + url: https://napari.org + img: /images/content_images/v_napari.png + alttext: A multi-dimensionan image made in napari. + - + url: https://vispy.org/gallery/index.html + img: /images/content_images/v_vispy.png + alttext: A Voronoi diagram made in vispy. + content: + - + text: NumPy is an essential component in the burgeoning [Python visualization landscape](https://pyviz.org/overviews/index.html), which includes [Matplotlib](https://matplotlib.org), [Seaborn](https://seaborn.pydata.org), [Plotly](https://plot.ly), [Altair](https://altair-viz.github.io), [Bokeh](https://docs.bokeh.org/en/latest/), [Holoviz](https://holoviz.org), [Vispy](http://vispy.org), [Napari](https://napari.org/), and [PyVista](https://docs.pyvista.org/), to name a few. + - + text: NumPy's accelerated processing of large arrays allows researchers to visualize datasets far larger than native Python could handle. diff --git a/content/ru/teams.md b/content/ru/teams.md new file mode 100644 index 0000000000..91cf5ca399 --- /dev/null +++ b/content/ru/teams.md @@ -0,0 +1,22 @@ +--- +title: NumPy Teams +sidebar: false +--- + +We are an international team on a mission to support scientific and research communities worldwide by building quality, open-source software. [Join us]({{< relref "/contribute" >}})! + +{{< include-html "static/gallery/maintainers.html" >}} + +{{< include-html "static/gallery/docs-team.html" >}} + +{{< include-html "static/gallery/web-team.html" >}} + +{{< include-html "static/gallery/triage-team.html" >}} + +{{< include-html "static/gallery/survey-team.html" >}} + +{{< include-html "static/gallery/emeritus-maintainers.html" >}} + +# Governance + +For the list of the Steering Council members, please see [here](https://numpy.org/about/). diff --git a/content/ru/user-survey-2020.md b/content/ru/user-survey-2020.md new file mode 100644 index 0000000000..b4349bcb7d --- /dev/null +++ b/content/ru/user-survey-2020.md @@ -0,0 +1,18 @@ +--- +title: 2020 NUMPY COMMUNITY SURVEY +sidebar: false +--- + +In 2020, the NumPy survey team in partnership with students and faculty from a Master’s course in Survey Methodology jointly hosted by the University of Michigan and the University of Maryland conducted the first official NumPy community survey. Over 1,200 users from 75 countries participated to help us map out a landscape of the NumPy community and voiced their thoughts about the future of the project. + +{{< figure >}} +src = '/surveys/NumPy_usersurvey_2020_report_cover.png' alt = 'Cover page of the 2020 NumPy user survey report, titled "NumPy Community Survey 2020 - results"' width = '250' +{{< /figure >}} + +**[Download the report](/surveys/NumPy_usersurvey_2020_report.pdf)** to take a closer look at the survey findings. + + +For the highlights, check out **[this infographic](https://github.com/numpy/numpy-surveys/blob/master/images/2020NumPysurveyresults_community_infographic.pdf)**. + +Ready for a deep dive? Visit **https://numpy.org/user-survey-2020-details/**. + diff --git a/content/ru/user-surveys.md b/content/ru/user-surveys.md new file mode 100644 index 0000000000..89a2aa0460 --- /dev/null +++ b/content/ru/user-surveys.md @@ -0,0 +1,10 @@ +--- +title: NUMPY USER SURVEYS +sidebar: false +--- + +**2020** The NumPy survey team in partnership with students and faculty from the University of Michigan and the University of Maryland conducted the first official NumPy community survey. Find the survey results [here](https://numpy.org/user-survey-2020/). + +**2021** The collected data is currently being analyzed. + +If you have any questions or suggestions for the past or future surveys, please open an issue [here](https://github.com/numpy/numpy-surveys/issues). diff --git a/content/zh/404.md b/content/zh/404.md new file mode 100644 index 0000000000..0f27a36ee9 --- /dev/null +++ b/content/zh/404.md @@ -0,0 +1,8 @@ +--- +title: 404 +sidebar: false +--- + +抱歉······ 目标网页并不存在。 + +如果您认为这个页面应该展示些什么东西，请在 GitHub 上面 [发起一个 issue](https://github.com/numpy/numpy.org/issues). diff --git a/content/zh/_index.md b/content/zh/_index.md new file mode 100644 index 0000000000..be88f9e642 --- /dev/null +++ b/content/zh/_index.md @@ -0,0 +1,49 @@ +--- +title: null +--- + +{{< grid columns="1 2 2 3" >}} + +[[item]] +type = 'card' +title = 'Powerful N-dimensional arrays' +body = ''' +Fast and versatile, the NumPy vectorization, indexing, and broadcasting concepts are the de-facto standards of array computing today. +''' + +[[item]] +type = 'card' +title = 'Numerical computing tools' +body = ''' +NumPy offers comprehensive mathematical functions, random number generators, linear algebra routines, Fourier transforms, and more. +''' + +[[item]] +type = 'card' +title = 'Open source' +body = ''' +Distributed under a liberal [BSD license](https://github.com/numpy/numpy/blob/main/LICENSE.txt), NumPy is developed and maintained [publicly on GitHub](https://github.com/numpy/numpy) by a vibrant, responsive, and diverse [community](/community). +''' + +[[item]] +type = 'card' +title = 'Interoperable' +body = ''' +NumPy supports a wide range of hardware and computing platforms, and plays well with distributed, GPU, and sparse array libraries. +''' + +[[item]] +type = 'card' +title = 'Performant' +body = ''' +The core of NumPy is well-optimized C code. Enjoy the flexibility of Python with the speed of compiled code. +''' + +[[item]] +type = 'card' +title = 'Easy to use' +body = ''' +NumPy's high level syntax makes it accessible and productive for programmers from any background or experience level. +''' + +{{< /grid>}} diff --git a/content/zh/about.md b/content/zh/about.md new file mode 100644 index 0000000000..243f7083ae --- /dev/null +++ b/content/zh/about.md @@ -0,0 +1,90 @@ +--- +title: About Us +sidebar: false +--- + +NumPy is an open source project that enables numerical computing with Python. It was created in 2005 building on the early work of the Numeric and Numarray libraries. NumPy will always be 100% open source software and free for all to use. It is released under the liberal terms of the [modified BSD license](https://github.com/numpy/numpy/blob/main/LICENSE.txt). + +NumPy is developed in the open on GitHub, through the consensus of the NumPy and wider scientific Python community. For more information on our governance approach, please see our [Governance Document](https://www.numpy.org/devdocs/dev/governance/index.html). + + +## Steering Council + +The NumPy Steering Council is the project's governing body. Its role is to ensure, through working with and serving the broader NumPy community, the long-term sustainability of the project, both as a software package and community. The NumPy Steering Council currently consists of the following members (in alphabetical order, by last name): + +- Sebastian Berg +- Ralf Gommers +- Charles Harris +- Stephan Hoyer +- Inessa Pawson +- Matti Picus +- Stéfan van der Walt +- Melissa Weber Mendonça +- Eric Wieser + +Emeritus: + +- Alex Griffing (2015-2017) +- Allan Haldane (2015-2021) +- Marten van Kerkwijk (2017-2019) +- Travis Oliphant (project founder, 2005-2012) +- Nathaniel Smith (2012-2021) +- Julian Taylor (2013-2021) +- Jaime Fernández del Río (2014-2021) +- Pauli Virtanen (2008-2021) + +To contact the NumPy Steering Council, please email numpy-team@googlegroups.com. + +## Teams + +The NumPy project leadership is actively working on diversifying contribution pathways to the project.
NumPy currently has the following teams: + +- development +- documentation +- triage +- website +- survey +- translations +- sprint mentors +- optimization +- funding and grants + +See the [Team](/teams) page for more info. + +## NumFOCUS Subcommittee + +- Charles Harris +- Ralf Gommers +- Inessa Pawson +- Sebastian Berg +- External member: Thomas Caswell + +## Sponsors + +NumPy receives direct funding from the following sources: +{{< sponsors >}} + + +## Institutional Partners + +Institutional Partners are organizations that support the project by employing people that contribute to NumPy as part of their job. Current Institutional Partners include: + +- UC Berkeley (Stéfan van der Walt) +- Quansight (Nathan Goldbaum, Ralf Gommers, Matti Picus, Melissa Weber Mendonça) +- NVIDIA (Sebastian Berg) + +{{< partners >}} + + +## Donate + +If you have found NumPy useful in your work, research, or company, please consider a donation to the project commensurate with your resources. Any amount helps! All donations will be used strictly to fund the development of NumPy’s open source software, documentation, and community. + +NumPy is a Sponsored Project of NumFOCUS, a 501(c)(3) nonprofit charity in the United States. NumFOCUS provides NumPy with fiscal, legal, and administrative support to help ensure the health and sustainability of the project. Visit [numfocus.org](https://numfocus.org) for more information. + +Donations to NumPy are managed by [NumFOCUS](https://numfocus.org). For donors in the United States, your gift is tax-deductible to the extent provided by law. As with any donation, you should consult with your tax advisor about your particular tax situation. + +NumPy's Steering Council will make the decisions on how to best use any funds received. Technical and infrastructure priorities are documented on the [NumPy Roadmap](https://www.numpy.org/neps/index.html#roadmap). + +{{}} + diff --git a/content/zh/arraycomputing.md b/content/zh/arraycomputing.md new file mode 100644 index 0000000000..abd29d11c1 --- /dev/null +++ b/content/zh/arraycomputing.md @@ -0,0 +1,21 @@ +--- +title: Array Computing +sidebar: false +--- + +*Array computing is the foundation of statistical, mathematical, scientific computing in various contemporary data science and analytics applications such as data visualization, digital signal processing, image processing, bioinformatics, machine learning, AI, and several others.* + +Large scale data manipulation and transformation depends on efficient, high-performance array computing. The language of choice for data analytics, machine learning, and productive numerical computing is **Python.** + +**Num**erical **Py**thon or NumPy is its de-facto standard Python programming language library that supports large, multi-dimensional arrays and matrices, and comes with a vast collection of high-level mathematical functions to operate on these arrays. + +Since the launch of NumPy in 2006, Pandas appeared on the landscape in 2008, and it was not until a couple of years ago that several array computing libraries showed up in succession, crowding the array computing landscape. Many of these newer libraries mimic NumPy-like features and capabilities, and pack newer algorithms and features geared towards machine learning and artificial intelligence applications. + +arraycl + +**Array computing** is based on **arrays** data structures. *Arrays* are used to organize vast amounts of data such that a related set of values can be easily sorted, searched, mathematically manipulated, and transformed easily and quickly. + +Array computing is *unique* as it involves operating on the data array *at once*. What this means is that any array operation applies to an entire set of values in one shot. This vectorized approach provides speed and simplicity by enabling programmers to code and operate on aggregates of data, without having to use loops of individual scalar operations. diff --git a/content/zh/case-studies/blackhole-image.md b/content/zh/case-studies/blackhole-image.md new file mode 100644 index 0000000000..22334ebed0 --- /dev/null +++ b/content/zh/case-studies/blackhole-image.md @@ -0,0 +1,80 @@ +--- +title: "Case Study: First Image of a Black Hole" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/blackhole.jpg' title = 'Black Hole M87' alt = 'black hole image' attribution = '(Image Credits: Event Horizon Telescope Collaboration)' attributionlink = 'https://www.jpl.nasa.gov/images/universe/20190410/blackhole20190410.jpg' +{{< /figure >}} + +{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="Katie Bouman, *Assistant Professor, Computing & Mathematical Sciences, Caltech*" +> }} Imaging the M87 Black Hole is like trying to see something that is by definition impossible to see. +> +> {{< /blockquote >}} + +## A telescope the size of the earth + +The [Event Horizon telescope (EHT)](https://eventhorizontelescope.org) is an array of eight ground-based radio telescopes forming a computational telescope the size of the earth, studing the universe with unprecedented sensitivity and resolution. The huge virtual telescope, which uses a technique called very-long-baseline interferometry (VLBI), has an angular resolution of [20 micro-arcseconds][resolution] — enough to read a newspaper in New York from a sidewalk café in Paris! + +### Key Goals and Results + +* **A New View of the Universe:** The groundwork for the EHT's groundbreaking image had been laid 100 years earlier when [Sir Arthur Eddington][eddington] yielded the first observational support of Einstein's theory of general relativity. + +* **The Black Hole:** EHT was trained on a supermassive black hole approximately 55 million light-years from Earth, lying at the center of the galaxy Messier 87 (M87) in the Virgo galaxy cluster. Its mass is 6.5 billion times the Sun's. It had been studied for [over 100 years](https://www.jpl.nasa.gov/news/news.php?feature=7385), but never before had a black hole been visually observed. + +* **Comparing Observations to Theory:** From Einstein’s general theory of relativity, scientists expected to find a shadow-like region caused by gravitational bending and capture of light. Scientists could use it to measure the black hole's enormous mass. + +### The Challenges + +* **Computational scale** + + EHT poses massive data-processing challenges, including rapid atmospheric phase fluctuations, large recording bandwidth, and telescopes that are widely dissimilar and geographically dispersed. + +* **Too much information** + + Each day EHT generates over 350 terabytes of observations, stored on helium-filled hard drives. Reducing the volume and complexity of this much data is enormously difficult. + +* **Into the unknown** + + When the goal is to see something never before seen, how can scientists be confident the image is correct? + +{{< figure >}} +src = '/images/content_images/cs/dataprocessbh.png' title = 'EHT Data Processing Pipeline' alt = 'data pipeline' align = 'center' attribution = '(Diagram Credits: The Astrophysical Journal, Event Horizon Telescope Collaboration)' attributionlink = 'https://iopscience.iop.org/article/10.3847/2041-8213/ab0c57' +{{< /figure >}} + +## NumPy’s Role + +What if there's a problem with the data? Or perhaps an algorithm relies too heavily on a particular assumption. Will the image change drastically if a single parameter is changed? + +The EHT collaboration met these challenges by having independent teams evaluate the data, using both established and cutting-edge image reconstruction techniques. When results proved consistent, they were combined to yield the first-of-a-kind image of the black hole. + +Their work illustrates the role the scientific Python ecosystem plays in advancing science through collaborative data analysis. + +{{< figure >}} +src = '/images/content_images/cs/bh_numpy_role.png' alt = 'role of numpy' title = 'The role of NumPy in Black Hole imaging' +{{< /figure >}} + +For example, the [`eht-imaging`][ehtim] Python package provides tools for simulating and performing image reconstruction on VLBI data. NumPy is at the core of array data processing used in this package, as illustrated by the partial software dependency chart below. + +{{< figure >}} +src = '/images/content_images/cs/ehtim_numpy.png' alt = 'ehtim dependency map highlighting numpy' title = 'Software dependency chart of ehtim package highlighting NumPy' +{{< /figure >}} + +Besides NumPy, many other packages, such as [SciPy](https://www.scipy.org) and [Pandas](https://pandas.io), are part of the data processing pipeline for imaging the black hole. The standard astronomical file formats and time/coordinate transformations were handled by [Astropy][astropy], while [Matplotlib][mpl] was used in visualizing data throughout the analysis pipeline, including the generation of the final image of the black hole. + +## Summary + +The efficient and adaptable n-dimensional array that is NumPy's central feature enabled researchers to manipulate large numerical datasets, providing a foundation for the first-ever image of a black hole. A landmark moment in science, it gives stunning visual evidence of Einstein’s theory. The achievement encompasses not only technological breakthroughs but also international collaboration among over 200 scientists and some of the world's best radio observatories. Innovative algorithms and data processing techniques, improving upon existing astronomical models, helped unfold a mystery of the universe. + +{{< figure >}} +src = '/images/content_images/cs/numpy_bh_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} + +[resolution]: https://eventhorizontelescope.org/press-release-april-10-2019-astronomers-capture-first-image-black-hole + +[eddington]: https://en.wikipedia.org/wiki/Eddington_experiment + +[ehtim]: https://github.com/achael/eht-imaging + +[astropy]: https://www.astropy.org/ +[mpl]: https://matplotlib.org/ diff --git a/content/zh/case-studies/cricket-analytics.md b/content/zh/case-studies/cricket-analytics.md new file mode 100644 index 0000000000..d0be271f50 --- /dev/null +++ b/content/zh/case-studies/cricket-analytics.md @@ -0,0 +1,72 @@ +--- +title: "Case Study: Cricket Analytics, the game changer!" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/ipl-stadium.png' title = 'IPLT20, the biggest Cricket Festival in India' alt = 'Indian Premier League Cricket cup and stadium' attribution = '(Image credits: IPLT20 (cup and logo) & Akash Yadav (stadium))' attributionlink = 'https://unsplash.com/@aksh1802' +{{< /figure >}} + +{{< blockquote cite="https://www.scoopwhoop.com/sports/ms-dhoni/" by="M S Dhoni, *International Cricket Player, ex-captain, Indian Team, plays for Chennai Super Kings in IPL*" +> }} You don't play for the crowd, you play for the country. +> +> {{< /blockquote >}} + +## About Cricket + +It would be an understatement to state that Indians love cricket. The game is played in just about every nook and cranny of India, rural or urban, popular with the young and the old alike, connecting billions in India unlike any other sport. Cricket enjoys lots of media attention. There is a significant amount of [money](https://www.statista.com/topics/4543/indian-premier-league-ipl/) and fame at stake. Over the last several years, technology has literally been a game changer. Audiences are spoilt for choice with streaming media, tournaments, affordable access to mobile based live cricket watching, and more. + +The Indian Premier League (IPL) is a professional Twenty20 cricket league, founded in 2008. It is one of the most attended cricketing events in the world, valued at [$6.7 billion](https://en.wikipedia.org/wiki/Indian_Premier_League) in 2019. + +Cricket is a game of numbers - the runs scored by a batsman, the wickets taken by a bowler, the matches won by a cricket team, the number of times a batsman responds in a certain way to a kind of bowling attack, etc. The capability to dig into cricketing numbers for both improving performance and studying the business opportunities, overall market, and economics of cricket via powerful analytics tools, powered by numerical computing software such as NumPy, is a big deal. Cricket analytics provides interesting insights into the game and predictive intelligence regarding game outcomes. + +Today, there are rich and almost infinite troves of cricket game records and statistics available, e.g., [ESPN cricinfo](https://stats.espncricinfo.com/ci/engine/stats/index.html) and [cricsheet](https://cricsheet.org). These and several such cricket databases have been used for [cricket analysis](https://www.researchgate.net/publication/336886516_Data_visualization_and_toss_related_analysis_of_IPL_teams_and_batsmen_performances) using the latest machine learning and predictive modelling algorithms. Media and entertainment platforms along with professional sports bodies associated with the game use technology and analytics for determining key metrics for improving match winning chances: + +* batting performance moving average, +* score forecasting, +* gaining insights into fitness and performance of a player against different opposition, +* player contribution to wins and losses for making strategic decisions on team composition + +{{< figure >}} +src = '/images/content_images/cs/cricket-pitch.png' title = 'Cricket Pitch, the focal point in the field' alt = 'A cricket pitch with bowler and batsmen' align = 'center' attribution = '(Image credit: Debarghya Das)' attributionlink = 'http://debarghyadas.com/files/IPLpaper.pdf' +{{< /figure >}} + +### Key Data Analytics Objectives + +* Sports data analytics are used not only in cricket but many [other sports](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx) for improving the overall team performance and maximizing winning chances. +* Real-time data analytics can help in gaining insights even during the game for changing tactics by the team and by associated businesses for economic benefits and growth. +* Besides historical analysis, predictive models are harnessed to determine the possible match outcomes that require significant number crunching and data science know-how, visualization tools and capability to include newer observations in the analysis. + +{{< figure >}} +src = '/images/content_images/cs/player-pose-estimator.png' alt = 'pose estimator' title = 'Cricket Pose Estimator' attribution = '(Image credit: connect.vin)' attributionlink = 'https://connect.vin/2019/05/ai-for-cricket-batsman-pose-analysis/' +{{< /figure >}} + +### The Challenges + +* **Data Cleaning and preprocessing** + + IPL has expanded cricket beyond the classic test match format to a much larger scale. The number of matches played every season across various formats has increased and so has the data, the algorithms, newer sports data analysis technologies and simulation models. Cricket data analysis requires field mapping, player tracking, ball tracking, player shot analysis, and several other aspects involved in how the ball is delivered, its angle, spin, velocity, and trajectory. All these factors together have increased the complexity of data cleaning and preprocessing. + +* **Dynamic Modeling** + + In cricket, just like any other sport, there can be a large number of variables related to tracking various numbers of players on the field, their attributes, the ball, and several possibilities of potential actions. The complexity of data analytics and modeling is directly proportional to the kind of predictive questions that are put forth during analysis and are highly dependent on data representation and the model. Things get even more challenging in terms of computation, data comparisons when dynamic cricket play predictions are sought such as what would have happened if the batsman had hit the ball at a different angle or velocity. + +* **Predictive Analytics Complexity** + + Much of the decision making in cricket is based on questions such as "how often does a batsman play a certain kind of shot if the ball delivery is of a particular type", or "how does a bowler change his line and length if the batsman responds to his delivery in a certain way". This kind of predictive analytics query requires highly granular dataset availability and the capability to synthesize data and create generative models that are highly accurate. + +## NumPy’s Role in Cricket Analytics + +Sports Analytics is a thriving field. Many researchers and companies [use NumPy](https://adtmag.com/blogs/dev-watch/2017/07/sports-analytics.aspx) and other PyData packages like Scikit-learn, SciPy, Matplotlib, and Jupyter, besides using the latest machine learning and AI techniques. NumPy has been used for various kinds of cricket related sporting analytics such as: + +* **Statistical Analysis:** NumPy's numerical capabilities help estimate the statistical significance of observational data or match events in the context of various player and game tactics, estimating the game outcome by comparison with a generative or static model. [Causal analysis](https://amplitude.com/blog/2017/01/19/causation-correlation) and [big data approaches](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4996805/) are used for tactical analysis. + +* **Data Visualization:** Data graphing and [visualization](https://towardsdatascience.com/advanced-sports-visualization-with-pandas-matplotlib-and-seaborn-9c16df80a81b) provide useful insights into relationship between various datasets. + +## Summary + +Sports Analytics is a game changer when it comes to how professional games are played, especially how strategic decision making happens, which until recently was primarily done based on “gut feeling" or adherence to past traditions. NumPy forms a solid foundation for a large set of Python packages which provide higher level functions related to data analytics, machine learning, and AI algorithms. These packages are widely deployed to gain real-time insights that help in decision making for game-changing outcomes, both on field as well as to draw inferences and drive business around the game of cricket. Finding out the hidden parameters, patterns, and attributes that lead to the outcome of a cricket match helps the stakeholders to take notice of game insights that are otherwise hidden in numbers and statistics. + +{{< figure >}} +src = '/images/content_images/cs/numpy_ca_benefits.png' alt = 'Diagram showing benefits of using NumPy for cricket analytics' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} diff --git a/content/zh/case-studies/deeplabcut-dnn.md b/content/zh/case-studies/deeplabcut-dnn.md new file mode 100644 index 0000000000..257905de91 --- /dev/null +++ b/content/zh/case-studies/deeplabcut-dnn.md @@ -0,0 +1,102 @@ +--- +title: "Case Study: DeepLabCut 3D Pose Estimation" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/mice-hand.gif' title = 'Analyzing mice hand-movement using DeepLapCut' alt = 'micehandanim' attribution = '(Source: www.deeplabcut.org )' attributionlink = 'http://www.mousemotorlab.org/deeplabcut' +{{< /figure >}} + +{{< blockquote cite="https://news.harvard.edu/gazette/story/newsplus/harvard-researchers-awarded-czi-open-source-award/" by="Alexander Mathis, *Assistant Professor, École polytechnique fédérale de Lausanne* ([EPFL](https://www.epfl.ch/en/))" +> }} Open Source Software is accelerating Biomedicine. DeepLabCut enables automated video analysis of animal behavior using Deep Learning. +> +> {{< /blockquote >}} + +## About DeepLabCut + +[DeepLabCut](https://github.com/DeepLabCut/DeepLabCut) is an open source toolbox that empowers researchers at hundreds of institutions worldwide to track behaviour of laboratory animals, with very little training data, at human-level accuracy. With DeepLabCut technology, scientists can delve deeper into the scientific understanding of motor control and behavior across animal species and timescales. + +Several areas of research, including neuroscience, medicine, and biomechanics, use data from tracking animal movement. DeepLabCut helps in understanding what humans and other animals are doing by parsing actions that have been recorded on film. Using automation for laborious tasks of tagging and monitoring, along with deep neural network based data analysis, DeepLabCut makes scientific studies involving observing animals, such as primates, mice, fish, flies etc., much faster and more accurate. + +{{< figure >}} +src = '/images/content_images/cs/race-horse.gif' title = 'Colored dots track the positions of a racehorse’s body part' alt = 'horserideranim' attribution = '(Source: Mackenzie Mathis)' +{{< /figure >}} + +DeepLabCut's non-invasive behavioral tracking of animals by extracting the poses of animals is crucial for scientific pursuits in domains such as biomechanics, genetics, ethology & neuroscience. Measuring animal poses non-invasively from video - without markers - in dynamically changing backgrounds is computationally challenging, both technically as well as in terms of resource needs and training data required. + +DeepLabCut allows researchers to estimate the pose of the subject, efficiently enabling them to quantify the behavior through a Python based software toolkit. With DeepLabCut, researchers can identify distinct frames from videos, digitally label specific body parts in a few dozen frames with a tailored GUI, and then the deep learning based pose estimation architectures in DeepLabCut learn how to pick out those same features in the rest of the video and in other similar videos of animals. It works across species of animals, from common laboratory animals such as flies and mice to more unusual animals like [cheetahs][cheetah-movement]. + +DeepLabCut uses a principle called [transfer learning](https://arxiv.org/pdf/1909.11229), which greatly reduces the amount of training data required and speeds up the convergence of the training period. Depending on the needs, users can pick different network architectures that provide faster inference (e.g. MobileNetV2), which can also be combined with real-time experimental feedback. DeepLabCut originally used the feature detectors from a top-performing human pose estimation architecture, called [DeeperCut](https://arxiv.org/abs/1605.03170), which inspired the name. The package now has been significantly changed to include additional architectures, augmentation methods, and a full front-end user experience. Furthermore, to support large-scale biological experiments DeepLabCut provides active learning capabilities so that users can increase the training set over time to cover edge cases and make their pose estimation algorithm robust within the specific context. + +Recently, the [DeepLabCut model zoo](http://www.mousemotorlab.org/dlc-modelzoo) was introduced, which provides pre-trained models for various species and experimental conditions from facial analysis in primates to dog posture. This can be run for instance in the cloud without any labeling of new data, or neural network training, and no programming experience is necessary. + +### Key Goals and Results + +* **Automation of animal pose analysis for scientific studies:** + + The primary objective of DeepLabCut technology is to measure and track posture of animals in a diverse settings. This data can be used, for example, in neuroscience studies to understand how the brain controls movement, or to elucidate how animals socially interact. Researchers have observed a [tenfold performance boost](https://www.biorxiv.org/content/10.1101/457242v1) with DeepLabCut. Poses can be inferred offline at up to 1200 frames per second (FPS). + +* **Creation of an easy-to-use Python toolkit for pose estimation:** + + DeepLabCut wanted to share their animal pose-estimation technology in the form of an easy to use tool that can be adopted by researchers easily. So they have created a complete, easy-to-use Python toolbox with project management features as well. These enable not only automation of pose-estimation but also managing the project end-to-end by helping the DeepLabCut Toolkit user right from the dataset collection stage to creating shareable and reusable analysis pipelines. + + Their [toolkit][DLCToolkit] is now available as open source. + + A typical DeepLabCut Workflow includes: + + - creation and refining of training sets via active learning + - creation of tailored neural networks for specific animals and scenarios + - code for large-scale inference on videos + - draw inferences using integrated visualization tools + +{{< figure >}} +src = '/images/content_images/cs/deeplabcut-toolkit-steps.png' title = 'Pose estimation steps with DeepLabCut' alt = 'dlcsteps' align = 'center' attribution = '(Source: DeepLabCut)' attributionlink = 'https://twitter.com/DeepLabCut/status/1198046918284210176/photo/1' +{{< /figure >}} + +### The Challenges + +* **Speed** + + Fast processing of animal behavior videos in order to measure their behavior and at the same time make scientific experiments more efficient, accurate. Extracting detailed animal poses for laboratory experiments, without markers, in dynamically changing backgrounds, can be challenging, both technically as well as in terms of resource needs and training data required. Coming up with a tool that is easy to use without the need for skills such as computer vision expertise that enables scientists to do research in more real-world contexts, is a non-trivial problem to solve. + +* **Combinatorics** + + Combinatorics involves assembly and integration of movement of multiple limbs into individual animal behavior. Assembling keypoints and their connections into individual animal movements and linking them across time is a complex process that requires heavy-duty numerical analysis, especially in case of multi-animal movement tracking in experiment videos. + +* **Data Processing** + + Last but not the least, array manipulation - processing large stacks of arrays corresponding to various images, target tensors and keypoints is fairly challenging. + +{{< figure >}} +src = '/images/content_images/cs/pose-estimation.png' title = 'Pose estimation variety and complexity' alt = 'challengesfig' align = 'center' attribution = '(Source: Mackenzie Mathis)' attributionlink = 'https://www.biorxiv.org/content/10.1101/476531v1.full.pdf' +{{< /figure >}} + +## NumPy's Role in meeting Pose Estimation Challenges + +NumPy addresses DeepLabCut technology's core need of numerical computations at high speed for behavioural analytics. Besides NumPy, DeepLabCut employs various Python software that utilize NumPy at their core, such as [SciPy](https://www.scipy.org), [Pandas](https://pandas.pydata.org), [matplotlib](https://matplotlib.org), [Tensorpack](https://github.com/tensorpack/tensorpack), [imgaug](https://github.com/aleju/imgaug), [scikit-learn](https://scikit-learn.org/stable/), [scikit-image](https://scikit-image.org) and [Tensorflow](https://www.tensorflow.org). + +The following features of NumPy played a key role in addressing the image processing, combinatorics requirements and need for fast computation in DeepLabCut pose estimation algorithms: + +* Vectorization +* Masked Array Operations +* Linear Algebra +* Random Sampling +* Reshaping of large arrays + +DeepLabCut utilizes NumPy’s array capabilities throughout the workflow offered by the toolkit. In particular, NumPy is used for sampling distinct frames for human annotation labeling, and for writing, editing and processing annotation data. Within TensorFlow the neural network is trained by DeepLabCut technology over thousands of iterations to predict the ground truth annotations from frames. For this purpose, target densities (scoremaps) are created to cast pose estimation as a image-to-image translation problem. To make the neural networks robust, data augmentation is employed, which requires the calculation of target scoremaps subject to various geometric and image processing steps. To make training fast, NumPy’s vectorization capabilities are leveraged. For inference, the most likely predictions from target scoremaps need to extracted and one needs to efficiently “link predictions to assemble individual animals”. + +{{< figure >}} +src = '/images/content_images/cs/deeplabcut-workflow.png' title = 'DeepLabCut Workflow' alt = 'workflow' attribution = '(Source: Mackenzie Mathis)' attributionlink = 'https://www.researchgate.net/figure/DeepLabCut-work-flow-The-diagram-delineates-the-work-flow-as-well-as-the-directory-and_fig1_329185962' +{{< /figure >}} + +## Summary + +Observing and efficiently describing behavior is a core tenant of modern ethology, neuroscience, medicine, and technology. [DeepLabCut](http://orga.cvss.cc/wp-content/uploads/2019/05/NathMathis2019.pdf) allows researchers to estimate the pose of the subject, efficiently enabling them to quantify the behavior. With only a small set of training images, the DeepLabCut Python toolbox allows training a neural network to within human level labeling accuracy, thus expanding its application to not only behavior analysis in the laboratory, but to potentially also in sports, gait analysis, medicine and rehabilitation studies. Complex combinatorics, data processing challenges faced by DeepLabCut algorithms are addressed through the use of NumPy's array manipulation capabilities. + +{{< figure >}} +src = '/images/content_images/cs/numpy_dlc_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} + +[cheetah-movement]: https://www.technologynetworks.com/neuroscience/articles/interview-a-deeper-cut-into-behavior-with-mackenzie-mathis-327618 + +[DLCToolkit]: https://github.com/DeepLabCut/DeepLabCut diff --git a/content/zh/case-studies/gw-discov.md b/content/zh/case-studies/gw-discov.md new file mode 100644 index 0000000000..08f54b7547 --- /dev/null +++ b/content/zh/case-studies/gw-discov.md @@ -0,0 +1,77 @@ +--- +title: "Case Study: Discovery of Gravitational Waves" +sidebar: false +--- + +{{< figure >}} +src = '/images/content_images/cs/gw_sxs_image.png' title = 'Gravitational Waves' alt = 'binary coalesce black hole generating gravitational waves' attribution = '(Image Credits: The Simulating eXtreme Spacetimes (SXS) Project at LIGO)' attributionlink = 'https://youtu.be/Zt8Z_uzG71o' +{{< /figure >}} + +{{< blockquote cite="https://www.youtube.com/watch?v=BIvezCVcsYs" by="David Shoemaker, *LIGO Scientific Collaboration*" >}} The scientific Python ecosystem is critical infrastructure for the research done at LIGO. +{{< /blockquote >}} + +## About [Gravitational Waves](https://www.nationalgeographic.com/news/2017/10/what-are-gravitational-waves-ligo-astronomy-science/) and [LIGO](https://www.ligo.caltech.edu) + +Gravitational waves are ripples in the fabric of space and time, generated by cataclysmic events in the universe such as collision and merging of two black holes or coalescing binary stars or supernovae. Observing GW can not only help in studying gravity but also in understanding some of the obscure phenomena in the distant universe and its impact. + +The [Laser Interferometer Gravitational-Wave Observatory (LIGO)](https://www.ligo.caltech.edu) was designed to open the field of gravitational-wave astrophysics through the direct detection of gravitational waves predicted by Einstein’s General Theory of Relativity. It comprises two widely separated interferometers within the United States — one in Hanford, Washington and the other in Livingston, Louisiana — operated in unison to detect gravitational waves. Each of them has multi-kilometer-scale gravitational wave detectors that use laser interferometry. The LIGO Scientific Collaboration (LSC), is a group of more than 1000 scientists from universities around the United States and in 14 other countries supported by more than 90 universities and research institutes; approximately 250 students actively contributing to the collaboration. The new LIGO discovery is the first observation of gravitational waves themselves, made by measuring the tiny disturbances the waves make to space and time as they pass through the earth. It has opened up new astrophysical frontiers that explore the warped side of the universe—objects and phenomena that are made from warped spacetime. + + +### Key Objectives + +* Though its [mission](https://www.ligo.caltech.edu/page/what-is-ligo) is to detect gravitational waves from some of the most violent and energetic processes in the Universe, the data LIGO collects may have far-reaching effects on many areas of physics including gravitation, relativity, astrophysics, cosmology, particle physics, and nuclear physics. +* Crunch observed data via numerical relativity computations that involves complex maths in order to discern signal from noise, filter out relevant signal and statistically estimate significance of observed data +* Data visualization so that the binary / numerical results can be comprehended. + + + +### The Challenges + +* **Computation** + + Gravitational Waves are hard to detect as they produce a very small effect and have tiny interaction with matter. Processing and analyzing all of LIGO's data requires a vast computing infrastructure.After taking care of noise, which is billions of times of the signal, there is still very complex relativity equations and huge amounts of data which present a computational challenge: [O(10^7) CPU hrs needed for binary merger analyses](https://youtu.be/7mcHknWWzNI) spread on 6 dedicated LIGO clusters + +* **Data Deluge** + + As observational devices become more sensitive and reliable, the challenges posed by data deluge and finding a needle in a haystack rise multi-fold. LIGO generates terabytes of data every day! Making sense of this data requires an enormous effort for each and every detection. For example, the signals being collected by LIGO must be matched by supercomputers against hundreds of thousands of templates of possible gravitational-wave signatures. + +* **Visualization** + + Once the obstacles related to understanding Einstein’s equations well enough to solve them using supercomputers are taken care of, the next big challenge was making data comprehensible to the human brain. Simulation modeling as well as signal detection requires effective visualization techniques. Visualization also plays a role in lending more credibility to numerical relativity in the eyes of pure science aficionados, who did not give enough importance to numerical relativity until imaging and simulations made it easier to comprehend results for a larger audience. Speed of complex computations and rendering, re-rendering images and simulations using latest experimental inputs and insights can be a time consuming activity that challenges researchers in this domain. + +{{< figure >}} +src = '/images/content_images/cs/gw_strain_amplitude.png' alt = 'gravitational waves strain amplitude' title = 'Estimated gravitational-wave strain amplitude from GW150914' attribution = '(Graph Credits: Observation of Gravitational Waves from a Binary Black Hole Merger, ResearchGate Publication)' attributionlink = 'https://www.researchgate.net/publication/293886905_Observation_of_Gravitational_Waves_from_a_Binary_Black_Hole_Merger' +{{< /figure >}} + +## NumPy’s Role in the Detection of Gravitational Waves + +Gravitational waves emitted from the merger cannot be computed using any technique except brute force numerical relativity using supercomputers. The amount of data LIGO collects is as incomprehensibly large as gravitational wave signals are small. + +NumPy, the standard numerical analysis package for Python, was utilized by the software used for various tasks performed during the GW detection project at LIGO. NumPy helped in solving complex maths and data manipulation at high speed. Here are some examples: + +* [Signal Processing](https://www.uv.es/virgogroup/Denoising_ROF.html): Glitch detection, [Noise identification and Data Characterization](https://ep2016.europython.eu/media/conference/slides/pyhton-in-gravitational-waves-research-communities.pdf) (NumPy, scikit-learn, scipy, matplotlib, pandas, pyCharm) +* Data retrieval: Deciding which data can be analyzed, figuring out whether it contains a signal - needle in a haystack +* Statistical analysis: estimate the statistical significance of observational data, estimating the signal parameters (e.g. masses of stars, spin velocity, and distance) by comparison with a model. +* Visualization of data + - Time series + - Spectrograms +* Compute Correlations +* Key [Software](https://github.com/lscsoft) developed in GW data analysis such as [GwPy](https://gwpy.github.io/docs/stable/overview.html) and [PyCBC](https://pycbc.org) uses NumPy and AstroPy under the hood for providing object based interfaces to utilities, tools, and methods for studying data from gravitational-wave detectors. + +{{< figure >}} +src = '/images/content_images/cs/gwpy-numpy-dep-graph.png' alt = 'gwpy-numpy depgraph' title = 'Dependency graph showing how GwPy package depends on NumPy' +{{< /figure >}} + +---- + +{{< figure >}} +src = '/images/content_images/cs/PyCBC-numpy-dep-graph.png' alt = 'PyCBC-numpy depgraph' title = 'Dependency graph showing how PyCBC package depends on NumPy' +{{< /figure >}} + +## Summary + +GW detection has enabled researchers to discover entirely unexpected phenomena while providing new insight into many of the most profound astrophysical phenomena known. Number crunching and data visualization is a crucial step that helps scientists gain insights into data gathered from the scientific observations and understand the results. The computations are complex and cannot be comprehended by humans unless it is visualized using computer simulations that are fed with the real observed data and analysis. NumPy along with other Python packages such as matplotlib, pandas, and scikit-learn is [enabling researchers](https://www.gw-openscience.org/events/GW150914/) to answer complex questions and discover new horizons in our understanding of the universe. + +{{< figure >}} +src = '/images/content_images/cs/numpy_gw_benefits.png' alt = 'numpy benefits' title = 'Key NumPy Capabilities utilized' +{{< /figure >}} diff --git a/content/zh/citing-numpy.md b/content/zh/citing-numpy.md new file mode 100644 index 0000000000..5bb5d791b4 --- /dev/null +++ b/content/zh/citing-numpy.md @@ -0,0 +1,35 @@ +--- +title: Citing NumPy +sidebar: false +--- + +If NumPy has been significant in your research, and you would like to acknowledge the project in your academic publication, we suggest citing the following paper: + +* Harris, C.R., Millman, K.J., van der Walt, S.J. et al. _Array programming with NumPy_. Nature 585, 357–362 (2020). DOI: [10.1038/s41586-020-2649-2](https://doi.org/10.1038/s41586-020-2649-2). ([Publisher link](https://www.nature.com/articles/s41586-020-2649-2)). + +_In BibTeX format:_ + + ``` +@Article{ harris2020array, + title = {Array programming with {NumPy}}, + author = {Charles R. Harris and K. Jarrod Millman and St{\'{e}}fan J. + van der Walt and Ralf Gommers and Pauli Virtanen and David + Cournapeau and Eric Wieser and Julian Taylor and Sebastian + Berg and Nathaniel J. Smith and Robert Kern and Matti Picus + and Stephan Hoyer and Marten H. van Kerkwijk and Matthew + Brett and Allan Haldane and Jaime Fern{\'{a}}ndez del + R{\'{i}}o and Mark Wiebe and Pearu Peterson and Pierre + G{\'{e}}rard-Marchant and Kevin Sheppard and Tyler Reddy and + Warren Weckesser and Hameer Abbasi and Christoph Gohlke and + Travis E. Oliphant}, + year = {2020}, + month = sep, + journal = {Nature}, + volume = {585}, + number = {7825}, + pages = {357--362}, + doi = {10.1038/s41586-020-2649-2}, + publisher = {Springer Science and Business Media {LLC}}, + url = {https://doi.org/10.1038/s41586-020-2649-2} +} +``` diff --git a/content/zh/code-of-conduct.md b/content/zh/code-of-conduct.md new file mode 100644 index 0000000000..5ee1f4bcbe --- /dev/null +++ b/content/zh/code-of-conduct.md @@ -0,0 +1,83 @@ +--- +title: NumPy Code of Conduct +sidebar: false +aliases: + - /conduct.html +--- + +### Introduction + +This Code of Conduct applies to all spaces managed by the NumPy project, including all public and private mailing lists, issue trackers, wikis, blogs, Twitter, and any other communication channel used by our community. The NumPy project does not organise in-person events, however events related to our community should have a code of conduct similar in spirit to this one. + +This Code of Conduct should be honored by everyone who participates in the NumPy community formally or informally, or claims any affiliation with the project, in any project-related activities and especially when representing the project, in any role. + +This code is not exhaustive or complete. It serves to distill our common understanding of a collaborative, shared environment and goals. Please try to follow this code in spirit as much as in letter, to create a friendly and productive environment that enriches the surrounding community. + +### Specific Guidelines + +We strive to: + +1. Be open. We invite anyone to participate in our community. We prefer to use public methods of communication for project-related messages, unless discussing something sensitive. This applies to messages for help or project-related support, too; not only is a public support request much more likely to result in an answer to a question, it also ensures that any inadvertent mistakes in answering are more easily detected and corrected. +2. Be empathetic, welcoming, friendly, and patient. We work together to resolve conflict, and assume good intentions. We may all experience some frustration from time to time, but we do not allow frustration to turn into a personal attack. A community where people feel uncomfortable or threatened is not a productive one. +3. Be collaborative. Our work will be used by other people, and in turn we will depend on the work of others. When we make something for the benefit of the project, we are willing to explain to others how it works, so that they can build on the work to make it even better. Any decision we make will affect users and colleagues, and we take those consequences seriously when making decisions. +4. Be inquisitive. Nobody knows everything! Asking questions early avoids many problems later, so we encourage questions, although we may direct them to the appropriate forum. We will try hard to be responsive and helpful. +5. Be careful in the words that we choose. We are careful and respectful in our communication, and we take responsibility for our own speech. Be kind to others. Do not insult or put down other participants. We will not accept harassment or other exclusionary behaviour, such as: + * Violent threats or language directed against another person. + * Sexist, racist, or otherwise discriminatory jokes and language. + * Posting sexually explicit or violent material. + * Posting (or threatening to post) other people’s personally identifying information (“doxing”). + * Sharing private content, such as emails sent privately or non-publicly, or unlogged forums such as IRC channel history, without the sender’s consent. + * Personal insults, especially those using racist or sexist terms. + * Unwelcome sexual attention. + * Excessive profanity. Please avoid swearwords; people differ greatly in their sensitivity to swearing. + * Repeated harassment of others. In general, if someone asks you to stop, then stop. + * Advocating for, or encouraging, any of the above behaviour. + +### Diversity Statement + +The NumPy project welcomes and encourages participation by everyone. We are committed to being a community that everyone enjoys being part of. Although we may not always be able to accommodate each individual’s preferences, we try our best to treat everyone kindly. + +No matter how you identify yourself or how others perceive you: we welcome you. Though no list can hope to be comprehensive, we explicitly honour diversity in: age, culture, ethnicity, genotype, gender identity or expression, language, national origin, neurotype, phenotype, political beliefs, profession, race, religion, sexual orientation, socioeconomic status, subculture and technical ability, to the extent that these do not conflict with this code of conduct. + +Though we welcome people fluent in all languages, NumPy development is conducted in English. + +Standards for behaviour in the NumPy community are detailed in the Code of Conduct above. Participants in our community should uphold these standards in all their interactions and help others to do so as well (see next section). + +### Reporting Guidelines + +We know that it is painfully common for internet communication to start at or devolve into obvious and flagrant abuse. We also recognize that sometimes people may have a bad day, or be unaware of some of the guidelines in this Code of Conduct. Please keep this in mind when deciding on how to respond to a breach of this Code. + +For clearly intentional breaches, report those to the Code of Conduct Committee (see below). For possibly unintentional breaches, you may reply to the person and point out this code of conduct (either in public or in private, whatever is most appropriate). If you would prefer not to do that, please feel free to report to the Code of Conduct Committee directly, or ask the Committee for advice, in confidence. + +You can report issues to the NumPy Code of Conduct Committee at numpy-conduct@googlegroups.com. + +Currently, the Committee consists of: + +* Stefan van der Walt +* Melissa Weber Mendonça +* Rohit Goswami + +If your report involves any members of the Committee, or if they feel they have a conflict of interest in handling it, then they will recuse themselves from considering your report. Alternatively, if for any reason you feel uncomfortable making a report to the Committee, then you can also contact senior NumFOCUS staff at [conduct@numfocus.org](https://numfocus.org/code-of-conduct#persons-responsible). + +### Incident reporting resolution & Code of Conduct enforcement + +_This section summarizes the most important points, more details can be found in_ [NumPy Code of Conduct - How to follow up on a report](report-handling-manual). + +We will investigate and respond to all complaints. The NumPy Code of Conduct Committee and the NumPy Steering Committee (if involved) will protect the identity of the reporter, and treat the content of complaints as confidential (unless the reporter agrees otherwise). + +In case of severe and obvious breaches, e.g. personal threat or violent, sexist or racist language, we will immediately disconnect the originator from NumPy communication channels; please see the manual for details. + +In cases not involving clear severe and obvious breaches of this Code of Conduct the process for acting on any received Code of Conduct violation report will be: + +1. acknowledge report is received, +2. reasonable discussion/feedback, +3. mediation (if feedback didn’t help, and only if both reporter and reportee agree to this), +4. enforcement via transparent decision (see [Resolutions](report-handling-manual/#resolutions)) by the Code of Conduct Committee. + +The Committee will respond to any report as soon as possible, and at most within 72 hours. + +### Endnotes + +We are thankful to the groups behind the following documents, from which we drew content and inspiration: + +- [The SciPy Code of Conduct](https://docs.scipy.org/doc/scipy/dev/conduct/code_of_conduct.html) diff --git a/content/zh/community.md b/content/zh/community.md new file mode 100644 index 0000000000..5034fba239 --- /dev/null +++ b/content/zh/community.md @@ -0,0 +1,66 @@ +--- +title: Community +sidebar: false +--- + +NumPy is a community-driven open source project developed by a diverse group of [contributors](/teams/). The NumPy leadership has made a strong commitment to creating an open, inclusive, and positive community. Please read the [NumPy Code of Conduct](/code-of-conduct) for guidance on how to interact with others in a way that makes the community thrive. + +We offer several communication channels to learn, share your knowledge and connect with others within the NumPy community. + + +## Participate online + +The following are ways to engage directly with the NumPy project and community. _Please note that we encourage users and community members to support each other for usage questions - see [Get Help](/gethelp)._ + + +### [NumPy mailing list](https://mail.python.org/mailman/listinfo/numpy-discussion) + +This list is the main forum for longer-form discussions, like adding new features to NumPy, making changes to the NumPy Roadmap, and all kinds of project-wide decision making. Announcements about NumPy, such as for releases, developer meetings, sprints or conference talks are also made on this list. + +On this list please use bottom posting, reply to the list (rather than to another sender), and don't reply to digests. A searchable archive of this list is available [here](https://mail.python.org/archives/list/numpy-discussion@python.org/). + +*** + +### [GitHub issue tracker](https://github.com/numpy/numpy/issues) + +- For bug reports (e.g. "`np.arange(3).shape` returns `(5,)`, when it should return `(3,)`"); +- documentation issues (e.g. "I found this section unclear"); +- and feature requests (e.g. "I would like to have a new interpolation method in `np.percentile`"). + +_Please note that GitHub is not the right place to report a security vulnerability. If you think you have found a security vulnerability in NumPy, please report it [here](https://tidelift.com/docs/security)._ + +*** + +### [Slack](https://numpy-team.slack.com) + +A real-time chat room to ask questions about _contributing_ to NumPy. This is a private space, specifically meant for people who are hesitant to bring up their questions or ideas on a large public mailing list or GitHub. Please see [here](https://numpy.org/devdocs/dev/index.html#contributing-to-numpy) for more details and how to get an invite. + + +## Study Groups and Meetups + +If you would like to find a local meetup or study group to learn more about NumPy and the wider ecosystem of Python packages for data science and scientific computing, we recommend exploring the [PyData meetups](https://www.meetup.com/pro/pydata/) (150+ meetups, 100,000+ members). + +NumPy also organizes in-person sprints for its team and interested contributors occasionally. These are typically planned several months in advance and will be announced on the [mailing list](https://mail.python.org/mailman/listinfo/numpy-discussion) and [Twitter](https://twitter.com/numpy_team). + + +## Conferences + +The NumPy project doesn't organize its own conferences. The conferences that have traditionally been most popular with NumPy maintainers, contributors and users are the SciPy and PyData conference series: + +- [SciPy US](https://conference.scipy.org) +- [EuroSciPy](https://www.euroscipy.org) +- [SciPy Latin America](https://www.scipyla.org) +- [SciPy India](https://scipy.in) +- [SciPy Japan](https://conference.scipy.org) +- [PyData conferences](https://pydata.org/event-schedule/) (15-20 events a year spread over many countries) + +Many of these conferences include tutorial days that cover NumPy and/or sprints where you can learn how to contribute to NumPy or related open source projects. + + +## Join the NumPy community + +To thrive, the NumPy project needs your expertise and enthusiasm. Not a coder? Not a problem! There are many ways to contribute to NumPy. + +If you are interested in becoming a NumPy contributor (yay!) we recommend checking out our [Contribute](/contribute) page. + +Also, feel free to stop by and say hi at one of our community meetings. To keep track of them, check out our events calendar [here](https://scientific-python.org/calendars/). diff --git a/content/zh/config.yaml b/content/zh/config.yaml new file mode 100644 index 0000000000..4aaf75b2e6 --- /dev/null +++ b/content/zh/config.yaml @@ -0,0 +1,140 @@ +languageName: English +params: + description: Why NumPy? Powerful n-dimensional arrays. Numerical computing tools. Interoperable. Performant. Open source. + navbarlogo: + image: logo.svg + text: NumPy + link: / + hero: + #Main hero title + title: NumPy + #Hero subtitle (optional) + subtitle: The fundamental package for scientific computing with Python + #Button text + buttontext: "Latest release: NumPy 1.26. View all releases" + #Where the main hero button links to + buttonlink: "/news/#releases" + #Hero image (from static/images/___) + image: logo.svg + shell: + title: placeholder + intro: + - + title: Try NumPy + text: Use the interactive shell to try NumPy in the browser + docslink: Don't forget to check out the docs. + casestudies: + title: CASE STUDIES + features: + - + title: First Image of a Black Hole + text: How NumPy, together with libraries like SciPy and Matplotlib that depend on NumPy, enabled the Event Horizon Telescope to produce the first ever image of a black hole + img: /images/content_images/case_studies/blackhole.png + alttext: First image of a black hole. It is an orange circle in a black background. + url: /case-studies/blackhole-image + - + title: Detection of Gravitational Waves + text: In 1916, Albert Einstein predicted gravitational waves; 100 years later their existence was confirmed by LIGO scientists using NumPy. + img: /images/content_images/case_studies/gravitional.png + alttext: Two orbs orbiting each other. They are displacing gravity around them. + url: /case-studies/gw-discov + - + title: Sports Analytics + text: Cricket Analytics is changing the game by improving player and team performance through statistical modelling and predictive analytics. NumPy enables many of these analyses. + img: /images/content_images/case_studies/sports.jpg + alttext: Cricket ball on green field. + url: /case-studies/cricket-analytics + - + title: Pose Estimation using deep learning + text: DeepLabCut uses NumPy for accelerating scientific studies that involve observing animal behavior for better understanding of motor control, across species and timescales. + img: /images/content_images/case_studies/deeplabcut.png + alttext: Cheetah pose analysis + url: /case-studies/deeplabcut-dnn + tabs: + title: ECOSYSTEM + section5: false + navbar: + - + title: Install + url: /install + - + title: Documentation + url: https://numpy.org/doc/stable + - + title: Learn + url: /learn + - + title: Community + url: /community + - + title: About Us + url: /about + - + title: News + url: /news + - + title: Contribute + url: /contribute + footer: + logo: logo.svg + socialmediatitle: "" + socialmedia: + - + link: https://github.com/numpy/numpy + icon: github + - + link: https://www.youtube.com/channel/UCguIL9NZ7ybWK5WQ53qbHng + icon: youtube + - + link: https://twitter.com/numpy_team + icon: twitter + quicklinks: + column1: + title: "" + links: + - + text: Install + link: /install + - + text: Documentation + link: https://numpy.org/doc/stable + - + text: Learn + link: /learn + - + text: Citing Numpy + link: /citing-numpy + - + text: Roadmap + link: https://numpy.org/neps/roadmap.html + column2: + links: + - + text: About us + link: /about + - + text: Community + link: /community + - + text: User surveys + link: /user-surveys + - + text: Contribute + link: /contribute + - + text: Code of conduct + link: /code-of-conduct + column3: + links: + - + text: Get help + link: /gethelp + - + text: Terms of use + link: /terms + - + text: Privacy + link: /privacy + - + text: Press kit + link: /press-kit diff --git a/content/zh/contribute.md b/content/zh/contribute.md new file mode 100644 index 0000000000..6efff53624 --- /dev/null +++ b/content/zh/contribute.md @@ -0,0 +1,66 @@ +--- +title: Contribute to NumPy +sidebar: false +--- + +The NumPy project welcomes your expertise and enthusiasm! Your choices aren't limited to programming, as you can see below there are many areas where we need **your** help. + +If you're unsure where to start or how your skills fit in, _reach out!_ You can ask on the [mailing list](https://mail.python.org/mailman/listinfo/numpy-discussion) or [GitHub](http://github.com/numpy/numpy) (open an [issue](https://github.com/numpy/numpy/issues) or comment on a relevant issue). + +Those are our preferred channels (open source is open by nature), but if you prefer to talk privately, contact our community coordinators at or on [Slack](https://numpy-team.slack.com) (write for an invite). + +We also have a biweekly _community call_, details of which are announced on the [mailing list](https://mail.python.org/mailman/listinfo/numpy-discussion). You are very welcome to join. If you are new to contributing to open source, we also highly recommend reading [this guide](https://opensource.guide/how-to-contribute/). + +Our community aspires to treat everyone equally and to value all contributions. We have a [Code of Conduct](/code-of-conduct) to foster an open and welcoming environment. + +### Writing code + +Programmers, this [guide](https://numpy.org/devdocs/dev/index.html#development-process-summary) explains how to contribute to the NumPy codebase.
Check out also our [YouTube channel](https://www.youtube.com/playlist?list=PLCK6zCrcN3GXBUUzDr9L4__LnXZVtaIzS) for additional advice. + + +### Reviewing pull requests +The project has more than 250 open pull requests -- meaning many potential improvements and many open-source contributors waiting for feedback. If you're a developer who knows NumPy, you can help even if you're not familiar with the codebase. You can: +* summarize a long-running discussion +* triage documentation PRs +* test proposed changes + + +### Developing educational materials + +NumPy's [User Guide](https://numpy.org/devdocs) is undergoing rehabilitation. We're in need of new tutorials, how-to's, and deep-dive explanations, and the site needs restructuring. Opportunities aren't limited to writers. We'd also welcome worked examples, notebooks, and videos. [NEP 44 — Restructuring the NumPyDocumentation](https://numpy.org/neps/nep-0044-restructuring-numpy-docs.html) lays out our ideas -- and you may have others. + + +### Issue triaging + +The [NumPy issue tracker](https://github.com/numpy/numpy/issues) has a _lot_ of open issues. Some are no longer valid, some should be prioritized, and some would make good issues for new contributors. You can: + +* check if older bugs are still present +* find duplicate issues and link related ones +* add good self-contained reproducers to issues +* label issues correctly (this requires triage rights -- just ask) + +Please just dive in. + + +### Website development + +We've just revamped our website, but we're far from done. If you love web development, these [issues](https://github.com/numpy/numpy.org/issues?q=is%3Aissue+is%3Aopen+label%3Adesign) list some of our unmet needs -- and feel free to share your own ideas. + + +### Graphic design + +We can barely begin to list the contributions a graphic designer can make here. Our docs are parched for illustration; our growing website craves images -- opportunities abound. + + +### Translating website content + +We plan multiple translations of [numpy.org](https://numpy.org) to make NumPy accessible to users in their native language. Volunteer translators are at the heart of this effort. See [here](https://numpy.org/neps/nep-0028-website-redesign.html#translation-multilingual-i18n) for background; comment on [this GitHub issue](https://github.com/numpy/numpy.org/issues/55) to sign up. + + +### Community coordination and outreach + +Through community contact we share our work more widely and learn where we're falling short. We're eager to get more people involved in efforts like our [Twitter](https://twitter.com/numpy_team) account, organizing NumPy [code sprints](https://scisprints.github.io/), a newsletter, and perhaps a blog. + +### Fundraising + +NumPy was all-volunteer for many years, but as its importance grew it became clear that to ensure stability and growth we'd need financial support. [This SciPy'19 talk](https://www.youtube.com/watch?v=dBTJD_FDVjU) explains how much difference that support has made. Like all the nonprofit world, we're constantly searching for grants, sponsorships, and other kinds of support. We have a number of ideas and of course we welcome more. Fundraising is a scarce skill here -- we'd appreciate your help. diff --git a/content/zh/gethelp.md b/content/zh/gethelp.md new file mode 100644 index 0000000000..da3bff67a5 --- /dev/null +++ b/content/zh/gethelp.md @@ -0,0 +1,20 @@ +--- +title: Get Help +sidebar: false +--- + +**Development issues:** For NumPy development-related matters (e.g., bug reports), please see [Community](/community). + +**User questions:** The best way to get help is to post your question to a site like [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) or [Reddit](https://www.reddit.com/r/Numpy/). We wish we could keep an eye on these sites, or answer questions directly, but the volume is a little overwhelming! + +### [StackOverflow](http://stackoverflow.com/questions/tagged/numpy) + +A forum for asking usage questions, e.g. "How do I do X in NumPy?”. Please [use the `#numpy` tag](https://stackoverflow.com/help/tagging) + +*** + +### [Reddit](https://www.reddit.com/r/Numpy/) + +Another forum for usage questions. + +*** diff --git a/content/zh/history.md b/content/zh/history.md new file mode 100644 index 0000000000..aa669d375b --- /dev/null +++ b/content/zh/history.md @@ -0,0 +1,21 @@ +--- +title: History of NumPy +sidebar: false +--- + +NumPy is a foundational Python library that provides array data structures and related fast numerical routines. When started, the library had little funding, and was written mainly by graduate students—many of them without computer science education, and often without a blessing of their advisors. To even imagine that a small group of “rogue” student programmers could upend the already well-established ecosystem of research software—backed by millions in funding and many hundreds of highly qualified engineers — was preposterous. Yet, the philosophical motivations behind a fully open tool stack, in combination with the excited, friendly community with a singular focus, have proven auspicious in the long run. Nowadays, NumPy is relied upon by scientists, engineers, and many other professionals around the world. For example, the published scripts used in the analysis of gravitational waves import NumPy, and the M87 black hole imaging project directly cites NumPy. + +For the in-depth account on milestones in the development of NumPy and related libraries please see [arxiv.org](https://arxiv.org/abs/1907.10121). + +If you’d like to obtain a copy of the original Numeric and Numarray libraries, follow the links below: + +[Download Page for *Numeric*](https://sourceforge.net/projects/numpy/files/Old%20Numeric/)* + +[Download Page for *Numarray*](https://sourceforge.net/projects/numpy/files/Old%20Numarray/)* + +*Please note that these older array packages are no longer maintained, and users are strongly advised to use NumPy for any array-related purposes or refactor any pre-existing code to utilize the NumPy library. + +### Historic Documentation + +[Download *`Numeric'* Manual](static/numeric-manual.pdf) + diff --git a/content/zh/install.md b/content/zh/install.md new file mode 100644 index 0000000000..f64a17841d --- /dev/null +++ b/content/zh/install.md @@ -0,0 +1,126 @@ +--- +title: Installing NumPy +sidebar: false +--- + +The only prerequisite for installing NumPy is Python itself. If you don't have Python yet and want the simplest way to get started, we recommend you use the [Anaconda Distribution](https://www.anaconda.com/download) - it includes Python, NumPy, and many other commonly used packages for scientific computing and data science. + +NumPy can be installed with `conda`, with `pip`, with a package manager on macOS and Linux, or [from source](https://numpy.org/devdocs/building). For more detailed instructions, consult our [Python and NumPy installation guide](#python-numpy-install-guide) below. + +**CONDA** + +If you use `conda`, you can install NumPy from the `defaults` or `conda-forge` channels: + +```bash +# Best practice, use an environment rather than install in the base env +conda create -n my-env +conda activate my-env +# If you want to install from conda-forge +conda config --env --add channels conda-forge +# The actual install command +conda install numpy +``` + +**PIP** + +If you use `pip`, you can install NumPy with: + +```bash +pip install numpy +``` +Also when using pip, it's good practice to use a virtual environment - see [Reproducible Installs](#reproducible-installs) below for why, and [this guide](https://dev.to/bowmanjd/python-tools-for-managing-virtual-environments-3bko#howto) for details on using virtual environments. + + + + +# Python and NumPy installation guide + +Installing and managing packages in Python is complicated, there are a number of alternative solutions for most tasks. This guide tries to give the reader a sense of the best (or most popular) solutions, and give clear recommendations. It focuses on users of Python, NumPy, and the PyData (or numerical computing) stack on common operating systems and hardware. + +## Recommendations + +We'll start with recommendations based on the user's experience level and operating system of interest. If you're in between "beginning" and "advanced", please go with "beginning" if you want to keep things simple, and with "advanced" if you want to work according to best practices that go a longer way in the future. + +### Beginning users + +On all of Windows, macOS, and Linux: + +- Install [Anaconda](https://www.anaconda.com/download) (it installs all packages you need and all other tools mentioned below). +- For writing and executing code, use notebooks in [JupyterLab](https://jupyterlab.readthedocs.io/en/stable/index.html) for exploratory and interactive computing, and [Spyder](https://www.spyder-ide.org/) or [Visual Studio Code](https://code.visualstudio.com/) for writing scripts and packages. +- Use [Anaconda Navigator](https://docs.anaconda.com/anaconda/navigator/) to manage your packages and start JupyterLab, Spyder, or Visual Studio Code. + + +### Advanced users + +#### Conda + +- Install [Miniforge](https://github.com/conda-forge/miniforge). +- Keep the `base` conda environment minimal, and use one or more [conda environments](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html) to install the package you need for the task or project you're working on. + +#### Alternative if you prefer pip/PyPI + +For users who know, from personal preference or reading about the main differences between conda and pip below, they prefer a pip/PyPI-based solution, we recommend: +- Install Python from [python.org](https://www.python.org/downloads/), [Homebrew](https://brew.sh/), or your Linux package manager. +- Use [Poetry](https://python-poetry.org/) as the most well-maintained tool that provides a dependency resolver and environment management capabilities in a similar fashion as conda does. + + +## Python package management + +Managing packages is a challenging problem, and, as a result, there are lots of tools. For web and general purpose Python development there's a whole [host of tools](https://packaging.python.org/guides/tool-recommendations/) complementary with pip. For high-performance computing (HPC), [Spack](https://github.com/spack/spack) is worth considering. For most NumPy users though, [conda](https://conda.io/en/latest/) and [pip](https://pip.pypa.io/en/stable/) are the two most popular tools. + + +### Pip & conda + +The two main tools that install Python packages are `pip` and `conda`. Their functionality partially overlaps (e.g. both can install `numpy`), however, they can also work together. We'll discuss the major differences between pip and conda here - this is important to understand if you want to manage packages effectively. + +The first difference is that conda is cross-language and it can install Python, while pip is installed for a particular Python on your system and installs other packages to that same Python install only. This also means conda can install non-Python libraries and tools you may need (e.g. compilers, CUDA, HDF5), while pip can't. + +The second difference is that pip installs from the Python Packaging Index (PyPI), while conda installs from its own channels (typically "defaults" or "conda-forge"). PyPI is the largest collection of packages by far, however, all popular packages are available for conda as well. + +The third difference is that conda is an integrated solution for managing packages, dependencies and environments, while with pip you may need another tool (there are many!) for dealing with environments or complex dependencies. + + + +### Reproducible installs + +As libraries get updated, results from running your code can change, or your code can break completely. It's important to be able to reconstruct the set of packages and versions you're using. Best practice is to: + +1. use a different environment per project you're working on, +2. record package names and versions using your package installer; each has its own metadata format for this: + - Conda: [conda environments and environment.yml](https://docs.conda.io/projects/conda/en/latest/user-guide/tasks/manage-environments.html) + - Pip: [virtual environments](https://docs.python.org/3/tutorial/venv.html) and [requirements.txt](https://pip.readthedocs.io/en/latest/user_guide/#requirements-files) + - Poetry: [virtual environments and pyproject.toml](https://python-poetry.org/docs/basic-usage/) + + + +## NumPy packages & accelerated linear algebra libraries + +NumPy doesn't depend on any other Python packages, however, it does depend on an accelerated linear algebra library - typically [Intel MKL](https://software.intel.com/en-us/mkl) or [OpenBLAS](https://www.openblas.net/). Users don't have to worry about installing those (they're automatically included in all NumPy install methods). Power users may still want to know the details, because the used BLAS can affect performance, behavior and size on disk: + +- The NumPy wheels on PyPI, which is what pip installs, are built with OpenBLAS. The OpenBLAS libraries are included in the wheel. This makes the wheel larger, and if a user installs (for example) SciPy as well, they will now have two copies of OpenBLAS on disk. + +- In the conda defaults channel, NumPy is built against Intel MKL. MKL is a separate package that will be installed in the users' environment when they install NumPy. + +- In the conda-forge channel, NumPy is built against a dummy "BLAS" package. When a user installs NumPy from conda-forge, that BLAS package then gets installed together with the actual library - this defaults to OpenBLAS, but it can also be MKL (from the defaults channel), or even [BLIS](https://github.com/flame/blis) or reference BLAS. + +- The MKL package is a lot larger than OpenBLAS, it's about 700 MB on disk while OpenBLAS is about 30 MB. + +- MKL is typically a little faster and more robust than OpenBLAS. + +Besides install sizes, performance and robustness, there are two more things to consider: + +- Intel MKL is not open source. For normal use this is not a problem, but if a user needs to redistribute an application built with NumPy, this could be an issue. +- Both MKL and OpenBLAS will use multi-threading for function calls like `np.dot`, with the number of threads being determined by both a build-time option and an environment variable. Often all CPU cores will be used. This is sometimes unexpected for users; NumPy itself doesn't auto-parallelize any function calls. It typically yields better performance, but can also be harmful - for example when using another level of parallelization with Dask, scikit-learn or multiprocessing. + + +## Troubleshooting + +If your installation fails with the message below, see [Troubleshooting ImportError](https://numpy.org/doc/stable/user/troubleshooting-importerror.html). + +``` +IMPORTANT: PLEASE READ THIS FOR ADVICE ON HOW TO SOLVE THIS ISSUE! + +Importing the numpy c-extensions failed. This error can happen for +different reasons, often due to issues with your setup. +``` + diff --git a/content/zh/learn.md b/content/zh/learn.md new file mode 100644 index 0000000000..4f9fa53ae3 --- /dev/null +++ b/content/zh/learn.md @@ -0,0 +1,76 @@ +--- +title: Learn +sidebar: false +--- + +For the **official NumPy documentation** visit [numpy.org/doc/stable](https://numpy.org/doc/stable). + +*** + +Below is a curated collection of educational resources, both for self-learning and teaching others, developed by NumPy contributors and vetted by the community. + +## Beginners + +There's a ton of information about NumPy out there. If you are just starting, we'd strongly recommend the following: + + **Tutorials** + +* [NumPy Quickstart Tutorial](https://numpy.org/devdocs/user/quickstart.html) +* [NumPy Tutorials](https://numpy.org/numpy-tutorials) A collection of tutorials and educational materials in the format of Jupyter Notebooks developed and maintained by the NumPy Documentation team. To submit your own content, visit the [numpy-tutorials repository on GitHub](https://github.com/numpy/numpy-tutorials). +* [NumPy Illustrated: The Visual Guide to NumPy *by Lev Maximov*](https://betterprogramming.pub/3b1d4976de1d?sk=57b908a77aa44075a49293fa1631dd9b) +* [Scientific Python Lectures](https://lectures.scientific-python.org/) Besides covering NumPy, these lectures offer a broader introduction to the scientific Python ecosystem. +* [NumPy: the absolute basics for beginners](https://numpy.org/devdocs/user/absolute_beginners.html) +* [NumPy tutorial *by Nicolas Rougier*](https://github.com/rougier/numpy-tutorial) +* [Stanford CS231 *by Justin Johnson*](http://cs231n.github.io/python-numpy-tutorial/) +* [NumPy User Guide](https://numpy.org/devdocs) + + **Books** + +* [Guide to NumPy *by Travis E. Oliphant*](https://web.mit.edu/dvp/Public/numpybook.pdf) This is a free version 1 from 2006. For the latest copy (2015) see [here](https://www.barnesandnoble.com/w/guide-to-numpy-travis-e-oliphant-phd/1144670472). +* [From Python to NumPy *by Nicolas P. Rougier*](https://www.labri.fr/perso/nrougier/from-python-to-numpy/) +* [Elegant SciPy](https://www.amazon.com/Elegant-SciPy-Art-Scientific-Python/dp/1491922877) *by Juan Nunez-Iglesias, Stefan van der Walt, and Harriet Dashnow* + +You may also want to check out the [Goodreads list](https://www.goodreads.com/shelf/show/python-scipy) on the subject of "Python+SciPy." Most books there are about the "SciPy ecosystem," which has NumPy at its core. + + **Videos** + +* [Introduction to Numerical Computing with NumPy](http://youtu.be/ZB7BZMhfPgk) *by Alex Chabot-Leclerc* + +*** + +## Advanced + +Try these advanced resources for a better understanding of NumPy concepts like advanced indexing, splitting, stacking, linear algebra, and more. + + **Tutorials** + +* [100 NumPy Exercises](http://www.labri.fr/perso/nrougier/teaching/numpy.100/index.html) *by Nicolas P. Rougier* +* [An Introduction to NumPy and Scipy](https://engineering.ucsb.edu/~shell/che210d/numpy.pdf) *by M. Scott Shell* +* [Numpy Medkits](http://mentat.za.net/numpy/numpy_advanced_slides/) *by Stéfan van der Walt* +* [NumPy Tutorials](https://numpy.org/numpy-tutorials) A collection of tutorials and educational materials in the format of Jupyter Notebooks developed and maintained by the NumPy Documentation team. To submit your own content, visit the [numpy-tutorials repository on GitHub](https://github.com/numpy/numpy-tutorials). + + **Books** + +* [Python Data Science Handbook](https://www.amazon.com/Python-Data-Science-Handbook-Essential/dp/1098121228) *by Jake Vanderplas* +* [Python for Data Analysis](https://www.amazon.com/Python-Data-Analysis-Wrangling-Jupyter/dp/109810403X) *by Wes McKinney* +* [Numerical Python: Scientific Computing and Data Science Applications with Numpy, SciPy, and Matplotlib](https://www.amazon.com/Numerical-Python-Scientific-Applications-Matplotlib/dp/1484242459) *by Robert Johansson* + + **Videos** + +* [Advanced NumPy - broadcasting rules, strides, and advanced indexing](https://www.youtube.com/watch?v=cYugp9IN1-Q) *by Juan Nunez-Iglesias* + +*** + +## NumPy Talks + +* [The Future of NumPy Indexing](https://www.youtube.com/watch?v=o0EacbIbf58) *by Jaime Fernández* (2016) +* [Evolution of Array Computing in Python](https://www.youtube.com/watch?v=HVLPJnvInzM&t=10s) *by Ralf Gommers* (2019) +* [NumPy: what has changed and what is going to change?](https://www.youtube.com/watch?v=YFLVQFjRmPY) *by Matti Picus* (2019) +* [Inside NumPy](https://www.youtube.com/watch?v=dBTJD_FDVjU) *by Ralf Gommers, Sebastian Berg, Matti Picus, Tyler Reddy, Stefan van der Walt, Charles Harris* (2019) +* [Brief Review of Array Computing in Python](https://www.youtube.com/watch?v=f176j2g2eNc) *by Travis Oliphant* (2019) + +*** + +## Citing NumPy + +If NumPy has been significant in your research, and you would like to acknowledge the project in your academic publication, please see [this citation information](/citing-numpy). diff --git a/content/zh/news.md b/content/zh/news.md new file mode 100644 index 0000000000..793619c0d1 --- /dev/null +++ b/content/zh/news.md @@ -0,0 +1,279 @@ +--- +title: News +sidebar: false +newsHeader: "NumPy 2.0 release date: June 16" +date: 2024-05-23 +--- + +### NumPy 2.0 release date: June 16 + +_23 May, 2024_ -- We are excited to announce that NumPy 2.0 is planned to be released on June 16, 2024. This release has been over a year in the making, and is the first major release since 2006. Importantly, in addition to many new features and performance improvement, it contains **breaking changes** to the ABI as well as the Python and C APIs. It is likely that downstream packages and end user code needs to be adapted - if you can, please verify whether your code works with NumPy `2.0.0rc2`. **Please see the following for more details:** + +- The [NumPy 2.0 migration guide](https://numpy.org/devdocs/numpy_2_0_migration_guide.html) +- The [2.0.0 release notes](https://numpy.org/devdocs/release/2.0.0-notes.html) +- Announcement issue for status updates: [numpy#24300](https://github.com/numpy/numpy/issues/24300) + + +### NumFOCUS end of the year fundraiser +_Dec 19, 2023_ -- NumFOCUS has teamed up with PyCharm during their EOY campaign to offer a 30% discount on first-time PyCharm licenses. All year-one revenue from PyCharm purchases from now until December 23rd, 2023 will go directly to the NumFOCUS programs. + +Use unique URL that will allow to track purchases https://lp.jetbrains.com/support-data-science/ or a coupon code ISUPPORTDATASCIENCE  + +### NumPy 1.26.0 released + +_Sep 16, 2023_ -- [NumPy 1.26.0](https://numpy.org/doc/stable/release/1.26.0-notes.html) is now available. The highlights of the release are: + +* Python 3.12.0 support. +* Cython 3.0.0 compatibility. +* Use of the Meson build system +* Updated SIMD support +* f2py fixes, meson and bind(x) support +* Support for the updated Accelerate BLAS/LAPACK library + +The NumPy 1.26.0 release is a continuation of the 1.25.x series that marks the transition to the Meson build system and provision of support for Cython 3.0.0. A total of 20 people contributed to this release and 59 pull requests were merged. + +The Python versions supported by this release are 3.9-3.12. + +### numpy.org is now available in Japanese and Portuguese + +_Aug 2, 2023_ -- numpy.org is now available in 2 additional languages: Japanese and Portuguese. This wouldn’t be possible without our dedicated volunteers: + +_Portuguese:_ +* Melissa Weber Mendonça (melissawm) +* Ricardo Prins (ricardoprins) +* Getúlio Silva (getuliosilva) +* Julio Batista Silva (jbsilva) +* Alexandre de Siqueira (alexdesiqueira) +* Alexandre B A Villares (villares) +* Vini Salazar (vinisalazar) + +_Japanese:_ +* Atsushi Sakai (AtsushiSakai) +* KKunai +* Tom Kelly (TomKellyGenetics) +* Yuji Kanagawa (kngwyu) +* Tetsuo Koyama (tkoyama010) + +The work on the translation infrastructure is supported with funding from CZI. + +Looking ahead, we’d love to translate the website into more languages. If you’d like to help, please connect with the NumPy Translations Team on Slack: https://join.slack.com/t/numpy-team/shared_invite/zt-1gokbq56s-bvEpo10Ef7aHbVtVFeZv2w. (Look for the #translations channel.) We are also building a Translations Team who will be working on localizing documentation and educational content across the Scientific Python ecosystem. If this piqued your interest, join us on the Scientific Python Discord: https://discord.gg/khWtqY6RKr. (Look for the #translation channel.) + +### NumPy 1.25.0 released + +_Jun 17, 2023_ -- [NumPy 1.25.0](https://numpy.org/doc/stable/release/1.25.0-notes.html) is now available. The highlights of the release are: + +* Support for MUSL, there are now MUSL wheels. +* Support for the Fujitsu C/C++ compiler. +* Object arrays are now supported in einsum. +* Support for the inplace matrix multiplication (`@=`). + +The NumPy 1.25.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase the execution speed, and clarify the documentation. There has also been preparatory work for the future NumPy 2.0.0, resulting in a large number of new and expired deprecations. + +A total of 148 people contributed to this release and 530 pull requests were merged. + +The Python versions supported by this release are 3.9-3.11. + +### Fostering an Inclusive Culture: Call for Participation + +_May 10, 2023_ -- Fostering an Inclusive Culture: Call for Participation + +How can we be better when it comes to diversity and inclusion? Read the report and find out how to get involved [here](https://contributor-experience.org/docs/posts/dei-report/). + +### NumPy documentation team leadership transition + +_Jan 6, 2023_ –- Mukulika Pahari and Ross Barnowski are appointed as the new NumPy documentation team leads replacing Melissa Mendonça. We thank Melissa for all her contributions to the NumPy official documentation and educational materials, and Mukulika and Ross for stepping up. + +### NumPy 1.24.0 released + +_Dec 18, 2022_ -- [NumPy 1.24.0](https://numpy.org/doc/stable/release/1.24.0-notes.html) is now available. The highlights of the release are: + +* New "dtype" and "casting" keywords for stacking functions. +* New F2PY features and fixes. +* Many new deprecations, check them out. +* Many expired deprecations, + +The NumPy 1.24.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase execution speed, and clarify the documentation. There are a large number of new and expired deprecations due to changes in dtype promotion and cleanups. It is the work of 177 contributors spread over 444 pull requests. The supported Python versions are 3.8-3.11. + +### Numpy 1.23.0 released + +_Jun 22, 2022_ -- [NumPy 1.23.0](https://numpy.org/doc/stable/release/1.23.0-notes.html) is now available. The highlights of the release are: + +* Implementation of `loadtxt` in C, greatly improving its performance. +* Exposure of DLPack at the Python level for easy data exchange. +* Changes to the promotion and comparisons of structured dtypes. +* Improvements to f2py. + +The NumPy 1.23.0 release continues the ongoing work to improve the handling and promotion of dtypes, increase the execution speed, clarify the documentation, and expire old deprecations. It is the work of 151 contributors spread over 494 pull requests. The Python versions supported by this release 3.8-3.10. Python 3.11 will be supported when it reaches the rc stage. + +### NumFOCUS DEI research study: call for participation + +_Apr 13, 2022_ -- NumPy is working with [NumFOCUS](http://numfocus.org/) on a [research project](https://numfocus.org/diversity-inclusion-disc/a-pivotal-time-in-numfocuss-project-aimed-dei-efforts?eType=EmailBlastContent&eId=f41a86c3-60d4-4cf9-86cf-58eb49dc968c) funded by the [Gordon & Betty Moore Foundation](https://www.moore.org/) to understand the barriers to participation that contributors, particularly those from historically underrepresented groups, face in the open-source software community. The research team would like to talk to new contributors, project developers and maintainers, and those who have contributed in the past about their experiences joining and contributing to NumPy. + +**Interested in sharing your experiences?** + +Please complete this brief [“Participant Interest” form](https://numfocus.typeform.com/to/WBWVJSqe) which contains additional information on the research goals, privacy, and confidentiality considerations. Your participation will be valuable to the growth and sustainability of diverse and inclusive open-source software communities. Accepted participants will participate in a 30-minute interview with a research team member. + +### Numpy 1.22.0 release + +_Dec 31, 2021_ -- [NumPy 1.22.0](https://numpy.org/doc/stable/release/1.22.0-notes.html) is now available. The highlights of the release are: + +* Type annotations of the main namespace are essentially complete. Upstream is a moving target, so there will likely be further improvements, but the major work is done. This is probably the most user visible enhancement in this release. +* A preliminary version of the proposed [array API Standard](https://data-apis.org/array-api/latest/) is provided (see [NEP 47](https://numpy.org/neps/nep-0047-array-api-standard.html)). This is a step in creating a standard collection of functions that can be used across libraries such as CuPy and JAX. +* NumPy now has a DLPack backend. DLPack provides a common interchange format for array (tensor) data. +* New methods for `quantile`, `percentile`, and related functions. The new methods provide a complete set of the methods commonly found in the literature. +* The universal functions have been refactored to implement most of [NEP 43](https://numpy.org/neps/nep-0043-extensible-ufuncs.html). This also unlocks the ability to experiment with the future DType API. +* A new configurable memory allocator for use by downstream projects. + +NumPy 1.22.0 is a big release featuring the work of 153 contributors spread over 609 pull requests. The Python versions supported by this release are 3.8-3.10. + +### Advancing an inclusive culture in the scientific Python ecosystem + +_August 31, 2021_ -- We are happy to announce the Chan Zuckerberg Initiative has [awarded a grant](https://chanzuckerberg.com/newsroom/czi-awards-16-million-for-foundational-open-source-software-tools-essential-to-biomedicine/) to support the onboarding, inclusion, and retention of people from historically marginalized groups on scientific Python projects, and to structurally improve the community dynamics for NumPy, SciPy, Matplotlib, and Pandas. + +As a part of [CZI's Essential Open Source Software for Science program](https://chanzuckerberg.com/eoss/), this [Diversity & Inclusion supplemental grant](https://cziscience.medium.com/advancing-diversity-and-inclusion-in-scientific-open-source-eaabe6a5488b) will support the creation of dedicated Contributor Experience Lead positions to identify, document, and implement practices to foster inclusive open-source communities. This project will be led by Melissa Mendonça (NumPy), with additional mentorship and guidance provided by Ralf Gommers (NumPy, SciPy), Hannah Aizenman and Thomas Caswell (Matplotlib), Matt Haberland (SciPy), and Joris Van den Bossche (Pandas). + +This is an ambitious project aiming to discover and implement activities that should structurally improve the community dynamics of our projects. By establishing these new cross-project roles, we hope to introduce a new collaboration model to the Scientific Python communities, allowing community-building work within the ecosystem to be done more efficiently and with greater outcomes. We also expect to develop a clearer picture of what works and what doesn't in our projects to engage and retain new contributors, especially from historically underrepresented groups. Finally, we plan on producing detailed reports on the actions executed, explaining how they have impacted our projects in terms of representation and interaction with our communities. + +The two-year project is expected to start by November 2021, and we are excited to see the results from this work! [You can read the full proposal here](https://figshare.com/articles/online_resource/Advancing_an_inclusive_culture_in_the_scientific_Python_ecosystem/16548063). + +### 2021 NumPy survey + +_July 12, 2021_ -- At NumPy, we believe in the power of our community. 1,236 NumPy users from 75 countries participated in our inaugural survey last year. The survey findings gave us a very good understanding of what we should focus on for the next 12 months. + +It’s time for another survey, and we are counting on you once again. It will take about 15 minutes of your time. Besides English, the survey questionnaire is available in 8 additional languages: Bangla, French, Hindi, Japanese, Mandarin, Portuguese, Russian, and Spanish. + +Follow the link to get started: https://berkeley.qualtrics.com/jfe/form/SV_aaOONjgcBXDSl4q. + + +### Numpy 1.21.0 release + +_Jun 23, 2021_ -- [NumPy 1.21.0](https://numpy.org/doc/stable/release/1.21.0-notes.html) is now available. The highlights of the release are: + +- continued SIMD work covering more functions and platforms, +- initial work on the new dtype infrastructure and casting, +- universal2 wheels for Python 3.8 and Python 3.9 on Mac, +- improved documentation, +- improved annotations, +- new `PCG64DXSM` bitgenerator for random numbers. + +This NumPy release is the result of 581 merged pull requests contributed by 175 people. The Python versions supported for this release are 3.7-3.9, support for Python 3.10 will be added after Python 3.10 is released. + + +### 2020 NumPy survey results + +_Jun 22, 2021_ -- In 2020, the NumPy survey team in partnership with students and faculty from the University of Michigan and the University of Maryland conducted the first official NumPy community survey. Find the survey results here: https://numpy.org/user-survey-2020/. + + +### Numpy 1.20.0 release + +_Jan 30, 2021_ -- [NumPy 1.20.0](https://numpy.org/doc/stable/release/1.20.0-notes.html) is now available. This is the largest NumPy release to date, thanks to 180+ contributors. The two most exciting new features are: +- Type annotations for large parts of NumPy, and a new `numpy.typing` submodule containing `ArrayLike` and `DtypeLike` aliases that users and downstream libraries can use when adding type annotations in their own code. +- Multi-platform SIMD compiler optimizations, with support for x86 (SSE, AVX), ARM64 (Neon), and PowerPC (VSX) instructions. This yielded significant performance improvements for many functions (examples: [sin/cos](https://github.com/numpy/numpy/pull/17587), [einsum](https://github.com/numpy/numpy/pull/18194)). + +### Diversity in the NumPy project + +_Sep 20, 2020_ -- We wrote a [statement on the state of, and discussion on social media around, diversity and inclusion in the NumPy project](/diversity_sep2020). + + +### First official NumPy paper published in Nature! + +_Sep 16, 2020_ -- We are pleased to announce the publication of [the first official paper on NumPy](https://www.nature.com/articles/s41586-020-2649-2) as a review article in Nature. This comes 14 years after the release of NumPy 1.0. The paper covers applications and fundamental concepts of array programming, the rich scientific Python ecosystem built on top of NumPy, and the recently added array protocols to facilitate interoperability with external array and tensor libraries like CuPy, Dask, and JAX. + + +### Python 3.9 is coming, when will NumPy release binary wheels? + +_Sept 14, 2020_ -- Python 3.9 will be released in a few weeks. If you are an early adopter of Python versions, you may be dissapointed to find that NumPy (and other binary packages like SciPy) will not have binary wheels ready on the day of the release. It is a major effort to adapt the build infrastructure to a new Python version and it typically takes a few weeks for the packages to appear on PyPI and conda-forge. In preparation for this event, please make sure to +- update your `pip` to version 20.1 at least to support `manylinux2010` and `manylinux2014` +- use [`--only-binary=numpy`](https://pip.pypa.io/en/stable/reference/pip_install/#cmdoption-only-binary) or `--only-binary=:all:` to prevent `pip` from trying to build from source. + + +### Numpy 1.19.2 release + +_Sep 10, 2020_ -- [NumPy 1.19.2](https://numpy.org/devdocs/release/1.19.2-notes.html) is now available. This latest release in the 1.19 series fixes several bugs, prepares for the [upcoming Cython 3.x release](http://docs.cython.org/en/latest/src/changes.html) and pins setuptools to keep distutils working while upstream modifications are ongoing. The aarch64 wheels are built with the latest manylinux2014 release that fixes the problem of differing page sizes used by different linux distros. + +### The inaugural NumPy survey is live! + +_Jul 2, 2020_ -- This survey is meant to guide and set priorities for decision-making about the development of NumPy as software and as a community. The survey is available in 8 additional languages besides English: Bangla, Hindi, Japanese, Mandarin, Portuguese, Russian, Spanish and French. + +Please help us make NumPy better and take the survey [here](https://umdsurvey.umd.edu/jfe/form/SV_8bJrXjbhXf7saAl). + + +### NumPy has a new logo! + +_Jun 24, 2020_ -- NumPy now has a new logo: + +NumPy logo + +The logo is a modern take on the old one, with a cleaner design. Thanks to Isabela Presedo-Floyd for designing the new logo, as well as to Travis Vaught for the old logo that served us well for 15+ years. + + +### NumPy 1.19.0 release + +_Jun 20, 2020_ -- NumPy 1.19.0 is now available. This is the first release without Python 2 support, hence it was a "clean-up release". The minimum supported Python version is now Python 3.6. An important new feature is that the random number generation infrastructure that was introduced in NumPy 1.17.0 is now accessible from Cython. + + +### Season of Docs acceptance + +_May 11, 2020_ -- NumPy has been accepted as one of the mentor organizations for the Google Season of Docs program. We are excited about the opportunity to work with a technical writer to improve NumPy's documentation once again! For more details, please see [the official Season of Docs site](https://developers.google.com/season-of-docs/) and our [ideas page](https://github.com/numpy/numpy/wiki/Google-Season-of-Docs-2020-Project-Ideas). + + +### NumPy 1.18.0 release + +_Dec 22, 2019_ -- NumPy 1.18.0 is now available. After the major changes in 1.17.0, this is a consolidation release. It is the last minor release that will support Python 3.5. Highlights of the release includes the addition of basic infrastructure for linking with 64-bit BLAS and LAPACK libraries, and a new C-API for `numpy.random`. + +Please see the [release notes](https://github.com/numpy/numpy/releases/tag/v1.18.0) for more details. + + +### NumPy receives a grant from the Chan Zuckerberg Initiative + +_Nov 15, 2019_ -- We are pleased to announce that NumPy and OpenBLAS, one of NumPy's key dependencies, have received a joint grant for $195,000 from the Chan Zuckerberg Initiative through their [Essential Open Source Software for Science program](https://chanzuckerberg.com/eoss/) that supports software maintenance, growth, development, and community engagement for open source tools critical to science. + +This grant will be used to ramp up the efforts in improving NumPy documentation, website redesign, and community development to better serve our large and rapidly growing user base, and ensure the long-term sustainability of the project. While the OpenBLAS team will focus on addressing sets of key technical issues, in particular thread-safety, AVX-512, and thread-local storage (TLS) issues, as well as algorithmic improvements in ReLAPACK (Recursive LAPACK) on which OpenBLAS depends. + +More details on our proposed initiatives and deliverables can be found in the [full grant proposal](https://figshare.com/articles/Proposal_NumPy_OpenBLAS_for_Chan_Zuckerberg_Initiative_EOSS_2019_round_1/10302167). The work is scheduled to start on Dec 1st, 2019 and continue for the next 12 months. + + + + +## Releases + +Here is a list of NumPy releases, with links to release notes. Bugfix releases (only the `z` changes in the `x.y.z` version number) have no new features; minor releases (the `y` increases) do. + +- NumPy 1.26.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.4)) -- _5 Feb 2024_. +- NumPy 1.26.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.3)) -- _2 Jan 2024_. +- NumPy 1.26.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.2)) -- _12 Nov 2023_. +- NumPy 1.26.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.1)) -- _14 Oct 2023_. +- NumPy 1.26.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.26.0)) -- _16 Sep 2023_. +- NumPy 1.25.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.25.2)) -- _31 Jul 2023_. +- NumPy 1.25.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.25.1)) -- _8 Jul 2023_. +- NumPy 1.24.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.4)) -- _26 Jun 2023_. +- NumPy 1.25.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.25.0)) -- _17 Jun 2023_. +- NumPy 1.24.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.3)) -- _22 Apr 2023_. +- NumPy 1.24.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.2)) -- _5 Feb 2023_. +- NumPy 1.24.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.1)) -- _26 Dec 2022_. +- NumPy 1.24.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.24.0)) -- _18 Dec 2022_. +- NumPy 1.23.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.5)) -- _19 Nov 2022_. +- NumPy 1.23.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.4)) -- _12 Oct 2022_. +- NumPy 1.23.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.3)) -- _9 Sep 2022_. +- NumPy 1.23.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.2)) -- _14 Aug 2022_. +- NumPy 1.23.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.1)) -- _8 Jul 2022_. +- NumPy 1.23.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.23.0)) -- _22 Jun 2022_. +- NumPy 1.22.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.4)) -- _20 May 2022_. +- NumPy 1.21.6 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.21.6)) -- _12 Apr 2022_. +- NumPy 1.22.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.3)) -- _7 Mar 2022_. +- NumPy 1.22.2 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.2)) -- _3 Feb 2022_. +- NumPy 1.22.1 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.1)) -- _14 Jan 2022_. +- NumPy 1.22.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.22.0)) -- _31 Dec 2021_. +- NumPy 1.21.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.21.5)) -- _19 Dec 2021_. +- NumPy 1.21.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.21.0)) -- _22 Jun 2021_. +- NumPy 1.20.3 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.20.3)) -- _10 May 2021_. +- NumPy 1.20.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.20.0)) -- _30 Jan 2021_. +- NumPy 1.19.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.19.5)) -- _5 Jan 2021_. +- NumPy 1.19.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.19.0)) -- _20 Jun 2020_. +- NumPy 1.18.4 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.18.4)) -- _3 May 2020_. +- NumPy 1.17.5 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.17.5)) -- _1 Jan 2020_. +- NumPy 1.18.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.18.0)) -- _22 Dec 2019_. +- NumPy 1.17.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.17.0)) -- _26 Jul 2019_. +- NumPy 1.16.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.16.0)) -- _14 Jan 2019_. +- NumPy 1.15.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.15.0)) -- _23 Jul 2018_. +- NumPy 1.14.0 ([release notes](https://github.com/numpy/numpy/releases/tag/v1.14.0)) -- _7 Jan 2018_. diff --git a/content/zh/press-kit.md b/content/zh/press-kit.md new file mode 100644 index 0000000000..2c8970bb29 --- /dev/null +++ b/content/zh/press-kit.md @@ -0,0 +1,8 @@ +--- +title: Press kit +sidebar: false +--- + +We would like to make it easy for you to include the NumPy project identity in your next academic paper, course materials, or presentation. + +You will find several high-resolution versions of the NumPy logo [here](https://github.com/numpy/numpy/tree/main/branding/logo). Note that by using the numpy.org resources, you accept the [NumPy Code of Conduct](/code-of-conduct). diff --git a/content/zh/privacy.md b/content/zh/privacy.md new file mode 100644 index 0000000000..6064e4c4f1 --- /dev/null +++ b/content/zh/privacy.md @@ -0,0 +1,8 @@ +--- +title: Privacy Policy +sidebar: false +--- + +**numpy.org** is operated by [NumFOCUS, Inc.](https://numfocus.org), the fiscal sponsor of the NumPy project. For the Privacy Policy of this website please refer to https://numfocus.org/privacy-policy. + +If you have any questions about the policy or NumFOCUS’s data collection, use, and disclosure practices, please contact the NumFOCUS staff at privacy@numfocus.org. diff --git a/content/zh/report-handling-manual.md b/content/zh/report-handling-manual.md new file mode 100644 index 0000000000..5586668cba --- /dev/null +++ b/content/zh/report-handling-manual.md @@ -0,0 +1,95 @@ +--- +title: NumPy Code of Conduct - How to follow up on a report +sidebar: false +--- + +This is the manual followed by NumPy’s Code of Conduct Committee. It’s used when we respond to an issue to make sure we’re consistent and fair. + +Enforcing the [Code of Conduct](/code-of-conduct) impacts our community today and for the future. It’s an action that we do not take lightly. When reviewing enforcement measures, the Code of Conduct Committee will keep the following values and guidelines in mind: + +* Act in a personal manner rather than impersonal. The Committee can engage the parties to understand the situation while respecting the privacy and any necessary confidentiality of reporters. However, sometimes it is necessary to communicate with one or more individuals directly: the Committee’s goal is to improve the health of our community rather than only produce a formal decision. +* Emphasize empathy for individuals rather than judging behavior, avoiding binary labels of “good” and “bad/evil”. Overt, clear-cut aggression and harassment exist, and we will address them firmly. But many scenarios that can prove challenging to resolve are those where normal disagreements devolve into unhelpful or harmful behavior from multiple parties. Understanding the full context and finding a path that re-engages all is hard, but ultimately the most productive for our community. +* We understand that email is a difficult medium and can be isolating. Receiving criticism over email, without personal contact, can be particularly painful. This makes it especially important to keep an atmosphere of open-minded respect for the views of others. It also means that we must be transparent in our actions, and that we will do everything in our power to make sure that all our members are treated fairly and with sympathy. +* Discrimination can be subtle and it can be unconscious. It can show itself as unfairness and hostility in otherwise ordinary interactions. We know that this does occur, and we will take care to look out for it. We would very much like to hear from you if you feel you have been treated unfairly, and we will use these procedures to make sure that your complaint is heard and addressed. +* Help increase engagement in good discussion practice: try to identify where discussion may have broken down, and provide actionable information, pointers, and resources that can lead to positive change on these points. +* Be mindful of the needs of new members: provide them with explicit support and consideration, with the aim of increasing participation from underrepresented groups in particular. +* Individuals come from different cultural backgrounds and native languages. Try to identify any honest misunderstandings caused by a non-native speaker and help them understand the issue and what they can change to avoid causing offence. Complex discussion in a foreign language can be very intimidating, and we want to grow our diversity also across nationalities and cultures. + + +## Mediation + +Voluntary informal mediation is a tool at our disposal. In contexts such as when two or more parties have all escalated to the point of inappropriate behavior (something sadly common in human conflict), it may be useful to facilitate a mediation process. This is only an example: the Committee can consider mediation in any case, mindful that the process is meant to be strictly voluntary and no party can be pressured to participate. If the Committee suggests mediation, it should: + +* Find a candidate who can serve as a mediator. +* Obtain the agreement of the reporter(s). The reporter(s) have complete freedom to decline the mediation idea or to propose an alternate mediator. +* Obtain the agreement of the reported person(s). +* Settle on the mediator: while parties can propose a different mediator than the suggested candidate, only if a common agreement is reached on all terms can the process move forward. +* Establish a timeline for mediation to complete, ideally within two weeks. + +The mediator will engage with all the parties and seek a resolution that is satisfactory to all. Upon completion, the mediator will provide a report (vetted by all parties to the process) to the Committee, with recommendations on further steps. The Committee will then evaluate these results (whether a satisfactory resolution was achieved or not) and decide on any additional action deemed necessary. + + +## How the Committee will respond to reports + +When the Committee (or a Committee member) receives a report, they will first determine whether the report is about a clear and severe breach (as defined below). If so, immediate action needs to be taken in addition to the regular report handling process. + + +## Clear and severe breach actions + +We know that it is painfully common for internet communication to start at or devolve into obvious and flagrant abuse. We will deal quickly with clear and severe breaches like personal threats, violent, sexist or racist language. + +When a member of the Code of Conduct Committee becomes aware of a clear and severe breach, they will do the following: + +* Immediately disconnect the originator from all NumPy communication channels. +* Reply to the reporter that their report has been received and that the originator has been disconnected. +* In every case, the moderator should make a reasonable effort to contact the originator, and tell them specifically how their language or actions qualify as a “clear and severe breach”. The moderator should also say that, if the originator believes this is unfair or they want to be reconnected to NumPy, they have the right to ask for a review, as below, by the Code of Conduct Committee. The moderator should copy this explanation to the Code of Conduct Committee. +* The Code of Conduct Committee will formally review and sign off on all cases where this mechanism has been applied to make sure it is not being used to control ordinary heated disagreement. + + +## Report handling + +When a report is sent to the Committee they will immediately reply to the reporter to confirm receipt. This reply must be sent within 72 hours, and the group should strive to respond much quicker than that. + +If a report doesn’t contain enough information, the Committee will obtain all relevant data before acting. The Committee is empowered to act on the Steering Council’s behalf in contacting any individuals involved to get a more complete account of events. + +The Committee will then review the incident and determine, to the best of their ability: + +* What happened. +* Whether this event constitutes a Code of Conduct violation. +* Who are the responsible party(ies). +* Whether this is an ongoing situation, and there is a threat to anyone’s physical safety. + +This information will be collected in writing, and whenever possible the group’s deliberations will be recorded and retained (i.e. chat transcripts, email discussions, recorded conference calls, summaries of voice conversations, etc). + +It is important to retain an archive of all activities of this Committee to ensure consistency in behavior and provide institutional memory for the project. To assist in this, the default channel of discussion for this Committee will be a private mailing list accessible to current and future members of the Committee as well as members of the Steering Council upon justified request. If the Committee finds the need to use off-list communications (e.g. phone calls for early/rapid response), it should in all cases summarize these back to the list so there’s a good record of the process. + +The Code of Conduct Committee should aim to have a resolution agreed upon within two weeks. In the event that a resolution can’t be determined in that time, the Committee will respond to the reporter(s) with an update and projected timeline for resolution. + + +## Resolutions + +The Committee must agree on a resolution by consensus. If the group cannot reach consensus and deadlocks for over a week, the group will turn the matter over to the Steering Council for resolution. + +Possible responses may include: + +* Taking no further action: + - if we determine no violations have occurred; + - if the matter has been resolved publicly while the Committee was considering responses. +* Coordinating voluntary mediation: if all involved parties agree, the Committee may facilitate a mediation process as detailed above. +* Remind publicly, and point out that some behavior/actions/language have been judged inappropriate and why in the current context, or can but hurtful to some people, requesting the community to self-adjust. +* A private reprimand from the Committee to the individual(s) involved. In this case, the group chair will deliver that reprimand to the individual(s) over email, cc’ing the group. +* A public reprimand. In this case, the Committee chair will deliver that reprimand in the same venue that the violation occurred, within the limits of practicality. E.g., the original mailing list for an email violation, but for a chat room discussion where the person/context may be gone, they can be reached by other means. The group may choose to publish this message elsewhere for documentation purposes. +* A request for a public or private apology, assuming the reporter agrees to this idea: they may at their discretion refuse further contact with the violator. The chair will deliver this request. The Committee may, if it chooses, attach “strings” to this request: for example, the group may ask a violator to apologize in order to retain one’s membership on a mailing list. +* A “mutually agreed upon hiatus” where the Committee asks the individual to temporarily refrain from community participation. If the individual chooses not to take a temporary break voluntarily, the Committee may issue a “mandatory cooling off period”. +* A permanent or temporary ban from some or all NumPy spaces (mailing lists, gitter.im, etc.). The group will maintain records of all such bans so that they may be reviewed in the future or otherwise maintained. + +Once a resolution is agreed upon, but before it is enacted, the Committee will contact the original reporter and any other affected parties and explain the proposed resolution. The Committee will ask if this resolution is acceptable, and must note feedback for the record. + +Finally, the Committee will make a report to the NumPy Steering Council (as well as the NumPy core team in the event of an ongoing resolution, such as a ban). + +The Committee will never publicly discuss the issue; all public statements will be made by the chair of the Code of Conduct Committee or the NumPy Steering Council. + + +## Conflicts of Interest + +In the event of any conflict of interest, a Committee member must immediately notify the other members, and recuse themselves if necessary. diff --git a/content/zh/tabcontents.yaml b/content/zh/tabcontents.yaml new file mode 100644 index 0000000000..d74cba9bce --- /dev/null +++ b/content/zh/tabcontents.yaml @@ -0,0 +1,373 @@ +params: + machinelearning: + paras: + - + para1: NumPy forms the basis of powerful machine learning libraries like [scikit-learn](https://scikit-learn.org) and [SciPy](https://www.scipy.org). As machine learning grows, so does the list of libraries built on NumPy. [TensorFlow’s](https://www.tensorflow.org) deep learning capabilities have broad applications — among them speech and image recognition, text-based applications, time-series analysis, and video detection. [PyTorch](https://pytorch.org), another deep learning library, is popular among researchers in computer vision and natural language processing. + para2: Statistical techniques called [ensemble](https://towardsdatascience.com/ensemble-methods-bagging-boosting-and-stacking-c9214a10a205) methods such as binning, bagging, stacking, and boosting are among the ML algorithms implemented by tools such as [XGBoost](https://xgboost.readthedocs.io/), [LightGBM](https://lightgbm.readthedocs.io/en/latest/), and [CatBoost](https://catboost.ai) — one of the fastest inference engines. [Yellowbrick](https://www.scikit-yb.org/en/latest/) and [Eli5](https://eli5.readthedocs.io/en/latest/) offer machine learning visualizations. + arraylibraries: + intro: + - + text: NumPy's API is the starting point when libraries are written to exploit innovative hardware, create specialized array types, or add capabilities beyond what NumPy provides. + headers: + - + text: Array Library + - + text: Capabilities & Application areas + libraries: + - + title: Dask + text: Distributed arrays and advanced parallelism for analytics, enabling performance at scale. + img: /images/content_images/arlib/dask.png + alttext: Dask + url: https://dask.org/ + - + title: CuPy + text: NumPy-compatible array library for GPU-accelerated computing with Python. + img: /images/content_images/arlib/cupy.png + alttext: CuPy + url: https://cupy.dev + - + title: JAX + text: "Composable transformations of NumPy programs: differentiate, vectorize, just-in-time compilation to GPU/TPU." + img: /images/content_images/arlib/jax_logo_250px.png + alttext: JAX + url: https://jax.readthedocs.io/ + - + title: Xarray + text: Labeled, indexed multi-dimensional arrays for advanced analytics and visualization. + img: /images/content_images/arlib/xarray.png + alttext: xarray + url: https://xarray.pydata.org/en/stable/index.html + - + title: Sparse + text: NumPy-compatible sparse array library that integrates with Dask and SciPy's sparse linear algebra. + img: /images/content_images/arlib/sparse.png + alttext: sparse + url: https://sparse.pydata.org/en/latest/ + - + title: PyTorch + text: Deep learning framework that accelerates the path from research prototyping to production deployment. + img: /images/content_images/arlib/pytorch-logo-dark.svg + alttext: PyTorch + url: https://pytorch.org/ + - + title: TensorFlow + text: An end-to-end platform for machine learning to easily build and deploy ML powered applications. + img: /images/content_images/arlib/tensorflow-logo.svg + alttext: TensorFlow + url: https://www.tensorflow.org + - + title: Arrow + text: A cross-language development platform for columnar in-memory data and analytics. + img: /images/content_images/arlib/arrow.png + alttext: arrow + url: https://arrow.apache.org/ + - + title: xtensor + text: Multi-dimensional arrays with broadcasting and lazy computing for numerical analysis. + img: /images/content_images/arlib/xtensor.png + alttext: xtensor + url: https://github.com/xtensor-stack/xtensor-python + - + title: Awkward Array + text: Manipulate JSON-like data with NumPy-like idioms. + img: /images/content_images/arlib/awkward.svg + alttext: awkward + url: https://awkward-array.org/ + - + title: uarray + text: Python backend system that decouples API from implementation; unumpy provides a NumPy API. + img: /images/content_images/arlib/uarray.png + alttext: uarray + url: https://uarray.org/en/latest/ + - + title: tensorly + text: Tensor learning, algebra and backends to seamlessly use NumPy, PyTorch, TensorFlow or CuPy. + img: /images/content_images/arlib/tensorly.png + alttext: tensorly + url: http://tensorly.org/stable/home.html + scientificdomains: + intro: + - + text: Nearly every scientist working in Python draws on the power of NumPy. + - + text: "NumPy brings the computational power of languages like C and Fortran to Python, a language much easier to learn and use. With this power comes simplicity: a solution in NumPy is often clear and elegant." + libraries: + - + title: Quantum Computing + alttext: A computer chip. + img: /images/content_images/sc_dom_img/quantum_computing.svg + links: + - + url: http://qutip.org + label: QuTiP + - + url: https://pyquil-docs.rigetti.com/en/stable + label: PyQuil + - + url: https://qiskit.org + label: Qiskit + - + url: https://pennylane.ai + label: PennyLane + - + title: Statistical Computing + alttext: A line graph with the line moving up. + img: /images/content_images/sc_dom_img/statistical_computing.svg + links: + - + url: https://pandas.pydata.org/ + label: Pandas + - + url: https://www.statsmodels.org/ + label: statsmodels + - + url: https://xarray.pydata.org/en/stable/ + label: Xarray + - + url: https://seaborn.pydata.org/ + label: Seaborn + - + title: Signal Processing + alttext: A bar chart with positive and negative values. + img: /images/content_images/sc_dom_img/signal_processing.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://pywavelets.readthedocs.io/ + label: PyWavelets + - + url: https://python-control.org/ + label: python-control + - + url: https://hyperspy.org/ + label: HyperSpy + - + title: Image Processing + alttext: An photograph of the mountains. + img: /images/content_images/sc_dom_img/image_processing.svg + links: + - + url: https://scikit-image.org/ + label: Scikit-image + - + url: https://opencv.org/ + label: OpenCV + - + url: https://mahotas.rtfd.io/ + label: Mahotas + - + title: Graphs and Networks + alttext: A simple graph. + img: /images/content_images/sc_dom_img/sd6.svg + links: + - + url: https://networkx.org/ + label: NetworkX + - + url: https://graph-tool.skewed.de/ + label: graph-tool + - + url: https://igraph.org/python/ + label: igraph + - + url: https://pygsp.rtfd.io/ + label: PyGSP + - + title: Astronomy + alttext: A telescope. + img: /images/content_images/sc_dom_img/astronomy_processes.svg + links: + - + url: https://www.astropy.org/ + label: AstroPy + - + url: https://sunpy.org/ + label: SunPy + - + url: https://spacepy.github.io/ + label: SpacePy + - + title: Cognitive Psychology + alttext: A human head with gears. + img: /images/content_images/sc_dom_img/cognitive_psychology.svg + links: + - + url: https://www.psychopy.org/ + label: PsychoPy + - + title: Bioinformatics + alttext: A strand of DNA. + img: /images/content_images/sc_dom_img/bioinformatics.svg + links: + - + url: https://biopython.org/ + label: BioPython + - + url: http://scikit-bio.org/ + label: Scikit-Bio + - + url: https://github.com/openvax/pyensembl + label: PyEnsembl + - + url: http://etetoolkit.org/ + label: ETE + - + title: Bayesian Inference + alttext: A graph with a bell-shaped curve. + img: /images/content_images/sc_dom_img/bayesian_inference.svg + links: + - + url: https://pystan.readthedocs.io/en/latest/ + label: PyStan + - + url: https://docs.pymc.io/ + label: PyMC3 + - + url: https://arviz-devs.github.io/arviz/ + label: ArviZ + - + url: https://emcee.readthedocs.io/ + label: emcee + - + title: Mathematical Analysis + alttext: Four mathematical symbols. + img: /images/content_images/sc_dom_img/mathematical_analysis.svg + links: + - + url: https://www.scipy.org/ + label: SciPy + - + url: https://www.sympy.org/ + label: SymPy + - + url: https://www.cvxpy.org/ + label: cvxpy + - + url: https://fenicsproject.org/ + label: FEniCS + - + title: Chemistry + alttext: A test tube. + img: /images/content_images/sc_dom_img/chemistry.svg + links: + - + url: https://cantera.org/ + label: Cantera + - + url: https://www.mdanalysis.org/ + label: MDAnalysis + - + url: https://github.com/rdkit/rdkit + label: RDKit + - + url: https://www.pybamm.org/ + label: PyBaMM + - + title: Geoscience + alttext: The Earth. + img: /images/content_images/sc_dom_img/geoscience.svg + links: + - + url: https://pangeo.io/ + label: Pangeo + - + url: https://simpeg.xyz/ + label: Simpeg + - + url: https://github.com/obspy/obspy/wiki + label: ObsPy + - + url: https://www.fatiando.org/ + label: Fatiando a Terra + - + title: Geographic Processing + alttext: A map. + img: /images/content_images/sc_dom_img/GIS.svg + links: + - + url: https://shapely.readthedocs.io/ + label: Shapely + - + url: https://geopandas.org/ + label: GeoPandas + - + url: https://python-visualization.github.io/folium + label: Folium + - + title: Architecture & Engineering + alttext: A microprocessor development board. + img: /images/content_images/sc_dom_img/robotics.svg + links: + - + url: https://compas.dev/ + label: COMPAS + - + url: https://cityenergyanalyst.com/ + label: City Energy Analyst + - + url: https://nortikin.github.io/sverchok/ + label: Sverchok + datascience: + intro: "NumPy lies at the core of a rich ecosystem of data science libraries. A typical exploratory data science workflow might look like:" + image1: + - + img: /images/content_images/ds-landscape.png + alttext: Diagram of Python Libraries. The five catagories are 'Extract, Transform, Load', 'Data Exploration', 'Data Modeling', 'Data Evaluation' and 'Data Presentation'. + image2: + - + img: /images/content_images/data-science.png + alttext: Diagram of three overlapping circles. The circles are labeled 'Mathematics', 'Computer Science' and 'Domain Expertise'. In the middle of the diagram, which has the three circles overlapping it, is an area labeled 'Data Science'. + examples: + - + text: "Extract, Transform, Load: [Pandas](https://pandas.pydata.org), [Intake](https://intake.readthedocs.io), [PyJanitor](https://pyjanitor-devs.github.io/pyjanitor/)" + - + text: "Exploratory analysis: [Jupyter](https://jupyter.org), [Seaborn](https://seaborn.pydata.org), [Matplotlib](https://matplotlib.org), [Altair](https://altair-viz.github.io)" + - + text: "Model and evaluate: [scikit-learn](https://scikit-learn.org), [statsmodels](https://www.statsmodels.org/stable/index.html), [PyMC3](https://docs.pymc.io), [spaCy](https://spacy.io)" + - + text: "Report in a dashboard: [Dash](https://plotly.com/dash), [Panel](https://panel.holoviz.org), [Voila](https://voila.readthedocs.io/)" + content: + - + text: For high data volumes, [Dask](https://dask.org) and [Ray](https://ray.io/) are designed to scale. Stable deployments rely on data versioning ([DVC](https://dvc.org)), experiment tracking ([MLFlow](https://mlflow.org)), and workflow automation ([Airflow](https://airflow.apache.org), [Dagster](https://dagster.io) and [Prefect](https://www.prefect.io)). + visualization: + images: + - + url: https://www.fusioncharts.com/blog/best-python-data-visualization-libraries + img: /images/content_images/v_matplotlib.png + alttext: A streamplot made in matplotlib + - + url: https://github.com/yhat/ggpy + img: /images/content_images/v_ggpy.png + alttext: A scatter-plot graph made in ggpy + - + url: https://www.journaldev.com/19692/python-plotly-tutorial + img: /images/content_images/v_plotly.png + alttext: A box-plot made in plotly + - + url: https://altair-viz.github.io/gallery/streamgraph.html + img: /images/content_images/v_altair.png + alttext: A streamgraph made in altair + - + url: https://seaborn.pydata.org + img: /images/content_images/v_seaborn.png + alttext: A pairplot of two types of graph, a plot-graph and a frequency graph made in seaborn" + - + url: https://docs.pyvista.org/ + img: /images/content_images/v_pyvista.png + alttext: A 3D volume rendering made in PyVista. + - + url: https://napari.org + img: /images/content_images/v_napari.png + alttext: A multi-dimensionan image made in napari. + - + url: https://vispy.org/gallery/index.html + img: /images/content_images/v_vispy.png + alttext: A Voronoi diagram made in vispy. + content: + - + text: NumPy is an essential component in the burgeoning [Python visualization landscape](https://pyviz.org/overviews/index.html), which includes [Matplotlib](https://matplotlib.org), [Seaborn](https://seaborn.pydata.org), [Plotly](https://plot.ly), [Altair](https://altair-viz.github.io), [Bokeh](https://docs.bokeh.org/en/latest/), [Holoviz](https://holoviz.org), [Vispy](http://vispy.org), [Napari](https://napari.org/), and [PyVista](https://docs.pyvista.org/), to name a few. + - + text: NumPy's accelerated processing of large arrays allows researchers to visualize datasets far larger than native Python could handle. diff --git a/content/zh/teams.md b/content/zh/teams.md new file mode 100644 index 0000000000..8a32f6e5ef --- /dev/null +++ b/content/zh/teams.md @@ -0,0 +1,22 @@ +--- +title: NumPy Teams +sidebar: false +--- + +我们是一个国际团队，通过建造优质开放源码软件，支持世界各地的科学和研究 个社区。 [加入我们]({{< relref "/contribute" >}})！ + +{{< include-html "static/gallery/maintainers.html" >}} + +{{< include-html "static/gallery/docs-team.html" >}} + +{{< include-html "static/gallery/web-team.html" >}} + +{{< include-html "static/gallery/triage-team.html" >}} + +{{< include-html "static/gallery/survey-team.html" >}} + +{{< include-html "static/gallery/emeritus-maintainers.html" >}} + +# Governance + +For the list of the Steering Council members, please see [here](https://numpy.org/about/). diff --git a/content/zh/user-survey-2020.md b/content/zh/user-survey-2020.md new file mode 100644 index 0000000000..b4349bcb7d --- /dev/null +++ b/content/zh/user-survey-2020.md @@ -0,0 +1,18 @@ +--- +title: 2020 NUMPY COMMUNITY SURVEY +sidebar: false +--- + +In 2020, the NumPy survey team in partnership with students and faculty from a Master’s course in Survey Methodology jointly hosted by the University of Michigan and the University of Maryland conducted the first official NumPy community survey. Over 1,200 users from 75 countries participated to help us map out a landscape of the NumPy community and voiced their thoughts about the future of the project. + +{{< figure >}} +src = '/surveys/NumPy_usersurvey_2020_report_cover.png' alt = 'Cover page of the 2020 NumPy user survey report, titled "NumPy Community Survey 2020 - results"' width = '250' +{{< /figure >}} + +**[Download the report](/surveys/NumPy_usersurvey_2020_report.pdf)** to take a closer look at the survey findings. + + +For the highlights, check out **[this infographic](https://github.com/numpy/numpy-surveys/blob/master/images/2020NumPysurveyresults_community_infographic.pdf)**. + +Ready for a deep dive? Visit **https://numpy.org/user-survey-2020-details/**. + diff --git a/content/zh/user-surveys.md b/content/zh/user-surveys.md new file mode 100644 index 0000000000..89a2aa0460 --- /dev/null +++ b/content/zh/user-surveys.md @@ -0,0 +1,10 @@ +--- +title: NUMPY USER SURVEYS +sidebar: false +--- + +**2020** The NumPy survey team in partnership with students and faculty from the University of Michigan and the University of Maryland conducted the first official NumPy community survey. Find the survey results [here](https://numpy.org/user-survey-2020/). + +**2021** The collected data is currently being analyzed. + +If you have any questions or suggestions for the past or future surveys, please open an issue [here](https://github.com/numpy/numpy-surveys/issues).