Repository navigation
OneSDK v0.0.4: Comprehensive Enhancement and Multi-Provider Optimization
We are excited to announce the release of OneSDK v0.0.4, featuring significant improvements across the board. This update brings enhanced functionality, expanded provider support, and improved documentation.
New Features and Improvements:
-
API Enhancements:
- Optimized API implementations for multiple providers, including Anthropic, Qwen, Doubao, Kimi, MiniMax, OpenAI, and Wenxin.
- Standardized error handling across all APIs for improved reliability and consistency.
- Implemented additional provider-specific features such as multimodal support, knowledge base management, and context caching.
-
Extended Provider Support:
- Added preliminary support for Ollama, Gemini, and Cohere (documentation provided, full testing pending).
- Expanded capabilities for existing providers, including new methods and improved integration.
-
Documentation Overhaul:
- Updated main README with a comprehensive provider support table, offering clear insight into each provider's capabilities.
- Enhanced individual provider documentation in the docs/ directory for improved clarity and completeness.
- Revised core SDK method descriptions for consistency across all providers.
-
Testing Improvements:
- Added corresponding test files for each optimized API, significantly increasing test coverage and reliability.
-
Package Management:
- Optimized setup.py for better dependency handling and package configuration.
-
Logging and Debugging:
- Improved logging capabilities for easier troubleshooting and debugging.
-
Performance Optimizations:
- Enhanced overall performance and reduced latency in API calls.
Known Issues:
- The APIs for Ollama, Gemini, and Cohere are still in development and may not be fully functional or optimized.
Upgrading:
To upgrade to OneSDK v0.0.4, use pip:
pip install --upgrade llm_onesdk
We recommend reviewing the updated documentation for any changes that might affect your existing implementations.
Future Plans:
- Complete testing and optimization for Ollama, Gemini, and Cohere APIs.
- Further enhance multi-modal capabilities across supported providers.
- Implement additional features based on user feedback and provider updates.
We appreciate your continued support and feedback. If you encounter any issues or have suggestions, please open an issue on our GitHub repository.
Thank you for using OneSDK!