This is a Retrieval-Augmented Generation (RAG) based chatbot designed to answer medical questions based on a provided set of documents. It's built with LangChain, Groq, and Streamlit.

Live Demo Link: https://medichat-bot.streamlit.app/

• Conversational AI: Ask medical questions in natural language.
• RAG Pipeline: Retrieves relevant information from medical documents to provide accurate, context-aware answers.
• Fast Inference: Powered by the high-speed Groq LPU™ Inference Engine.
• Built with Industry Standards: Modular code structure, dependency management, and ready for deployment.

The application follows a standard RAG architecture:

1. Data Ingestion: Medical documents (PDFs) are loaded and split into chunks.
2. Vector Store: Text chunks are converted into embeddings (using Hugging Face models) and stored in a FAISS vector store.
3. Retrieval: When a user asks a question, the system retrieves the most relevant chunks from the vector store.
4. Generation: The retrieved context and the user's question are passed to a Large Language Model (Groq's Llama4) to generate a final answer.

1. Clone the repository:

   git clone https://github.com/ihimanshu29/Medical-Chatbot.git

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows, use `venv\Scripts\activate`

3. Install dependencies:

   pip install -r requirements.txt

4. Set up environment variables:
   • Create a .env file in the root directory.
   • Add your API keys to the .env file:

     HUGGINGFACE_API_KEY="your_hf_key"
     GROQ_API_KEY="your_groq_key"
     

5. Build the knowledge base: (Ensure you have your PDF documents inside the data/ folder)

   python -m src.pipeline.training_pipeline

6. Run the Streamlit app:

   python -m streamlit run src/app.py

7. Project Structure:

   ├── data/
   │ └── medical_data.pdf
   ├── src/
   │ ├── **init**.py
   │ ├── components/
   │ │ ├── **init**.py
   │ │ ├── data_ingestion.py
   │ │ ├── llm_interface.py
   │ │ └── vector_store.py
   │ └── pipeline/
   │ ├── **init**.py
   │ ├── prediction_pipeline.py
   │ └── training_pipeline.py
   ├── vectorstore/
   │ └── db_faiss/
   ├── app.py
   ├── config.py
   ├── .env
   ├── .gitignore
   ├── README.md
   └── requirements.txt