Skip to content

Releases: sharanharsoor/chunking

Release v0.4.1

10 Oct 11:26

Choose a tag to compare

Chunking Strategy Library v0.4.1 Release

🚀 Major Updates
Version 0.4.1 brings significant improvements to the chunking strategy library, focusing on enhanced reliability, better cross-platform support, and comprehensive code coverage integration.

✨ New Features
Codecov Integration: Added comprehensive code coverage reporting with Codecov integration
Enhanced CI/CD: Improved GitHub Actions workflow with better dependency management
Platform Optimization: Streamlined support for Linux and macOS platforms

🔧 Improvements
Dependency Management: Fixed critical numpy import issues with conditional loading
CLI Enhancements: Improved Windows compatibility by replacing Unicode characters with ASCII alternatives
Test Robustness: Enhanced test suite with more flexible assertions and better error handling
Documentation: Updated platform support information and installation instructions

🐛 Bug Fixes
Critical Fix: Resolved ModuleNotFoundError for numpy in core embeddings module
Windows Compatibility: Fixed Unicode encoding issues in CLI output
Test Failures: Addressed timing assertion issues across different environments
File Handling: Improved cross-platform temporary file handling
Path Issues: Fixed Windows path separator problems in test files

Release v0.4.0

10 Oct 05:48

Choose a tag to compare

Release v0.4.0: Enhanced Documentation & Comprehensive Codebase Improvements
🎯 What's New
This release focuses on significantly improving documentation quality, enhancing codebase maintainability, and strengthening PIL/image type handling across the library.
📚 Documentation Enhancements

Improved Codebase Documentation: Comprehensive updates to API reference documentation, making it easier for developers to understand and integrate the chunking strategies
Enhanced README: Added clearer examples and use cases for all 40+ chunking strategies
Better Configuration Examples: Expanded configuration file examples with real-world scenarios

🔧 Technical Improvements

PIL/Image Type Hints: Fixed and enhanced type hints for image processing functions, improving IDE support and type checking
Coverage Improvements: Added .coverage ignore patterns to improve development workflow and test coverage accuracy
Code Quality: General codebase cleanup and refactoring for better maintainability

🐛 Bug Fixes

Fixed CI issues related to PIL/image type handling
Resolved coverage reporting inconsistencies

🚀 Performance

No breaking changes - fully backward compatible with v0.3.0
All existing features continue to work seamlessly

📦 Installation
bashpip install chunking-strategy==0.4.0

Or with all features

pip install chunking-strategy[all]==0.4.0

Release v0.3.0

24 Sep 19:05

Choose a tag to compare

v0.3.0 with all improvements

🚀 Release v0.2.0: Adaptive Chunking Learning System

19 Sep 20:47

Choose a tag to compare

🎯 Major Features:

AI-powered adaptive chunking with machine learning capabilities
Content profiling and performance learning system
Session persistence and historical optimization
30-50% better chunk quality through intelligent strategy selection

📚 Enhanced Documentation & Examples:

Comprehensive adaptive chunking guide (164 lines)
6 new example files including LangChain, Streamlit integrations
Complete performance monitoring and benchmarking demos
Interactive Streamlit web app with metrics

🐛 Bug Fixes:

• Fixed Streamlit app statistics import issue
• Enhanced file upload support for all file types
• Improved error handling and user experience
• Updated author information to "Sharan Harsoor"

🔧 Developer Experience:

list_strategies alias for better API usability
• Cleaned up redundant documentation files
• Production-ready adaptive chunking system
• Fixed PyPI badges for better visibility

📦 Installation:

pip install chunking-strategy==0.2.0

🚀 Try New Features:

# Try adaptive chunking learning demo
python examples/22_adaptive_chunking_learning_demo.py

# Run enhanced Streamlit app
streamlit run examples/19_streamlit_app_demo.py

🔗 What's New:

  • 22 comprehensive examples including adaptive learning
  • Enhanced Streamlit integration with performance metrics
  • LangChain RAG pipeline demonstration
  • Production-ready adaptive system with session persistence

Full Changelog: v0.2.0...v0.2.0

v0.1.0 - Initial Release

17 Sep 09:13

Choose a tag to compare

First stable release of chunking-strategy!

Features:

  • 40+ chunking strategies for text, documents, multimedia
  • Full CLI interface
  • Configuration support
  • Batch processing capabilities