Medical RAG Chatbot

A Retrieval-Augmented Generation (RAG) chatbot built with Flask and LangChain for answering medical queries using the GALE Encyclopedia of Medicine as the knowledge base.

Features

• Medical Knowledge Base: Uses the GALE Encyclopedia of Medicine (Second Edition) as the primary data source
• RAG Architecture: Combines retrieval and generation for accurate, context-aware responses
• Vector Database: FAISS-powered vector store for efficient similarity search
• Web Interface: Clean Flask-based web UI for chat interactions
• Session Management: Maintains conversation history during user sessions
• Logging: Comprehensive logging system for debugging and monitoring
• Hugging Face Integration: Powered by Mistral-7B-Instruct-v0.3 model

Architecture

├── app/
│   ├── application.py          # Flask web application
│   ├── components/
│   │   ├── data_loader.py      # PDF document loading
│   │   ├── embeddings.py       # Text embeddings generation
│   │   ├── llm.py             # Language model integration
│   │   ├── pdf_loader.py      # PDF processing utilities
│   │   ├── retriever.py       # QA chain and retrieval logic
│   │   └── vector_store.py    # FAISS vector database
│   ├── config/
│   │   └── config.py          # Configuration settings
│   ├── common/
│   │   ├── custom.py          # Custom exceptions
│   │   └── logger.py          # Logging utilities
│   └── template/
│       └── index.html         # Web interface
├── data/                      # Medical documents
├── vector_store/              # FAISS database files
└── logs/                      # Application logs

Prerequisites

• Python 3.8+
• Hugging Face account and API token
• Sufficient disk space for vector database storage

Installation

1. Clone the repository:

   git clone <repository-url>
   cd medical_rag

2. Install dependencies:

   pip install -r requirements.txt

3. Set up environment variables:

   export HF_TOKEN="your_huggingface_token"
   export HFHUB_TOKEN="your_huggingface_hub_token"

4. Install the package:

   pip install -e .

Usage

1. Start the Flask application:

   python app/application.py

2. Access the web interface: Open your browser and navigate to http://localhost:5000

3. Ask medical questions:

   • Type your medical query in the chat interface
   • The system will search the medical encyclopedia and provide relevant answers
   • Use the "Clear" button to reset the conversation

Configuration

Key settings in app/config/config.py:

• HF_REPO_ID: Hugging Face model identifier (default: Mistral-7B-Instruct-v0.3)
• CHUNK_SIZE: Text chunk size for document processing (default: 500)
• CHUNK_OVERLAP: Overlap between text chunks (default: 50)
• DB_FAISS_PATH: Vector database storage path

API Endpoints

• GET/POST /: Main chat interface
• GET /clear: Clear conversation history

Data Source

The chatbot uses "The GALE ENCYCLOPEDIA of MEDICINE SECOND" as its primary knowledge source, providing comprehensive medical information across various topics.

Logging

Application logs are stored in the logs/ directory with daily rotation. Log levels and formats can be configured in app/common/logger.py.

Limitations

• Responses are limited to information contained in the medical encyclopedia
• Not a substitute for professional medical advice
• Retrieval accuracy depends on query formulation and document chunking

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Run tests (if available)
5. Submit a pull request

License

This project is developed by Sudhanshu. Please check with the author for licensing terms.

Disclaimer

This chatbot is for educational and informational purposes only. It should not be used as a substitute for professional medical advice, diagnosis, or treatment. Always consult with qualified healthcare professionals for medical concerns.