Defects4J Mutant Coverage

A parallel mutant generator for Defects4J dataset projects that creates unique mutants with configurable mutation combinations and does comprehensive coverage analysis.

🚀 Features

• Parallel Processing: Generate multiple mutants simultaneously (configurable workers)
• Configurable Mutant Generation:
  • Percentage-based mutant selection (0-100%)
  • Multiple mutations per mutant (1-4 mutations)
  • Unique mutant combinations (no duplicates)
• Cross-Platform: Works on Windows, macOS, and Linux
• Reproducible Results: Deterministic mutant generation with random seeds
• Comprehensive Analysis:
  • Line coverage percentage
  • Method-level coverage details
  • Test failure analysis
  • CSV output with all metrics
• Flexible Project Selection: Process specific bugs or entire projects

Prerequisites

System Requirements

• Python 3.8 or higher
• Java 8 or higher
• Defects4J installed and in PATH

Usage

git clone https://github.com/SABR-Lab/CodeReasoning.git
cd CodeReasoning

# Generate mutants for Math bug 1
python main.py --project "Math-1" --percentage 50 --max-mutations 2

# Results are saved in:
ls ~/defects4j_mutants/

Command Line Arguments Understanding

Argument	Description	Default	Example
--project	Project(s) to process	Required	"Math-all", "Math-1", "Math-1,Math-2"
--percentage	Percentage of mutants to create (0-100)	50	--percentage 75
--max-mutations	Maximum mutations per mutant (1-4)	4	--max-mutations 2
--workers	Number of parallel workers	10	--workers 8

Project Selection Formats

• Single bug: "Math-1"
• Multiple bugs: "Math-1,Math-2,Lang-1"
• All bugs in project: "Math-all"
• Multiple projects: "Math-1,Lang-1,Time-1"

📊 Output

The generator creates the following output structure:

/tmp/mutated_codes/
├── Math_1f/                 # Original project checkout
├── Math_1_mutants/          # Mutant results directory
│   └── Math_1_mutant_coverage.csv  # Comprehensive results
├── Math_2f/
├── Math_2_mutants/
│   └── Math_2_mutant_coverage.csv
└── Math_All_Bugs_Merged.csv # Merged results for all Math bugs

CSV Output Columns

• Mutant: Unique mutant identifier
• Line Coverage %: Overall line coverage percentage
• Mutator: Mutation operator(s) applied
• Class Name: Target class
• Line Number: Modified line number(s)
• Target File: Source file path
• Test Counts: Total tests, failed tests
• Test Lists: All tests and failing tests
• Method Coverage: Line numbers covered per method