Skip to content

CONTRIBUTING.md

Latest commit

 

History

History
118 lines (86 loc) · 3.04 KB

CONTRIBUTING.md

File metadata and controls

118 lines (86 loc) · 3.04 KB
Raw

Contributing to FungalClassifier

Thank you for contributing. This document covers development setup, code standards, testing, and how to add new feature blocks or model types.

Development Setup

git clone https://github.com/your-org/fungal-classifier.git
cd fungal-classifier
conda env create -f environment.yml
conda activate fungal-classifier
pip install -e ".[dev]"

Install pre-commit hooks:

pip install pre-commit
pre-commit install

Code Standards

  • Python 3.11+
  • Formatter: ruff format (enforced by pre-commit)
  • Linter: ruff check — no warnings allowed in CI
  • Type hints: required for all public functions
  • Docstrings: NumPy-style for all public classes and functions
  • Line length: 100 characters

Run checks locally:

make lint
make format

Testing

# Fast unit tests (no slow marker)
make test-fast

# All tests including integration
make test

# Single test file
pytest tests/test_features.py -v

All PRs must pass the full test suite. Integration tests (@pytest.mark.slow) run on CI but not on every local commit.

Test coverage target: >80% for fungal_classifier/ core modules.

Adding a New Feature Block

  1. Create fungal_classifier/features/{block_name}.py
  2. Implement a build_{block_name}_matrix(paths, ...) -> pd.DataFrame function
  3. Add it to fungal_classifier/features/__init__.py
  4. Add the block handler to scripts/01_build_features.py
  5. Add default config section to configs/default.yaml
  6. Write unit tests in tests/test_features.py
  7. Add a section to docs/feature_engineering.md

Convention: All feature matrices must:

  • Have genome_id as the index name
  • Return np.float32 dtype
  • Accept a min_genome_freq parameter and filter rare features
  • Log their shape at INFO level

Adding a New Model Type

  1. Create fungal_classifier/models/{model_name}.py
  2. Implement sklearn-compatible fit(X, y) and predict(X) / predict_proba(X)
  3. Add to fungal_classifier/models/__init__.py
  4. Add a --model-type option to scripts/02_train.py if appropriate
  5. Write tests in tests/test_models.py

Pull Request Process

  1. Branch from main: git checkout -b feature/my-feature
  2. Write code and tests
  3. Run make test && make lint
  4. Update docs if you changed behaviour or added features
  5. Open a PR with a clear description of what changed and why
  6. A maintainer will review within a week

Reporting Issues

Please include:

  • Python version and OS
  • Conda environment (conda list)
  • Minimal reproducible example
  • Full error traceback

Roadmap

Planned additions:

  • features/secretome.py — signal peptide and GPI anchor features
  • features/synteny.py — gene cluster synteny features
  • models/graph_model.py — GNN over genome-scale metabolic networks
  • Multi-label classification (genomes with multiple ecological roles)
  • Active learning module for prioritising new genome sequencing
  • Integration with NCBI datasets API for auto-downloading annotation