- Added streaming service support to the Q&A agent for real-time token streaming.
- Updated `answer_question` method to stream responses token-by-token to the frontend.
- Modified `handle_qna_workflow` to handle both custom and values streaming modes.
- Enhanced state management to include streaming service for improved user experience.
- Updated the document saving endpoint to trigger reindexing after saving.
- Introduced a new Celery task for reindexing documents.
- Refactored the editor page to reflect the changes in the API endpoint and method.
- Added `validate_llm_config` function to `llm_service.py` for validating LLM configurations via test API calls.
- Integrated validation in `create_llm_config` and `update_llm_config` routes in `llm_config_routes.py`, raising HTTP exceptions for invalid configurations.
- Enhanced error handling to provide detailed feedback on configuration issues.
- Increased task time limits in celery_app.py for longer processing times.
- Enhanced pagination logic in NotionHistoryConnector to handle large result sets.
- Implemented batch commits every 10 documents across various indexers (Airtable, ClickUp, Confluence, Discord, GitHub, Google Calendar, Gmail, JIRA, Linear, Luma, Notion, Slack) to improve performance and reduce database load.
- Updated final commit logging for clarity on total documents processed.
- Added warning suppression for pdfminer warnings during Docling PDF processing
- Suppresses 'Cannot set gray non-stroke color' warnings that cause uploads to halt
- Temporarily sets pdfminer logger to ERROR level during document processing
- Fixes issue where files ~34MB would fail due to pdfminer warning spam
Resolves issue where PDF uploads would halt with repeated pdfminer warnings
- Added WindowsProactorEventLoopPolicy for better async subprocess support on Windows.
- Ensured proper cleanup of the event loop after task execution.
- Removed redundant asyncio import from the generate_chat_podcast_task function.
