Usage

Command-line Interface

Analyze a single Python file:

ml_smell_detector analyze path/to/your/file.py

Analyze all Python files in a directory:

ml_smell_detector analyze path/to/your/directory

Save reports to a custom output directory:

ml_smell_detector analyze path/to/your/directory --output-dir reports/

Skip specific directories during analysis:

ml_smell_detector analyze path/to/your/directory --ignore tests notebooks __pycache__

Options

path — path to a .py file or a directory to analyze.
--output-dir DIR — directory to write reports to (default: output/).
--ignore DIR [DIR ...] — directory names to skip during analysis.

Output

The tool writes two report files to the output directory (default: output/):

analysis_report.txt — human-readable report grouped by file and detector category
analysis_report.csv — machine-readable CSV with columns: Framework, Smell/Checker Name, How to Fix, Benefits, File Path, Location, Count

Use Cases

Code Review: Automatically check for common ML code smells during code reviews.
Continuous Integration: Integrate into your CI/CD pipeline to catch potential issues early.
Education: Teach best practices in ML code to students or junior developers.
Refactoring: Identify areas of improvement in existing ML codebases.
Project Onboarding: Quickly assess the quality of ML code in a new project.

CI/CD Integration

Add to your GitHub Actions workflow:

- name: Run ML smell detector
  run: |
    pip install ml-code-smell-detector
    ml_smell_detector analyze src/ --output-dir reports/
- name: Upload reports
  uses: actions/upload-artifact@v3
  with:
    name: ml-smell-reports
    path: reports/