Spam Detection API

A simple and fast Flask-based REST API that classifies incoming text messages as spam or ham (not spam) using a trained Machine Learning model built with scikit-learn.

---

Features

• RESTful API built with Flask
• Spam detection using Multinomial Naive Bayes and CountVectorizer
• Language restriction (supports only English)
• Model training script included
• Uses .venv for dependency isolation

---

File Structure

spam-detection-api/
├── app/
│   ├── __init__.py         # Flask app
│   ├── model.py            # App routes
│   ├── model.py            # Trained model function
├── models/
│   └── model.pkl           # Saved trained model
├── data/
│   └── spam.csv            # Dataset
├── train_model.py          # Train the model
├── run.py                  # Runs the Flask App
├── requirements.txt        # Project dependencies
└── README.md               # This file

---

Setup Instructions

You can use Docker to build and run this project or follow the steps to setup the project on your machine.

1. Clone the repo

git clone https://github.com/supr1yo/spam-detection-api.git
cd spam-detection-api

2. Set up the virtual environment

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

3. Install required packages

pip install -r requirements.txt

4. Add dataset (if missing)

Download the spam.csv dataset (e.g., from Kaggle) and place it in the data directory if missing.

---

Train the Model

python train_model.py

This script reads spam.csv, trains the model, and saves model.pkl inside the models/ directory.

---

Run the API

python run.py

The API will start at http://127.0.0.1:5000/

---

API Endpoint

POST /predict

Classifies a message as spam or not.

Request

{
    "message": "Free $50 steam points"
}

Response

{
    "label": "spam",
    "message": "Free $50 steam points"
}

Returns 400 Bad Request if:

• Message is missing from request body

---

Model Info

• Algorithm: Multinomial Naive Bayes
• Feature Extraction: CountVectorizer

---

License

MIT License. See LICENSE file for details.

---

Author

Built with 💙 by supr1yo