- Enhance error handling in control tool with more specific domain validation
- Modify list devices tool to use direct filtering instead of manual iteration
- Add more descriptive success messages for different device domains and services
- Simplify device state filtering logic in list devices tool
- Improve intent confidence calculation with more nuanced scoring
- Add comprehensive keyword and pattern matching for better intent detection
- Refactor confidence calculation to handle various input scenarios
- Implement more aggressive boosting for specific action keywords
- Adjust parameter extraction logic for more robust intent parsing
- Standardized error handling across tool implementations
- Improved return type consistency for tool execution results
- Simplified configuration parsing and type definitions
- Enhanced type safety for various configuration schemas
- Cleaned up and normalized tool response structures
- Updated SSE and event subscription tool implementations
- Updated package.json with latest dependency versions
- Added winston-daily-rotate-file type definitions
- Refactored import paths in tools and routes to use .js extensions
- Modified SSE routes to use getStatistics method
- Improved log rotation utility with more robust file cleanup
- Removed performance monitoring, tools service, and polyfills modules
- Moved formatToolCall helper function to test file
- Simplified imports in index.ts by removing polyfills import
- Cleaned up unnecessary files and consolidated code structure
- Refactored main application entry point to use centralized configuration
- Created modular route structure with separate files for different API endpoints
- Introduced app.config.ts for centralized environment variable management
- Moved tools and route logic into dedicated files
- Simplified index.ts and improved overall project organization
- Added comprehensive type definitions for tools and API interactions
- Introduced comprehensive NLP processing modules for intent classification, entity extraction, and context analysis
- Created AI router with rate-limited endpoints for command interpretation and execution
- Added prompt templates for different AI models with configurable system and user prompts
- Implemented robust type definitions for AI-related interfaces and schemas
- Enhanced security and error handling in AI processing pipeline
- Introduced WebSocket client for real-time Home Assistant event streaming
- Created context management system for tracking resource relationships and state
- Implemented performance monitoring and optimization utilities
- Added security middleware with token validation, rate limiting, and input sanitization
- Extended tool registry with enhanced tool registration and execution capabilities
- Expanded test coverage for new modules and added comprehensive test scenarios
- Improved type safety and added robust error handling across new modules
