🤖 Automated Program Repair (APR) Framework

Bug Detection and Repair: A comprehensive, machine-learning-driven framework designed to automatically detect, analyze, and repair bugs in source code. Stop debugging, start repairing!

🚀 Project Overview: The Mission

This project is a deep dive into Automated Program Repair (APR), addressing the critical need for faster, more reliable software development. Our core goals are:

• 1️⃣ Data Powerhouse: Create robust, high-quality datasets ($\text{BugNet, AoC}$) essential for training state-of-the-art bug detection and repair systems.
• 2️⃣ Algorithmic Excellence: Develop, implement, and rigorously evaluate cutting-edge algorithms for automated code repair.
• 3️⃣ Deep Insights: Provide analytical clarity into common bug patterns and effective repair strategies.

---

🏗️ Repository Blueprint: Where to Start

The project is thoughtfully structured to support research, data generation, and practical application:

Directory	🎯 Purpose	Why it's Awesome
bugnet	BugNet Dataset Generation	Scripts to create the comprehensive $\text{BugNet}$ dataset (buggy code $\leftrightarrow$ correct code pairs).
repair-pipeline	The Repair Demo	Practical, runnable demo applications showcasing our trained Python repair models in action.
aoc-dataset	AoC Dataset Source	Code used to generate the $\text{Advent of Code (AoC)}$ dataset for diverse testing.
hint	Natural Language Hint Generation	Contains the code to translate tricky bugs into user-friendly, natural language descriptions.
repair	Evaluation & Benchmarks	The frameworks used to benchmark and evaluate the performance of all implemented repair algorithms.

---

🛠️ Get Started in 30 Seconds!

Ready to fix bugs automatically? Setting up your environment is simple:

# 1. Create a virtual environment
python -m venv .venv

# 2. Activate the environment
source .venv/bin/activate

# 3. Install dependencies from the Makefile
make install

📖 Usage Guides: Explore Our Capabilities

Dive into the specific components you find most interesting!

• ✨ Execute the Repair Pipeline:

  👉 See the automatic repair models in action! Go to repair-pipeline documentation

• 📊 Explore the AoC Dataset:

  💡 Visualize and analyze one of our key testing datasets. See aoc-dataset visualization instructions

• 🧠 Generate Natural Language Hints:

  💬 Learn how the system explains why the bug exists. Start Hint Generation Analysis

• 🧪 Evaluate Repair Performance:

  📈 Run benchmarks on different repair algorithms using our datasets. Check the Evaluation Framework

• ➕ Generate the BugNet Dataset:

  💾 Learn how we built the core dataset used for training. Follow BugNet Generation Instructions

---

🌟 Project Context & Acknowledgments

This advanced program repair framework was proudly developed as a capstone project during the Intel Unnati internship program, focusing on applying cutting-edge Machine Learning techniques to solve complex Software Engineering challenges.

Core Contributors

A big thank you to the team who built this framework:

• Atharva Karval
• Samarth Patil
• Om Dalbhanjan

---

What's Next?

We welcome contributions and feedback! Have you found an exciting new repair algorithm? Open an Issue or a Pull Request!