metadata

dataset_info:
  features:
    - name: status
      dtype: string
    - name: repo_name
      dtype: string
    - name: repo_url
      dtype: string
    - name: issue_id
      dtype: int64
    - name: updated_files
      dtype: string
    - name: title
      dtype: string
    - name: body
      dtype: string
    - name: issue_url
      dtype: string
    - name: pull_url
      dtype: string
    - name: before_fix_sha
      dtype: string
    - name: after_fix_sha
      dtype: string
    - name: report_datetime
      dtype: timestamp[ns, tz=UTC]
    - name: language
      dtype: string
    - name: commit_datetime
      dtype: timestamp[us, tz=UTC]
  splits:
    - name: train
      num_bytes: 39772539.62553094
      num_examples: 11761
    - name: test
      num_bytes: 30776584.756565426
      num_examples: 12021
  download_size: 23878160
  dataset_size: 70549124.38209637
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
language:
  - en
tags:
  - code
  - bugs
  - patch
size_categories:
  - 10K<n<100K

image/png

Dataset Card for Dataset Name

BeetleBox

Dataset Details

The BeetleBox dataset is a comprehensive multi-language, multi-project dataset designed for bug localization research. It includes 26,321 bugs from 29 projects, covering five major programming languages: Java, Python, C++, JavaScript, and Go. The dataset was meticulously curated to ensure accuracy, with a manual analysis revealing an incorrect ground truth rate of only 0.06%.

Dataset Description

The BeetleBox dataset is a substantial multi-language, multi-project dataset specifically compiled for bug localization research. It features 26,321 bugs from 29 projects across five widely used programming languages: Java, Python, C++, JavaScript, and Go. This dataset was carefully curated to maintain accuracy, with a manual analysis revealing an incorrect ground truth rate of only 0.06%. The dataset includes detailed information for each bug report, such as the bug's status, repository name, repository URL, issue ID, a list of files updated during the fix, the bug report's title, body, pull request URL, issue URL, SHA values before and after the fix, and the dates and times of both the bug report and the fixing commit. The dataset was gathered from popular repositories on GitHub based on their star ratings and recent updates, ensuring active and well-maintained projects. Each repository's closed issues and corresponding fixing pull requests were linked using predefined GitHub keywords, and issues linked to multiple pull requests were filtered to ensure only a single, merged pull request was included.

Dataset Statistics

Language	Train	Test
C++	3,868	4,783
Go	758	400
Java	3,369	2,270
JavaScript	1,974	3,085
Python	3,215	2,599

Key Features

Multi-Language: Covers five major programming languages.
Multi-Project: Includes data from 29 different projects.
High Accuracy: Maintains a low incorrect ground truth rate of 0.06%.
Detailed Metadata: Provides extensive details for each bug report.

Repository: [More Information Needed]
Paper [optional]: [More Information Needed]
Demo [optional]: [More Information Needed]

--> -->