Encoding Version History Context for Better Code Representation

This repository contains source code of research paper "Encoding Version History Context for Better Code Representation", which is submitted to MSR 2024

Structure

The structure of our source code's repository is as follows:

• mining: contains our source code to extract version history context from Github;
  • version_history.py: script for data mining
• data: contain final data set, named "SeSaMe_VersionHistory_Callgraph.vFinal.json";
• astnn_*: contain script for ASTNN with different experiment settings
• codebert_*: contain script for CodeBERT with different experiment settings
• others: contains source code for:
  • preprocess_clone.py: contains source code for data preprocessing for Code Clone Detection
  • preprocess_class.py: contains source code for data preprocessing for Code Classification
• env.yml: contains the configuration for our enviroment.

Experiments

To replicate the result:

• for Code Clone Detection, please run the following commands

bash experiment_clone.sh

• for Code Classification, please run the following commands

bash experiment_class.sh

The experiment result will be stored in the file "result.txt"

📜 Citation

If you use our tool, please cite our paper as follows:

@inproceedings{nguyen2024encodingversionhistory,
  title={Encoding Version History Context for Better Code Representation},
  author={Nguyen, Huy and Treude, Christoph and Thongtanunam, Patanamon},
  booktitle={Proceedings of the 21st International Conference on Mining Software Repositories, 2024)},
  venue={Lisbon, Portugal}
  pages={x--y},
  year={2024}
}