From Commits to Corrections: Toward Lightweight Mining of Python Bug-Fix Patterns from GitHub

Abhishek Tiwari; Emil Spangenberg; Jens Christian Toftdahl; Joakim Leed; Oliver Feldborg Hansen; Torben Worm; Mikkel Baun Kjærgaard

doi:10.1145/3796563.3796573

From Commits to Corrections: Toward Lightweight Mining of Python Bug-Fix Patterns from GitHub

Abhishek Tiwari^*
, Emil Spangenberg
, Jens Christian Toftdahl
, Joakim Leed
, Oliver Feldborg Hansen
, Torben Worm
, Mikkel Baun Kjærgaard

^*Corresponding author for this work

University of Southern Denmark

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

Software repositories often contain a lot of historical data that provides insights into the introduction, identification, and resolution of software bugs. This data, if structured, can benefit various software fields, e.g., proactive bug detection. However, due to the largely unstructured nature of this data, extracting this information is challenging and often requires natural language processing and machine learning techniques. In this work, we propose a lightweight pipeline for mining data from open-sourced Python-specific GitHub repositories to link semantic patterns between bug types and their corresponding repairs. The pipeline is divided into six stages, comprising scraping, rule-based filtering, categorization, and data visualization. With an input of approximately 1.3 million scraped commits, the pipeline yields a rich dataset comprising 1,500 commits that showcase the extracted semantic patterns and map bug types to corresponding patches.

Original language	English
Title of host publication	Proceedings of the 19th Innovations in Software Engineering Conference, ISEC 2026
Editors	Anubha Jain, Vikas Bajpai, Rupak Majumdar, Subhajit Roy, Sangharatna Godboley, Santosh Singh Rathore
Publisher	Association for Computing Machinery
Publication date	29. Apr 2026
Article number	19
ISBN (Electronic)	9798400725036
DOIs	https://doi.org/10.1145/3796563.3796573
Publication status	Published - 29. Apr 2026
Event	19th Innovations in Software Engineering Conference, ISEC 2026 - Jaipur, India Duration: 19. Feb 2026 → 21. Feb 2026

Conference

Conference	19th Innovations in Software Engineering Conference, ISEC 2026
Country/Territory	India
City	Jaipur
Period	19/02/2026 → 21/02/2026

Bibliographical note

Publisher Copyright:
© 2026 Copyright held by the owner/author(s).

Keywords

Bug Fix Patterns
GitHub
Mining Software Repositories
Python

Documents & Links

10.1145/3796563.3796573

SDU Link

Check access to the material in SDU Library's search engine

Cite this

Tiwari, A., Spangenberg, E., Toftdahl, J. C., Leed, J., Hansen, O. F., Worm, T., & Kjærgaard, M. B. (2026). From Commits to Corrections: Toward Lightweight Mining of Python Bug-Fix Patterns from GitHub. In A. Jain, V. Bajpai, R. Majumdar, S. Roy, S. Godboley, & S. S. Rathore (Eds.), Proceedings of the 19th Innovations in Software Engineering Conference, ISEC 2026 Article 19 Association for Computing Machinery. https://doi.org/10.1145/3796563.3796573

Tiwari, Abhishek ; Spangenberg, Emil ; Toftdahl, Jens Christian et al. / From Commits to Corrections : Toward Lightweight Mining of Python Bug-Fix Patterns from GitHub. Proceedings of the 19th Innovations in Software Engineering Conference, ISEC 2026. editor / Anubha Jain ; Vikas Bajpai ; Rupak Majumdar ; Subhajit Roy ; Sangharatna Godboley ; Santosh Singh Rathore. Association for Computing Machinery, 2026.

@inproceedings{c7f7522858e5419c940e2584b43c33b6,

title = "From Commits to Corrections: Toward Lightweight Mining of Python Bug-Fix Patterns from GitHub",

abstract = "Software repositories often contain a lot of historical data that provides insights into the introduction, identification, and resolution of software bugs. This data, if structured, can benefit various software fields, e.g., proactive bug detection. However, due to the largely unstructured nature of this data, extracting this information is challenging and often requires natural language processing and machine learning techniques. In this work, we propose a lightweight pipeline for mining data from open-sourced Python-specific GitHub repositories to link semantic patterns between bug types and their corresponding repairs. The pipeline is divided into six stages, comprising scraping, rule-based filtering, categorization, and data visualization. With an input of approximately 1.3 million scraped commits, the pipeline yields a rich dataset comprising 1,500 commits that showcase the extracted semantic patterns and map bug types to corresponding patches.",

keywords = "Bug Fix Patterns, GitHub, Mining Software Repositories, Python",

author = "Abhishek Tiwari and Emil Spangenberg and Toftdahl, \{Jens Christian\} and Joakim Leed and Hansen, \{Oliver Feldborg\} and Torben Worm and Kj{\ae}rgaard, \{Mikkel Baun\}",

note = "Publisher Copyright: {\textcopyright} 2026 Copyright held by the owner/author(s).; 19th Innovations in Software Engineering Conference, ISEC 2026 ; Conference date: 19-02-2026 Through 21-02-2026",

year = "2026",

month = apr,

day = "29",

doi = "10.1145/3796563.3796573",

language = "English",

editor = "Anubha Jain and Vikas Bajpai and Rupak Majumdar and Subhajit Roy and Sangharatna Godboley and Rathore, \{Santosh Singh\}",

booktitle = "Proceedings of the 19th Innovations in Software Engineering Conference, ISEC 2026",

publisher = "Association for Computing Machinery",

address = "United States",

}

Tiwari, A, Spangenberg, E, Toftdahl, JC, Leed, J, Hansen, OF, Worm, T & Kjærgaard, MB 2026, From Commits to Corrections: Toward Lightweight Mining of Python Bug-Fix Patterns from GitHub. in A Jain, V Bajpai, R Majumdar, S Roy, S Godboley & SS Rathore (eds), Proceedings of the 19th Innovations in Software Engineering Conference, ISEC 2026., 19, Association for Computing Machinery, 19th Innovations in Software Engineering Conference, ISEC 2026, Jaipur, India, 19/02/2026. https://doi.org/10.1145/3796563.3796573

From Commits to Corrections: Toward Lightweight Mining of Python Bug-Fix Patterns from GitHub. / Tiwari, Abhishek; Spangenberg, Emil; Toftdahl, Jens Christian et al.
Proceedings of the 19th Innovations in Software Engineering Conference, ISEC 2026. ed. / Anubha Jain; Vikas Bajpai; Rupak Majumdar; Subhajit Roy; Sangharatna Godboley; Santosh Singh Rathore. Association for Computing Machinery, 2026. 19.

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - From Commits to Corrections

T2 - 19th Innovations in Software Engineering Conference, ISEC 2026

AU - Tiwari, Abhishek

AU - Spangenberg, Emil

AU - Toftdahl, Jens Christian

AU - Leed, Joakim

AU - Hansen, Oliver Feldborg

AU - Worm, Torben

AU - Kjærgaard, Mikkel Baun

PY - 2026/4/29

Y1 - 2026/4/29

N2 - Software repositories often contain a lot of historical data that provides insights into the introduction, identification, and resolution of software bugs. This data, if structured, can benefit various software fields, e.g., proactive bug detection. However, due to the largely unstructured nature of this data, extracting this information is challenging and often requires natural language processing and machine learning techniques. In this work, we propose a lightweight pipeline for mining data from open-sourced Python-specific GitHub repositories to link semantic patterns between bug types and their corresponding repairs. The pipeline is divided into six stages, comprising scraping, rule-based filtering, categorization, and data visualization. With an input of approximately 1.3 million scraped commits, the pipeline yields a rich dataset comprising 1,500 commits that showcase the extracted semantic patterns and map bug types to corresponding patches.

AB - Software repositories often contain a lot of historical data that provides insights into the introduction, identification, and resolution of software bugs. This data, if structured, can benefit various software fields, e.g., proactive bug detection. However, due to the largely unstructured nature of this data, extracting this information is challenging and often requires natural language processing and machine learning techniques. In this work, we propose a lightweight pipeline for mining data from open-sourced Python-specific GitHub repositories to link semantic patterns between bug types and their corresponding repairs. The pipeline is divided into six stages, comprising scraping, rule-based filtering, categorization, and data visualization. With an input of approximately 1.3 million scraped commits, the pipeline yields a rich dataset comprising 1,500 commits that showcase the extracted semantic patterns and map bug types to corresponding patches.

KW - Bug Fix Patterns

KW - GitHub

KW - Mining Software Repositories

KW - Python

U2 - 10.1145/3796563.3796573

DO - 10.1145/3796563.3796573

M3 - Article in proceedings

AN - SCOPUS:105038872484

BT - Proceedings of the 19th Innovations in Software Engineering Conference, ISEC 2026

A2 - Jain, Anubha

A2 - Bajpai, Vikas

A2 - Majumdar, Rupak

A2 - Roy, Subhajit

A2 - Godboley, Sangharatna

A2 - Rathore, Santosh Singh

PB - Association for Computing Machinery

Y2 - 19 February 2026 through 21 February 2026

ER -

Tiwari A, Spangenberg E, Toftdahl JC, Leed J, Hansen OF, Worm T et al. From Commits to Corrections: Toward Lightweight Mining of Python Bug-Fix Patterns from GitHub. In Jain A, Bajpai V, Majumdar R, Roy S, Godboley S, Rathore SS, editors, Proceedings of the 19th Innovations in Software Engineering Conference, ISEC 2026. Association for Computing Machinery. 2026. 19 doi: 10.1145/3796563.3796573

From Commits to Corrections: Toward Lightweight Mining of Python Bug-Fix Patterns from GitHub

Abstract

Conference

Bibliographical note

Keywords

Documents & Links

SDU Link

Fingerprint

Cite this