v2.3.0

Added

• High-level backup() and restore() helpers wrapping JSONL/CSV workflows
• MkDocs documentation scaffold with API reference, operations playbook, and migration guide
• Backup/restore coverage in the integration test suite

Fixed

• Enforced embedding dimension validation across add/update/search operations
• import_from_json() and import_from_csv() now respect skip_duplicates and emit clear errors when embeddings are missing

Documentation

• New migration guide outlining v2.3.0 changes
• Expanded README with backup helper examples and coverage instructions
• Requirements updated with MkDocs to build the documentation locally

CI

• Pytest coverage step now generates XML output and uploads coverage.xml as a GitHub Actions artifact

v2.2.0

Added

• Metadata filtering: New filter_by_metadata() method for efficient key-value filtering using SQLite's JSON_EXTRACT
• Metadata counting: New count_by_metadata() method to count records matching metadata filters
• Combined search: New similarity_search_with_filter() method combining vector similarity with metadata filtering
• Support for nested JSON paths in metadata filters (e.g., {"author.name": "Alice"})
• Pagination support in filter_by_metadata() with limit and offset parameters
• New validation function validate_metadata_filters() for secure filter validation
• New utility function build_metadata_where_clause() for safe SQL generation
• Comprehensive test coverage for metadata filtering (unit, integration, and security tests)
• New example advanced_metadata_queries.py demonstrating nested paths and complex queries
• Updated metadata_filtering.py example with new filtering methods
• Export/import functionality: New methods for data portability
  • export_to_json() - Export records to JSON Lines format
  • import_from_json() - Import records from JSON Lines format
  • export_to_csv() - Export records to CSV format
  • import_from_csv() - Import records from CSV format
• Support for filtered exports using metadata filters
• Batch processing for memory-efficient import/export of large datasets
• Optional embedding inclusion in exports
• New io module with import/export utilities
• Comprehensive tests for import/export functionality
• New example export_import_example.py demonstrating backup/restore workflows

Security

• SQL injection prevention in metadata filter keys
• Validation of JSON paths to prevent malicious queries
• Parameterized queries for all metadata filtering operations

Improved

• Data migration workflows now much easier
• Backup and restore capabilities for production use
• Interoperability with external systems via CSV/JSON

Documentation

• Added "Metadata Filtering" section to README with examples
• Added "Export/Import" section to README with examples
• Updated examples list in README
• Added comprehensive docstrings for new methods

v2.1.1

Changed

• Moved table name validation from __init__() to create_table() method
• Table name is now validated when creating the table schema, not during client initialization

Improved

• More flexible client initialization - validation happens at table creation time

v2.1.0

Added

• Connection pooling support via ConnectionPool class
• Thread-safe connection reuse for concurrent access scenarios
• Optional pool parameter in SQLiteVecClient constructor
• examples/connection_pool_example.py demonstrating pooled connections
• Comprehensive tests for connection pooling in tests/test_pool.py

Improved

• Better resource management for multi-threaded applications
• Connection reuse reduces overhead in high-concurrency scenarios
• Backward compatible - pooling is optional

v2.0.0

Changed

• BREAKING: Simplified update_many() - removed complex grouping logic (58% code reduction)
• BREAKING: Renamed get_by_id() → get() for cleaner API
• BREAKING: Renamed delete_by_id() → delete() for cleaner API

Removed

• BREAKING: Removed get_by_text() - use SQL queries or get_all() with filtering
• BREAKING: Removed get_by_metadata() - use SQL queries or get_all() with filtering
• BREAKING: Removed list_results() - use get_all() generator instead

Added

• Kept count() method for convenience (user request)

Improved

• 28% smaller codebase (650 → 467 lines)
• 15% fewer methods (20 → 17)
• Test coverage increased 89% → 92%
• Cleaner, more intuitive API
• Better code maintainability
• All core CRUD and bulk operations preserved

v1.2.0

Added

• Benchmarks module with comprehensive performance testing for CRUD operations
• Configurable benchmark suite via benchmarks/config.yaml
• Support for testing with different dataset sizes (100, 1K, 10K, 50K records)
• Benchmark operations: add, get_many, similarity_search, update_many, get_all, delete_many

Fixed

• Fixed delete_many() to handle large rowid lists by batching into chunks of 500 to avoid SQLite's "too many SQL variables" error

v1.0.1

Changed

• Patch release

v1.0.0

Added

• SQLiteVecClient class for CRUD and similarity search operations
• create_table() - Initialize schema with configurable dimensions and distance metrics
• add() - Insert texts with embeddings
• update() - Update existing records
• update_many() - Bulk update multiple records in a single transaction
• delete() - Remove records by rowid
• get() - Fetch single record by rowid
• get_many() - Fetch multiple records
• get_all() - Memory-efficient generator for iterating over all records
• get_by_text() - Find records by exact text match
• get_by_metadata() - Find records by metadata JSON match
• list_all() - List all records with pagination
• similarity_search() - Vector similarity search with configurable top_k
• count() - Get total record count
• clear() - Remove all records
• drop_table() - Delete table and index
• transaction() - Context manager for atomic multi-operation transactions
• Support for text, metadata (JSON), and float32 embeddings
• Automatic synchronization between base table and vec0 index via triggers
• Context manager support for automatic connection cleanup
• Distance metrics: cosine, L2, L1
• Python logging module integration with configurable log levels
• Type hints and dataclasses for typed results
• Comprehensive test suite with 91%+ coverage
• Complete CI/CD pipeline with GitHub Actions
• Enhanced GitHub release workflow with improved changelog extraction
• Better error handling in CI/CD pipeline
• Pre-commit hooks and code quality tools (ruff, mypy)
• CONTRIBUTING.md with contribution guidelines
• CHANGELOG.md for tracking changes
• TESTING.md with testing documentation
• SECURITY_IMPROVEMENTS.md documenting security enhancements
• Multiple examples in examples/ directory
• Updated batch_operations.py example with new features
• Comprehensive README with quick start guide

Security

• SQL injection prevention with table name validation
• Input validation for all parameters
• Custom exception classes for better error handling

Fixed

• PyPI publishing workflow with tag-based releases
• Changelog extraction in GitHub releases
• Changelog parsing in release notes generation