Tags: FuryCow/llm_chain
Tags
Release v0.5.5
* Major refactor:
* Core classes have been extensively refactored for improved modularity, clarity, and maintainability:
* Adopted SOLID principles throughout the codebase.
* Extracted interfaces for memory, tool management, and all builder components, now located in a dedicated interfaces namespace.
* Introduced a builders folder and builder pattern for prompt, memory context, tool responses, RAG documents, and retriever context.
* Improved dependency injection and separation of concerns, making the codebase easier to extend and test.
* Centralized error handling and configuration validation.
* Enhanced documentation and type signatures for all major classes.
* The public API remains minimal and idiomatic, with extensibility via interfaces and factories.
* Chain#ask method rewritten following Ruby best practices: now declarative, each pipeline stage is a private method, code is cleaner and easier to extend.
* All ToolManager creation and configuration is now done via a dedicated factory: ToolManagerFactory. Old calls (ToolManager.create_default_toolset, ToolManager.from_config) have been replaced with factory methods.
Version 0.5.4: Deepseek-Coder-V2 client support This release adds support for Deepseek-Coder-V2 models, a specialized code generation LLM. New features: - Deepseek-Coder-V2 client with optimized settings for code tasks - Support for multiple model variants (latest, 16b, 236b) - Seamless integration with existing tool ecosystem - Enhanced CLI with proper development mode detection The Deepseek-Coder-V2 models are particularly effective for: - Code generation and explanation - Algorithm implementation - Code review and optimization - Programming language translation
Release v0.5.2: Enhanced Stability and Diagnostics π Major Release Features: β’ π System Diagnostics - LLMChain.diagnose_system for health checks β’ βοΈ Configuration Validation - Early detection of setup issues β’ π Enhanced Error Handling - Retry logic with exponential backoff β’ π Robust WebSearch - Better timeout and graceful degradation β’ π» Improved CodeInterpreter - Support for 8+ code extraction formats β’ π Debug Logging - Structured logging with LLM_CHAIN_DEBUG=true π‘οΈ Reliability Improvements: β’ Automatic retry for network failures (3 attempts, exponential backoff) β’ Graceful fallback for offline scenarios β’ Better timeout handling (15-20s with proper error messages) β’ Enhanced validation with helpful warnings and suggestions β’ Improved memory management and resource cleanup π§ Technical Enhancements: β’ ConfigurationValidator class for comprehensive system checks β’ Enhanced code extraction (markdown, inline, Windows line endings) β’ Offline mode detection with internet connectivity checks β’ Structured error handling with specific exception types β’ Performance optimizations for chain creation π Documentation: β’ Updated README with comprehensive examples β’ Enhanced CHANGELOG with detailed feature descriptions β’ All new features documented and tested β’ Backward compatibility maintained π§ͺ Quality: β’ 100% integration test coverage β’ All README examples verified β’ Performance benchmarks passing β’ Production-ready stability This release significantly improves the stability and user experience of LLMChain.
Release Candidate v0.5.2-rc.1 π― Pre-production release for testing v0.5.2 features: β System diagnostics and health checks β Configuration validation β Enhanced error handling with retry logic β Robust WebSearch with graceful degradation β Improved CodeInterpreter with better code extraction β Structured logging and debug mode π§ͺ Ready for testing: β’ All integration tests passing β’ README examples verified β’ Backward compatibility maintained Next: Final release v0.5.2 after validation