🛡️ Persian Swear Detector

A robust and easy-to-use Python tool for detecting Persian (Farsi) offensive text using both rule-based and machine learning (ML) approaches.

🚀 Features

• Hybrid Detection: Combines rule-based and ML-based detection for high accuracy
• Confidence Scores: Provides confidence levels for predictions
• Persian Language Support: Handles Persian text preprocessing and normalization
• CLI Interface: Simple command-line interface for quick testing
• Model Persistence: Save and load trained models for fast deployment

📦 Project Structure

├── swear_detector.py         # Main detector script
├── requirements.txt          # Python dependencies
├── dataset/
│   └── dataset.json         # Labeled dataset for training (Offensive/Normal)
├── models/
│   └── model.pkl            # Trained ML model
├── Dockerfile               # Docker support
├── docker-compose.yml       # Docker Compose config
└── README.md               # Documentation

🛠️ Installation

1. Clone the repository
2. Install dependencies:

python -m venv venv
source venv/bin/activate
pip install -r requirements.txt

🚀 Usage

Run the detector:

python3 swear_detector.py

Each prediction includes:

• Original text
• Final prediction (Offensive/Normal)
• Confidence score
• ML confidence score (for offensive predictions)

📊 Dataset

The project uses a labeled dataset (dataset.json) containing:

• Offensive texts: Inappropriate or offensive content
• Normal texts: Regular, non-offensive content

🤖 Model

The system uses a hybrid approach:

1. Machine Learning: TF-IDF + Logistic Regression
2. Rule-based detection
3. Combined scoring for final prediction

🐳 Docker Support

Build and run with Docker:

docker-compose up --build