Minecpp 0.4

A tool to mine a GitHub repository and obtain a dataset containing a list of bug-fix pairs and related information.

MineCPP

MineCPP also known as Minecraft++ is an extension of Minecraft

Introduction

MineCPP - A tool to mine a GitHub repository and obtain a dataset containing a list of bug-fix pairs and related information. The tool, with the argument -U [GitHub URL], mines the repository and provides the output project_name.csv. The schema of project_name.csv contains 17 columns and each row in it represents a potential bug-fix pair.

Getting Started

MineCPP is a python based tool. Make sure python is installed before following the Installation guide.

Installation

MineCPP can be installed with a simple pip command.

# Installation command
pip install minecpp

All the dependencies are taken care by the installation.

Usage

MineCPP comes with three optional arguments:

optional arguments:
  -h, --help  show this help message and exit  
  --version   show version number and exit  
  -u U        Provide the GitHub repo link to anlyse the repository

A GitHub URL of the repository is enough to perform analysis on the repo. Command to run it on repository is:

minecpp -u https://github.com/SET-IITGN/Minecraft

Tool's Output

The output of the tool is a project_name.csv file. The schema of the file is:

• 'Before Bug fix': Represents the code snippet containing a bug.
• 'After Bug fix': Represents the code snippet after the bug is fixed.
• 'Location': Represents the line numbers. The 'after' field represents the line number where the bug is fixed, and 'before' represents the line number where the bug was found.
• 'Bug type': Represents the type of bug obtained from LLM using the git diff between the fixed commit and the buggy commit.
• 'Commit Message': Represents the author's description of the commit.
• 'File Path': Represents the path of the file in which the change is present or the bug is fixed.
• 'Test File': Denotes whether the test file is present for the bug. Here, 1 represents that the test file is present, and 0 represents that the test file is absent.
• 'Coding Effort': Represents the effort an author makes before a bug occurs (obtained from the AST of the source code).
• 'Constructs': Represents the type of constructs in which the bug occurred.
• 'Lizard Features Buggy': Denotes the cyclomatic complexity of the buggy file.
• 'Lizard Features Fixed': Denotes the cyclomatic complexity of the bug-fix file.
• 'BLEU', 'crystalBLEU_score', 'bert_score': Represent three different algorithms that estimate the similarity between buggy and fixed code. The similarity score lies in the range 0 to 1, where 1 indicates similarity, and 0 indicates dissimilarity.

Tool's GUI

The tool also provides a GUI to explore and analyse the dataset. It provides two features

• Dataset Visualization: This feature is used to view the dataset and it is interactive.
• Quantitative Analysis: This feature is used to show the quantative analysis of Coding Effort vs Bug-Fix pairs and Similarity Score vs Bug-Fix pairs.

Configuration

Python 3.8 or above
Needs C++ 14

Contributing

Conrtibutions are accepted. The contributions will be accepted only if they are suitable for the tool.

License

Apache License