Cortex Engines should be semantically named?

[dan-menlo]

Problem

• Right now, our engines are named cortex.llamacpp, cortex.tensorrt-llm, and cortex.onnx
• We also have cortex.python, which is not an engine but more of a runtime
• This is potentially confusing to Devs, as these are the Github repos
• This is also confusing to Users

Alternatives

• Semantic naming: e.g. llamacpp-engine, tensorrtllm-engine, onnx-engine
• For ONNX, should it be ONNXRuntime instead of ONNX?
• Is "*-runtime" better? (or should we stick to our nomenclature of
• What should we call python-engine? (it's not really an engine, but more of a Extension base template?)

Considerations

• Will we encounter any issues if we make the changes now? (e.g. any legacy users who depend on old naming/URL structure?)
• Should we consider all of them as extensions
  • There is a base C++ Extension format
  • There is a base Python Extension format (that packages its own python binary, i.e. cortex.python)

[vansangpfiev]

• If we only rename the engines, not the repositories, it will not break anything cc: @hiento09
• @dan-homebrew Could you please elaborate more on Should we consider all of them as extensions?