📰 Fake News Detector API

Detect fake news in seconds with a robust ML ensemble model—lightning-fast predictions, reliably hosted on Railway.

A production‑ready FastAPI microservice that predicts whether a news article is Fake or Real using an ensemble ML model (TF‑IDF + VotingClassifier). Fully Dockerised, CI‑ready, and live on Railway.

---

📑 Table of Contents

1. Features
2. Live Demo
3. Tech Stack
4. Project Structure
5. Local Setup
6. Docker Usage
7. API Reference
8. Testing
9. Environment Variables
10. License
11. Author
12. Contributing

---

✨ Features

• 🧠 Ensemble ML pipeline (TF‑IDF + VotingClassifier)
• ⚡ FastAPI backend with automatic Swagger UI (/docs)
• 🐳 Docker‑first design — build once, run anywhere
• 🧵 CI/CD ready with GitHub Actions and automated Codecov reports
• 🛡️ Typed request/response models & custom exception handling
• 🌐 CORS enabled for easy front‑end integration

---

🚀 Live Demo

Base URL: https://fake-news-detector-co.up.railway.app

$ curl https://fake-news-detector-co.up.railway.app/
{"message":"Fake News Detector API is up and running."}

Open /docs for an interactive Swagger UI.

Use the Swagger interface to test endpoints, send JSON payloads, and view real-time predictions

---

⚙️ Tech Stack

• FastAPI & Uvicorn
• scikit‑learn & joblib
• Python 3.11
• Docker & Railway
• Pytest + Codecov

---

📁 Project Structure

.
├── app/
│   ├── main.py              # FastAPI entry‑point
│   ├── predictor.py         # Prediction logic
│   ├── logging_config.py    # Centralised logger
│   ├── exceptions.py        # Custom exceptions
│   ├── schemas.py           # Pydantic models
│   └── utils/
│       └── loader.py        # Load model & vectorizer
├── models/
│   ├── ensemble_voting_model.pkl
│   └── tfidf_vectorizer.pkl
├── tests/                   # Unit & integration tests
├── Dockerfile
├── requirements.txt
├── .dockerignore
├── .github/                 # CI workflow
└── README.md

---

🔧 Local Setup

# 1. Clone
git clone https://github.com/XavierRolex/fake-news-detector.git
cd fake-news-detector

# 2. (Optional) Create & activate virtual env
python -m venv venv && source venv/bin/activate

# 3. Install deps
pip install -r requirements.txt

# 4. Create a .env file (see below)
echo "MODEL_PATH=models/ensemble_voting_model.pkl" >> .env
echo "VECTORIZER_PATH=models/tfidf_vectorizer.pkl" >> .env

# 5. Run the API
uvicorn app.main:app --reload

Visit http://127.0.0.1:8000/docs for Swagger UI.

---

🐳 Docker Usage

# Build image
docker build -t fake-news-api .

# Run container
docker run -p 8000:8000 fake-news-api

---

📡 API Reference

GET /

Health check
Response

{ "message": "Fake News Detector API is up and running." }

POST /predict

Predict if text is Fake or Real
Request

{ "text": "NASA confirms water on the Moon's surface." }

Response

{ "label": "Fake", "probability": 0.1099 }

---

🧪 Testing

pytest --cov=app

Coverage reports are uploaded automatically to Codecov via GitHub Actions.

---

📄 Environment Variables

Key	Description	Default
MODEL_PATH	Path to the trained model	models/ensemble_voting_model.pkl
VECTORIZER_PATH	Path to TF‑IDF vectorizer	models/tfidf_vectorizer.pkl
LOG_LEVEL	Logging level	INFO

Note: On Railway, these are configured via the Variables tab.

---

📝 License

Distributed under the MIT License. See LICENSE for details.

---

👤 Author

Nafis Anzum

• ✉️ [email protected]
• 🔗 github.com/XavierRolex
• 💼 LinkedIn

---

🌟 Contributing

Contributions are welcome! Please open an issue first to discuss your ideas, or submit a pull request.

---

© 2025 Nafis Anzum – All rights reserved.