v0.3.2 Release

What's added:

• Added new github, predict and train CLI commands
• Reworked the check and explain CLI commands
• Added GitHub integration for searching and reporting issues via CLI
• Added ML capabilities with an engine for error prediction and pattern training, including NeuralNetwork, ErrorEmbedding, FeatureExtractor, and MLEngine
• Added monitoring and prediction features: file watching, static analysis (SimpleChecker for AST-based checks), and ML-based error prediction with severity and confidence scoring
• Added storage and caching: ConfigManager for settings (e.g., AI providers, API keys), HistoryManager (SQLite-based for error history), PatternManager (JSON-based), and CacheManager
• Added data models: Error, Pattern, Prediction, and TrainingData dataclasses
• Added utilities: logging, path validation, and enhanced helpers (e.g., detect_all_errors for syntax and semantic issues)
• Added comprehensive testing: unit tests for models, parsers, predictor, trainer, GitHub; integration tests for CLI commands and end-to-end; performance tests for load time, memory, and prediction speed
• Added CI/CD workflow (pypi-publish.yml)
• Enhanced error explanation logic with structured output and pattern-based fixes
• Enhanced CLI usability with rich formatting, input validation, command grouping, and version display
• Reworked codebase structure
• Updated dependencies in requirements.txt and pyproject.toml
• Updated setup.py for new package structure, metadata, and simplified install_requires
• Updated init.py files for better exports and dynamic attribute access

What's removed:

• Removed monolithic parser (core/parser.py) and watcher (core/watcher.py), replaced with modular designs
• Removed old AI logic (ai/init.py, ai/prompts.py, ai/providers.py), moved to integrations/ai/
• Removed patterns directory (patterns/init.py and language-specific JSON files), integrated into storage/patterns.py