SingleStore Knowledge Graph Backend API

This is the backend API server for the SingleStore Knowledge Graph application. It can be deployed both locally and on Replit or Railway.

Local Development

Prerequisites

• Python 3.12.9
• SingleStore database access
• OpenAI API key
• Google Gemini API key

Setup

1. Create a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

2. Install dependencies:

pip install -r requirements.txt

3. Create a .env file in the root directory with:

OPENAI_API_KEY=your_openai_key
GEMINI_API_KEY=your_gemini_key
SINGLESTORE_HOST=your_host
SINGLESTORE_PORT=your_port
SINGLESTORE_USER=your_user
SINGLESTORE_PASSWORD=your_password
SINGLESTORE_DATABASE=your_database

4. Run the server locally:

./start_backend_services.sh  # On Windows: python main.py

The server will start at http://localhost:8000

Replit Deployment

1. Create a new Python repl on Replit
2. Upload all the files from this directory
3. Add environment variables in Replit Secrets
4. Click "Run" to deploy

The server will be available at your Replit URL (e.g., https://your-repl-name.username.repl.co)

Railway Deployment

1. Fork this repository
2. Create a new project on Railway
3. Connect your GitHub repository
4. Add the following environment variables in Railway:

   OPENAI_API_KEY=your_openai_key
   GEMINI_API_KEY=your_gemini_key
   SINGLESTORE_HOST=your_host
   SINGLESTORE_PORT=your_port
   SINGLESTORE_USER=your_user
   SINGLESTORE_PASSWORD=your_password
   SINGLESTORE_DATABASE=your_database
   API_KEY=your_secure_api_key

5. Deploy! Railway will automatically build and deploy your application

The server will be available at your Railway URL (e.g., https://your-app-name.railway.app)

Project Structure

.
├── .replit                 # Replit configuration
├── api/                    # API endpoints
├── config/                 # Configuration management
├── core/                   # Core business logic
├── db/                     # Database operations
├── processors/             # Document processors
├── search/                 # Search functionality
├── tasks/                  # Background tasks
├── utils/                  # Utility functions
├── main.py                # FastAPI application
├── requirements.txt       # Python dependencies
└── start_backend_services.sh  # Local startup script

API Endpoints

• /kbdata: Knowledge base statistics
• /config: System configuration
• /upload: Document upload
• /search: Document search
• /graph: Knowledge graph
• /task-status: Processing status
• /cancel-processing: Cancel tasks

API Documentation

Visit /docs or /redoc for complete API documentation.

Environment Variables

See .env.example for all required environment variables.

Notes

• Free tier Replit has limitations (512MB RAM, 500MB storage)
• Server sleeps after inactivity on free tier
• Consider "Always On" feature for production use