Software Engineering Lab of Beijing Institute of Technology (BIT) BITSE-Lab

The Software Engineering Lab of Beijing Institute of Technology (BITSE-Lab) maintains an open-source organization on GitHub, aiming to systematically present the lab’s research and engineering outcomes in the field of Software Engineering and Open Source Software (OSS), while providing a unified, open, and reproducible collaboration platform for research and teaching activities.

🔗 GitHub Organization: 👉 https://github.com/bitse-lab

This organization serves as a centralized hub for:

• 📦 Research code and experimental artifacts
• 🧪 Prototypes and proof-of-concept systems
• 🛠️ Research tools and engineering implementations

By leveraging open-source practices, BITSE-Lab actively promotes:

• ✅ Research reproducibility
• 🔁 Experimental replication and data sharing
• 🤝 Community collaboration and knowledge dissemination

https://github.com/bitse-lab/awesome-git-commit

Evaluating Generated Commit Messages with Large Language Models

This project investigates the problem of automated evaluation of generated commit messages. Traditional reference-based metrics such as BLEU, ROUGE-L, and METEOR exhibit significant limitations in this task due to the non one-to-one mapping between code changes and commit messages.

We systematically explore:

• Large Language Models (LLMs) as automatic evaluators
• Prompting strategies including Few-shot learning and Chain-of-Thought reasoning
• Evaluation robustness, fairness, and reproducibility

Our results show that LLM-based evaluators can achieve near-human-level evaluation quality, significantly outperforming conventional automatic metrics.

https://github.com/bitse-lab/conventional-commit-classification

A First Look at Conventional Commits Classification

This repository presents an empirical study on the Conventional Commits Specification (CCS) and its adoption in open-source projects.

Key contributions include:

• An empirical analysis of CCS usage in real-world repositories
• Identification of common misclassification patterns
• A refined and less overlapping definition set for commit types
• An automated approach for conventional commit classification

📦 The repository contains all datasets, models, and experimental code used in the study.

https://github.com/bitse-lab/python-supply-chain

Python Software Supply Chain Construction and Visualization Tool

This project is an open-source tool designed for Python developers, security researchers, and system administrators, focusing on:

• Python dependency resolution
• Software supply chain construction
• Vulnerability discovery and analysis

Key features:

• Accurate supply chain construction without installing packages
• Optimized pip dependency resolution algorithm
• Integrated CVE vulnerability lookup
• Interactive visualization through a front-end interface

https://github.com/bitse-lab/OSSMirror

This repository supports research on open-source ecosystem analysis, mirroring, and large-scale data collection, providing infrastructure for reproducible OSS experiments.

https://github.com/bitse-lab/OSS_Health

OSS Health Reporter is a system designed to evaluate and report the overall health of Open Source Software (OSS) projects through a comprehensive, multi-dimensional analysis framework.

The system assesses OSS project health from three complementary dimensions:

• 🧩 Software Dimension – code quality, evolution, and technical sustainability
• 👥 Community Dimension – contributor activity, collaboration, and project vitality
• 📈 Market Dimension – adoption, popularity, and ecosystem impact

To ensure objectivity and robustness, OSS Health Reporter employs:

• Entropy Weight Method to automatically determine the relative importance of each metric
• GPA-based aggregation model to compute a unified, interpretable health score

The BITSE-Lab GitHub organization also supports:

• 📘 Open source software engineering courses
• 🧑‍🎓 Undergraduate and graduate research training
• 🧪 Experimental replication and benchmarking
• 🤝 Academic and industrial collaboration

Students and researchers are welcome to explore, reproduce, and build upon our work.

We welcome contributions in the form of:

• Issues and pull requests
• Experimental reproduction and feedback
• Research collaboration
• Teaching and educational use

Please feel free to contact us through GitHub.

Each repository follows its own declared open-source license unless otherwise specified.