Testability Analysis - Distribution

This is a minimal distribution of the Testability Analysis framework, designed to evaluate the testability of AI-generated code changes (PRs) in software engineering contexts.

Setup

1. Prerequisites

• Python 3.8+
• pip (package manager)

2. Create a Virtual Environment

python3 -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install Dependencies

pip install -r requirements.txt

Usage

The distribution includes scripts to set up the dataset, run the analysis pipeline, and generate reports.

1. Set Up the Dataset

You can load the actual AIDev dataset from Hugging Face or use sample data for quick testing.

For full analysis (requires internet):

python scripts/setup_aidv_dataset.py

For a quick test (uses local sample data):

python scripts/setup_aidv_dataset.py --use-sample

2. Run the Analysis Pipeline

Execute the complete testability analysis on the loaded dataset.

python scripts/run_analysis_pipeline.py

3. Generate Reports & Visualizations

After the pipeline completes, you can generate markdown reports and visualizations.

Generate Reports:

python scripts/generate_reports.py

Generate Visualizations:

python scripts/generate_comprehensive_visualizations.py

Directory Structure

• src/: Core application logic and components.
• scripts/: Execution scripts for data setup, pipeline runs, and reporting.
• config/: Configuration files for industry standards.
• data/: (Generated) Raw and processed dataset files.
• reports/: (Generated) Detailed analysis and compliance reports.
• visualizations/: (Generated) Charts and visual metrics.

License

This project is licensed under the MIT License - see the LICENSE file for details.