How laybaatariq/langchain-medicalchatbot works
This is a medical-information chatbot with a React chat interface and a FastAPI backend. A user submits a message; the backend loads session history, routes and triages the question, then returns an emergency response, conversational answer, document-grounded answer, or general answer. The interface renders the reply and any sources, while chat history is stored in MongoDB and medical-document retrieval uses Qdrant. The README provides no capability detail; additional branches are based on the source index and tree, so some ingestion and frontend-to-backend links remain unverified.
Main entry points: medical_qa_chain.py Answer workflow · vector_store.py Vector search · App.jsx Chat app · chat.py Chat endpoint
Components 30
Chat interface
- Chat app
frontend/src/App.jsx - Chat state
frontend/src/hooks/useChat.js - API client
frontend/src/api.js - Message composer
frontend/src/components/Composer.jsx - Empty state
frontend/src/components/EmptyState.jsx - Message view
frontend/src/components/Message.jsx - Source display
frontend/src/components/Sources.jsx - Chat icons
frontend/src/components/Icons.jsx
API and sessions
- FastAPI app
backend/app/main.py - Chat endpoint
backend/app/api/routes/chat.py - Health endpoint
backend/app/api/routes/health.py - API schemas
backend/app/api/schemas.py - Chat history
backend/app/db/mongo.py
Answer orchestration
- Answer workflow
backend/app/chains/medical_qa_chain.py - Intent router
backend/app/chains/router.py - Answer prompts
backend/app/chains/system_prompts.py
Safety and knowledge
- Safety checks
backend/app/safety/emergency.py - Keyword filters
backend/app/safety/keyword_filters.py - Triage classifier
backend/app/safety/triage_classifier.py - Rules and doctors
backend/app/core/knowledge.py - Clinical rule data
backend/app/data/rules
Document retrieval
- Vector search
backend/app/rag/vector_store.py - Document ingestion
backend/app/rag/ingest.py - MedlinePlus ingestion
backend/app/rag/ingest_medlineplus.py - Medical source files
backend/app/data/medical_docs - MedlinePlus source
backend/app/data/medlineplus
Other components
- User
- App settings
backend/app/core/config.py - Qdrant Cloud
- MongoDB Atlas
Connections 36
- User → Message composer: enters question
- Chat app → Chat state: uses chat state
frontend/src/App.jsx - Chat app → Message composer: passes send handler
frontend/src/App.jsx - Chat app → Empty state: renders
frontend/src/App.jsx - Chat app → Message view: renders replies
frontend/src/App.jsx - Chat app → Chat icons: renders
frontend/src/App.jsx - Chat state → API client: calls
frontend/src/hooks/useChat.js - Message view → Source display: renders sources
frontend/src/components/Message.jsx - Message view → Chat icons: renders
frontend/src/components/Message.jsx - Message composer → Chat icons: renders
frontend/src/components/Composer.jsx - Empty state → Chat icons: renders
frontend/src/components/EmptyState.jsx - API client → Chat endpoint: sends chat request
- FastAPI app → Chat endpoint: registers route
backend/app/main.py - FastAPI app → Health endpoint: registers route
backend/app/main.py - FastAPI app → Chat history: sets indexes
backend/app/main.py - FastAPI app → Vector search: checks collection
backend/app/main.py - Chat endpoint → API schemas: uses models
backend/app/api/routes/chat.py - Health endpoint → API schemas: uses models
backend/app/api/routes/health.py - Chat endpoint → Answer workflow: requests answer
backend/app/api/routes/chat.py - Chat endpoint → Chat history: reads and saves history
backend/app/api/routes/chat.py - Chat history → MongoDB Atlas: stores chat records
- Answer workflow → Intent router: routes message
backend/app/chains/medical_qa_chain.py - Answer workflow → Answer prompts: uses prompts
backend/app/chains/medical_qa_chain.py - Answer workflow → Safety checks: checks urgency
backend/app/chains/medical_qa_chain.py - Answer workflow → Keyword filters: checks keywords
backend/app/chains/medical_qa_chain.py - Intent router → Safety checks: guesses language
backend/app/chains/router.py - Safety checks → Rules and doctors: loads red flags
backend/app/safety/emergency.py - Rules and doctors → Clinical rule data: loads rule files
backend/app/core/knowledge.py - Answer workflow → Vector search: searches passages
backend/app/chains/medical_qa_chain.py - Vector search → Qdrant Cloud: queries vectors
- Document ingestion → Vector search: upserts chunks
backend/app/rag/ingest.py - MedlinePlus ingestion → Vector search: upserts chunks
backend/app/rag/ingest_medlineplus.py - Medical source files → Document ingestion: source input
- MedlinePlus source → MedlinePlus ingestion: source input
- App settings → Chat history: configures connection
backend/app/db/mongo.py - App settings → Vector search: configures connection
backend/app/rag/vector_store.py