📦 Model Setup

🔒 The trained model files are not included in this repository due to GitHub's file size restrictions and security considerations.

Please download the following files from our shared drive:

👉 Download model files here： https://liveswinburneeduau-my.sharepoint.com/:f:/r/personal/104725051_student_swin_edu_au/Documents/url-checker-api?csf=1&web=1&e=ePqotC

Required files:

model.pkl – Trained machine learning model

label_encoder.pkl – Label encoder for decoding prediction outputs

Once downloaded, place both files into the root of the url-checker-api/ directory like so:

url-checker-api/ │ ├── app/ │ ├── init.py
│ ├── models.py │ ├── extensions.py
│ ├── routes/ │ │ ├── init.py │ │ ├── auth_routes.py │ │ ├── history_routes.py
│ │ ├── predict_routes.py │ │ ├── report_routes.py
│ │ └── train_routes.py
│ ├── services/ │ │ ├── jwt_utils.py
│ │ └── feature_extractor.py │ └── uploads/
│ ├── config.py
├── run.py
├── model.pkl
├── label_encoder.pkl
└── README.md

The backend will fail to start or make predictions if these files are missing.

🔐 URL Threat Classification API (Flask Backend)

This is a Flask API backend that takes a URL input and uses a hybrid machine learning model to classify the URL into one of four categories:

• benign
• phishing
• malware
• defacement

---

🚀 Features

• Accepts POST requests with a URL
• Extracts structural features from the URL
• Uses a trained ML model (model.pkl) and label encoder (label_encoder.pkl)
• Returns a predicted label from 4 threat classes
• CORS-enabled for integration with frontend (e.g., React)

---

📦 Dependencies

Install required Python packages:

pip install flask flask-cors pandas joblib PyJWT
pip install flask flask-cors flask-sqlalchemy pymysql werkzeug pyjwt

Start the Flask server

In the terminal:

python run.py

If successful, you’ll see output like:

• Running on http://127.0.0.1:5050/ (Press CTRL+C to quit)

create tables in database

python create_admin.py