Doc-bot 🩺 – Simple Medical Chatbot

Doc-bot is a chatbot that answers health-related questions using a medical PDF book.
Built with React, FastAPI, Gemini AI, and Pinecone.

Example: “What are the side effects of antidepressants?”

How It Works

• Frontend: React app (App.jsx) for chatting.

• Backend: FastAPI (app.py) searches the book & calls AI.

• AI Service (Node.js): geminiService.js uses the Gemini API to generate answers based on the retrieved context.

• Vector Search (Pinecone): Stores embeddings (vector representations) of the medical PDF's text chunks for fast similarity search.

• RAG Pipeline:

  • Document Processing – The PDF is split into small chunks (few hundred words) for better search accuracy.
  • Embeddings – Each chunk is converted into a numeric vector using an embedding model.
  • Storage – These vectors are stored in Pinecone’s vector database.
  • Query – When a user asks a question, it is converted to a vector embedding.
  • Retrieval – Pinecone finds the most similar text chunks from the stored vectors.
  • Answer Generation – The retrieved chunks are sent to Gemini AI along with the question to create a contextual, accurate answer.

Setup & Run

1.Clone the Repository

git clone https://github.com/Ayush2911ojha/Doc-bot.git
cd Doc-bot

2. In the Doc-bot folder (root), create a .env file:

PINECONE_API_KEY=your_pinecone_api_key
GEMINI_API_KEY=your_gemini_api_key
PORT=4000

3. Install & Run FastAPI Backend

python -m venv venv
venv\Scripts\activate

Install backend dependencies:

pip install -r requirements.txt

Start the FastAPI backend (from the root directory):

    
 uvicorn app:app --host 0.0.0.0 --port 8080

4. Install & Run Gemini Node.js Service

npm install express cors dotenv
node geminiService.js

5. Run React Frontend

cd frontend
npm install
npm start

Usage

Open http://localhost:5173
Ask questions like:

• Hello
• What are the side effects of antidepressants?

General Notes

• Ensure that your FastAPI and Node.js servers are running before using the frontend.

• You may need to adjust CORS settings or environment variables for local development.

• The backend expects the medical PDF to be chunked, embedded, and indexed in Pinecone during initial setup.