- Added background task processing for document uploads to improve responsiveness.
- Updated the DocumentProcessor to use synchronous processing.
- Introduced namespace configuration in VectorStore for better organization of stored embeddings.
- Enhanced logging to reflect changes in document processing and embedding storage.
- Replaced SQLite direct connections with SQLAlchemy ORM for better abstraction and maintainability.
- Introduced new database models for 'Analysis' and 'Metadata' with appropriate fields.
- Enhanced database initialization and session management.
- Updated methods for saving, retrieving, and deleting analysis and metadata records to use SQLAlchemy sessions.
- Removed debug print statements from analysis retrieval.
- Implemented a reranking method for issues and recommendations based on relevance to compliance.
- Updated the analysis HTML template to display issues with improved formatting and markdown support.
- Added support for processing DOCX files and extracting content.
- Updated database schema to include a combined 'issues_and_recommendations' field.
- Improved error handling during document uploads and analysis.
- Modified the analysis display to show issues and recommendations in a structured format.
- Adjusted API call parameters for better performance and error management.