This directory is used for small data sets and insights found while playing with the data, reading paper or conducting a different research.

Security

Security related commits. Also has reverted security commits, hence the leading to the security problem.

Performance

Commits dealing with performance issues (e.g. running time, memory).

Commit-labeling

A dataset of commits, labeled by various language models. By comparaing different models predictions on these commits one can use active learning to identify informative commits and further imrpve the models.

Regular expresions

A data set of 110k commits related to regular expressions.

Repository description

This reposotory contain database construction on the BigQuery GitHub scehmea. It was constructed as part of

The Corrective Commit Probability Code Quality Metric

Supplementary Materials of the "The Corrective Commit Probability Code Quality Metric" paper by Idan Amit and Dror G. Feitelson.

Please cite as

@misc{amit2020corrective,
    title={The Corrective Commit Probability Code Quality Metric},
    author={Idan Amit and Dror G. Feitelson},
    year={2020},
    eprint={2007.10912},
    archivePrefix={arXiv},
    primaryClass={cs.SE}
}

It was later extended as part of

Follow Your Nose - Which Code Smells are Worth Chasing?

Supplementary Materials of the "Follow Your Nose -- Which Code Smells are Worth Chasing?" paper by Idan Amit, Nili Ben Ezra, and Dror G. Feitelson.

Please cite as

@misc{amit2021follow,
      title={Follow Your Nose -- Which Code Smells are Worth Chasing?}, 
      author={Idan Amit and Nili Ben Ezra and Dror G. Feitelson},
      year={2021},
      eprint={2103.01861},
      archivePrefix={arXiv},
      primaryClass={cs.SE}
}

See here the linguistic commit classification

See here the analysis utilities

See here the general database construction

Research used the model of commit d15d54e Repository will keep advancing.

Versions

Live version is updating at https://github.com/evidencebp/sweets

Repository will keep advancing.