🩺 GenPAI — Generative Physician AI

GenPAI is a multi-component medical AI system designed to assist patients with non-emergency general health questions.
It combines a modern Angular frontend, a .NET Core API backend, a Python FastAPI microservice powered by OpenAI + LangChain, and Redis for advanced similarity search.

---

🌟 Project Structure

/genpai/ /frontend/ → Angular app (chat UI) /api-server/ → .NET Core Web API (routing + orchestration) /llm-server/ → Python FastAPI LLM service (OpenAI + LangChain) /redis/ → Redis vector store (RediSearch)

---

📦 Components

🖼 Frontend (Angular)

• Chat UI for users to submit symptoms and queries
• Displays live AI responses, dynamic loading dots, and conversation history
• Connects to the backend API via HTTP

⚙ Backend API (C# / .NET Core)

• Receives patient queries from the Angular frontend
• Calls Python FastAPI LLM service for generative responses
• Acts as the secure middle layer between frontend and AI logic

🧠 LLM Service (Python / FastAPI)

• Handles natural language questions and symptom queries
• Uses OpenAI GPT (with LangChain RAG support)
• Connects to Redis vector search for embedding-based retrieval

🏪 Redis (Vector Store)

• Stores medical knowledge chunks + embeddings
• Provides fast KNN similarity search (using RediSearch)

---

🔑 Environment & Config

✅ Included in version control:

• Angular app source
• .NET Core API (excluding sensitive configs)
• Python LLM FastAPI service (excluding .env)
• Example configs

❌ Ignored via .gitignore:

• llm-server/.env → holds your OPENAI_API_KEY
• llm-server/venv/ → Python virtual environment
• api-server/PhysicianAI.Api/bin/ and obj/ → .NET build artifacts
• api-server/PhysicianAI.Api/appsettings.Development.json → local dev secrets

---

⚡ Quick Start

1️⃣ Frontend (Angular)

cd frontend
npm install
npm run start

2️⃣ .NET API Server

cd api-server/PhysicianAI.Api
dotnet run

3️⃣ Python FastAPI LLM Server

cd llm-server
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
uvicorn app.main:app --host 0.0.0.0 --port 5000

4️⃣ Redis (with RediSearch)

docker run -p 6379:6379 redis/redis-stack-server:latest

🏗 Architecture Overview

[ Angular Frontend ] → [ .NET API Server ] → [ Python FastAPI LLM Service ] → [ Redis Vector DB ]

• Frontend sends HTTP requests to .NET API
• .NET API orchestrates backend calls, forwards queries to FastAPI
• FastAPI runs LLM + RAG pipeline, retrieves data from Redis
• Response flows back to the user chat window

🔐 Secrets Handling

• .env in Python → holds OPENAI_API_KEY
• appsettings.Development.json in .NET → holds dev environment configs (ignored in Git)

OPENAI_API_KEY=your-openai-api-key-here
REDIS_HOST=localhost
REDIS_PORT=6379

Example appsettings.Development.json:

{
  "Logging": {
    "LogLevel": {
      "Default": "Information"
    }
  },
  "AllowedHosts": "*"
}

✨ Features

✅ Handles non-emergency symptom Q&A ✅ Provides dynamic AI chat responses ✅ Integrates OpenAI GPT + LangChain RAG pipelines ✅ Uses Redis vector search for embedding queries ✅ Modular microservice architecture for scalability

📜 License

This project is under development. License terms will be added in a future release.