MediAssist-AI

MediAssist-AI is a production-ready Retrieval-Augmented Generation (RAG) chatbot for medical question answering. It combines LangChain, OpenAI, Pinecone vector search, and a Flask web interface to deliver context-aware responses from your medical knowledge base.

Built and maintained by Shraddha3838 (Coder Shraddha).

---

Features

• PDF ingestion pipeline with chunking and metadata normalization
• HuggingFace embeddings (all-MiniLM-L6-v2) stored in Pinecone
• GPT-powered RAG responses with safety-focused system prompts
• Responsive chat UI with real-time AJAX messaging
• Docker support and GitHub Actions CI/CD for AWS ECR + EC2 deployment
• Health check endpoint at /health

---

Tech Stack

Layer	Technology
Backend	Python, Flask
Orchestration	LangChain
LLM	OpenAI GPT
Vector DB	Pinecone
Embeddings	Sentence Transformers
Deployment	Docker, AWS ECR, EC2, GitHub Actions

---

Project Structure

mediassist-ai/
├── app.py                 # Flask application entrypoint
├── store_index.py         # Embedding upload script
├── src/
│   ├── config.py          # Environment-driven configuration
│   ├── helper.py          # PDF loading and text splitting utilities
│   └── prompt.py          # RAG system prompt
├── data/                  # Place medical PDF files here
├── templates/             # HTML templates
├── static/                # CSS assets
├── Dockerfile
└── .github/workflows/     # CI/CD pipeline

---

Getting Started

1. Clone the repository

git clone https://github.com/Shraddha3838/git.git
cd git

2. Create a virtual environment

conda create -n mediassist python=3.10 -y
conda activate mediassist

3. Install dependencies

pip install -r requirements.txt

4. Configure environment variables

Copy the example file and add your API keys:

cp .env.example .env

PINECONE_API_KEY=your_pinecone_key
OPENAI_API_KEY=your_openai_key

5. Add medical documents

Place your PDF files inside the data/ directory.

6. Build the vector index

python store_index.py

7. Run the application

python app.py

Open http://localhost:8080 in your browser.

---

Docker

docker build -t mediassist-ai .
docker run -p 8080:8080 --env-file .env mediassist-ai

---

AWS Deployment

1. Create an ECR repository and EC2 instance with Docker installed.
2. Configure a self-hosted GitHub Actions runner on EC2.
3. Add these repository secrets:
   • AWS_ACCESS_KEY_ID
   • AWS_SECRET_ACCESS_KEY
   • AWS_DEFAULT_REGION
   • ECR_REPO
   • PINECONE_API_KEY
   • OPENAI_API_KEY

Push to main to trigger automated build and deployment.

---

Disclaimer

MediAssist-AI is intended for educational and informational use only. It does not provide medical diagnosis or emergency guidance. Always consult a licensed healthcare provider for medical decisions.

---

Author

Shraddha3838 — Fullstack Developer & AIML Engineer
GitHub: https://github.com/Shraddha3838

---

License

This project is licensed under the MIT License. See LICENSE for details.