README.md

metadata

title: ML Prediction API
emoji: 🔮
colorFrom: green
colorTo: blue
sdk: docker
app_port: 7860
pinned: false
tags:
  - fastapi
  - xgboost
  - machine-learning
  - postgresql
license: mit

ML Prediction API

Production-ready POC for serving an XGBoost model via FastAPI with PostgreSQL logging.

Features

• FastAPI: High-performance API with automatic validation.
• PostgreSQL: Persistent logging of all inputs and predictions.
• XGBoost: Efficient model inference.
• Docker: Containerized for easy deployment.
• CI/CD: Automated testing and deployment via GitHub Actions.
• Authentication: Simple API Key protection.

Installation

Prerequisites

• Docker
• Python 3.10+
• PostgreSQL (if running locally without Docker Compose)

Local Setup

1. Clone the repository

   git clone <repo-url>
   cd <repo-name>
   

2. Install Dependencies

   pip install -r requirements.txt
   

3. Environment Variables Create a .env file or set variables:

   export DATABASE_URL="postgresql://user:password@localhost:5432/dbname"
   export API_KEY="your-secret-key"
   export MODEL_PATH="02_xgb_model_tuned.pkl"
   

4. Initialize Database

   python scripts/create_db.py
   

5. Run the API

   uvicorn app.main:app --reload
   

API Documentation

Once running, visit:

• Swagger UI: http://localhost:8000/docs
• ReDoc: http://localhost:8000/redoc

Endpoints

• POST /predict: Make a prediction. Requires X-API-KEY header.
• GET /health: Health check.
• GET /model/info: Model metadata. Requires X-API-KEY header.

Deployment

Hugging Face Spaces

This project is configured to deploy to Hugging Face Spaces using Docker.

1. Create a new Space on Hugging Face (Select Docker SDK).
2. Add the following Secrets in the Space settings:
   • DATABASE_URL: Connection string to your external PostgreSQL database.
   • API_KEY: Your desired API key.
3. The GitHub Action ci-cd.yml will automatically deploy changes from the main branch.

Architecture

• app/main.py: Entry point.
• app/routers: API route definitions.
• app/services: Business (ML inference).
• app/models: Database and Pydantic models.
• app/core: Configuration and DB setup.

Testing

Run tests with coverage:

pytest --cov=app tests/

GLHF !