LLM Medical Chatbot

A Retrieval-Augmented Generation (RAG) based medical chatbot leveraging Large Language Models (LLMs) and a vector database for accurate, context-aware medical question answering.

---

🚀 Project Overview

• Domain: Medical Question Answering
• Core Technologies: Python, React, Qdrant, Mistral-7B (or other LLMs)
• Features:
  • Retrieval-Augmented Generation (RAG) pipeline
  • Semantic search over medical knowledge bases (MedQuAD, PubMed, etc.)
  • Modern frontend for chat-based interaction
  • Scalable backend with RESTful APIs

---

📁 Repository Structure

• frontend/ — React-based web UI
• scripts/ — Data processing, retrieval, and backend scripts
• models/ — (Large files, download separately)
• data/ — (Large files, download separately)
• qdrant_data/ — (Vector DB files, download separately)

---

⚡️ Quickstart

1. Clone the Repository

git clone https://github.com/IshaanPotle/LLM_Medical-Chatbot.git
cd LLM_Medical-Chatbot

2. Install Dependencies

• Backend:
  • Python 3.8+
  • Install required packages (see scripts/requirements.txt or your script headers)
• Frontend:
  • Node.js 18+
  • cd frontend && npm install

3. Download Large Files

This repository does not include large files (models, data, vector DB) due to GitHub size limits.

Download links (all large files):

• Hugging Face Repository: https://huggingface.co/IshaanPotle27/llm_medical_chatbot/tree/main/.
  • Models: Download from the models/ folder
  • Processed Data: Download from the data/ folder
  • Qdrant Vector DB: Download from the qdrant_data/ folder

Instructions:

1. Download the files from the above Hugging Face link.
2. Place them in the corresponding folders (models/, data/, qdrant_data/).

4. Run the Backend

# Example (adjust as needed)
python scripts/backend_rag.py

5. Run the Frontend

cd frontend
npm run dev

---

🧩 Key Features

• Retrieval-Augmented Generation: Combines LLMs with semantic search for accurate, context-rich answers.
• Medical Knowledge Bases: Utilizes MedQuAD, PubMed, and other trusted sources.
• Scalable Architecture: Modular backend and frontend for easy extension.

---

🗂️ Data Sources

• MedQuAD
• PubMed

---

📝 Customization

• Swap in different LLMs or embedding models as needed.
• Add new data sources by updating the scripts in scripts/.

---

🤝 Contributing

Pull requests and issues are welcome!

---

📄 License

See MedQuAD-master/LICENSE.txt and project root for licensing details.

---

📬 Contact

For questions or collaboration, open an issue or contact IshaanPotle.