Bug Prediction System

A full-stack application that analyzes GitHub commits and predicts the likelihood of them introducing a bug. It leverages an XGBoost machine learning model trained on over 16,000 real-world Python and TypeScript commits to evaluate changes in real-time.

Features

• Commit Risk Analysis: Evaluates commits based on size, complexity, developer history, and timing features to assign a Bug Risk level (Low, Medium, High).
• Batch Processing: Supports predicting risks for multiple commits simultaneously (up to 500).
• User Authentication: Secure sign-up and login system for users.
• Search History: Authenticated users can save their repository scans and revisit their past bug prediction results.
• Leaderboard: A gamified public leaderboard ranking users by the number of repositories they have checked and bugs they've detected.

Tech Stack

Frontend (client/)

• Framework: React 19 + TypeScript + Vite
• Styling: RadianUI + Tailwind CSS + Radix UI + class-variance-authority
• Routing: React Router
• Data Visualization: Recharts
• Forms & Validation: React Hook Form + Zod

Backend (bug_prediction-system_v1/)

• Framework: FastAPI
• Machine Learning: Scikit-Learn, XGBoost, Pandas
• Database ORM: SQLAlchemy (Async)
• Model: Isotonic-calibrated XGBoost classifier

Project Structure

final-project/
├── bug_prediction-system_v1/   # FastAPI backend API and ML models
├── client/                     # React Vite frontend application
├── data/                       # Datasets used for training
├── diagram/                    # System architecture diagrams
├── docs/                       # Project documentation
└── notebook_versions/          # Jupyter notebooks for data exploration and model training

Getting Started

Prerequisites

• Node.js (v18+)
• Python (3.9+)

1. Backend Setup (Local & Hugging Face)

The backend code is actively hosted on Hugging Face Spaces: yube08/bug_prediction-system_v1.

To run the backend locally, navigate to the backend directory:

cd bug_prediction-system_v1

Install the dependencies:

pip install -r requirements.txt

Run the FastAPI server:

uvicorn main:app --reload

The local API will be available at http://localhost:8000. You can view the API documentation at http://localhost:8000/docs.

2. Frontend Setup

Navigate to the frontend directory:

cd client

Install the dependencies:

npm install

Note: By default, the frontend is configured (via client/.env) to connect to the Hugging Face deployed backend at https://yube08-bug-prediction-system-v1.hf.space. To connect to a local backend, update the VITE_SERVER_URL in the .env file to http://localhost:8000.

Run the development server:

npm run dev

The frontend application will be available at http://localhost:5173.