v0.0.3-alpha

🚀 What's New

REST API Integration

• Added new REST API endpoint /api/chat for programmatic access
• Implemented Bearer token authentication for API security
• Support for version-specific documentation queries
• Structured JSON responses with answer and source references

Architecture Improvements

• Refactored core logic into separate module for better maintainability
• Added shared functionality between API and UI components
• Improved error handling and response formatting
• Enhanced token validation and security

Code Organization

• Introduced proper package structure with __init__.py files
• Created dedicated core module for shared business logic
• Separated API routing and handling from core functionality
• Better type hinting and documentation

🔧 Technical Details

API Features

• Bearer token authentication
• JSON request/response format
• Version-specific document retrieval
• Source reference tracking
• Conversation history support

🔐 Environment Changes

• Added new environment variable:

BEARER_TOKENS=comma,separated,tokens

📚 Documentation

• Added API documentation to README
• Updated setup instructions
• Added API usage examples

Known Issues

• Large conversation histories may impact response time
• Some complex queries might timeout in high-load scenarios

v0.0.2-alpha

🎯 Overview

Major update to the documentation processing pipeline with improved version handling, smarter RST preprocessing, and enhanced semantic search capabilities. This release features a complete refactoring of the core components and introduces version-aware document processing and querying.

🔄 Major Changes

Architecture & Core Changes

• Complete refactoring of document processing pipeline
• Added version-aware document processing and retrieval
• Improved RST preprocessing with better directive handling
• Enhanced URL and anchor generation for better documentation linking
• Streamlined file processing workflow

Documentation Processing

• Added specialized RST content preprocessing
  • Improved handling of Odoo-specific RST directives
  • Better cleaning of navigation sections and file links
  • Smarter handling of metadata blocks
• Enhanced markdown chunking with better section preservation
• Improved header path creation and section anchoring
• Added version metadata to processed chunks

Search & Retrieval

• Version-specific document retrieval
• Optimized Supabase vector search function
• Improved embedding handling for better semantic search
• Enhanced context preparation for more relevant responses

✨ New Features

• Version-specific documentation fetching via pull_rawdata.sh
• Smart RST preprocessing with section anchor generation
• Version-aware document processing and search
• Improved conversation context handling
• Version selector in chat interface

🐛 Bug Fixes

• Fixed metadata handling in RST conversion
• Improved error handling in document processing
• Fixed section anchor generation
• Resolved version-specific URL generation issues
• Better handling of special RST directives

🔧 Technical Improvements

• Refactored codebase for better maintainability
• Improved type hints and documentation
• Enhanced error handling and logging
• Better memory management in document processing
• Optimized database operations

v0.0.1-alpha

Overview

This v0.0.1-alpha release introduces a complete Odoo documentation processing pipeline that converts HTML documentation to markdown, processes it for semantic search, and provides a chat interface for querying the documentation. The system uses OpenAI embeddings and GPT models to enable intelligent documentation search and Q&A capabilities.

Key Features

HTML to Markdown Conversion (convert_html_to_md.py)

• Recursive HTML file processing
• Structured directory preservation
• Meta-refresh detection and handling
• Parallel processing using asyncio

Document Processing (document_processor.py)

• Intelligent document chunking with header preservation
• OpenAI embeddings generation
• Automatic title and summary extraction
• Supabase database integration for vector search
• Support for processing both single files and directories

Interactive Chat Interface (streamlit_ui.py)

• Real-time semantic search
• Streaming responses
• Conversation history management
• Context-aware responses using relevant documentation chunks
• Clear conversation functionality

Known Limitations

• Maximum text chunk size of 8000 characters for embeddings
• Requires Supabase database with vector similarity search capability
• HTML conversion requires local file access