Enhance agent instructions and tool usage guidelines for lung cancer queries. Added critical instructions for handling greetings and non-medical questions, specifying that tools should not be used in these cases. Updated the medical guidelines knowledge tool documentation to reinforce its exclusive use for medical inquiries. Improved clarity in the agent's purpose and tool usage rules to ensure accurate responses.
Refine agent instructions for medical and non-medical queries. Clarified tool usage requirements, emphasizing that the "medical_guidelines_knowledge_tool" is mandatory for medical questions related to lung cancer, while non-medical questions should be answered directly without tools. Updated guidelines for side effect reporting and provider comparisons to enhance clarity and usability.
Enhance medical query handling by adding optional guideline parameter to API requests. Updated QueryRequest model to include 'guideline' field, allowing users to specify guideline providers (IASLC, NCCN, ASCO, ESMO, NICE) for tailored responses. Adjusted related functions to accommodate this new parameter. Updated documentation and examples accordingly.
Refactor project for Lung Cancer AI Advisor: update app and API descriptions, modify .gitignore to exclude Jupyter notebooks, and remove outdated deployment documentation. Delete unused files and enhance logging for better traceability.
Update .env.example with OpenAI and LangSmith configuration, modify app.py to dynamically set the port for deployment, enhance CORS middleware to support additional local development origins, and improve document retrieval settings for more comprehensive context in responses.
Enhance API security and functionality by adding authentication middleware and session management. Updated app.py to include the new auth router and integrated authentication checks for protected endpoints. Modified requirements.txt to include necessary libraries for session handling. Updated .env.example to include authentication credentials. Improved retrieval functions with query expansion for better medical term matching and enriched context in responses.
Enhance automatic validation process in agent.py. Updated response handling to provide context and next steps when guidelines are not found, ensuring professionalism. Modified validation function to run silently in the background, logging results for backend analysis without displaying them to the user. Improved error handling and logging for validation failures.
Add session management endpoints to health and medical routers. Implemented functionality to retrieve active sessions, clear specific sessions, and clear all sessions. Updated ask and ask_stream endpoints to support optional session IDs for improved conversation continuity.
Refactor agent and tools for session-based memory management and side effect reporting. Removed medical answer validation tool, added session memory management class, and enhanced side effect reporting with LLM classification. Updated agent functions to support session IDs for improved conversation tracking.
