Develop and Test PhilanthroPy

Run the test suite before you propose a change. These are the same checks that run in GitHub CI, so passing them locally means your work is ready to review. The suite guards the reliability of the donor analytics models.

The single command that runs everything below is make ci. The sections here are for running one piece at a time.

Prerequisites

Install the development dependencies:

pip install -e ".[dev]"

Running Tests Locally

The suite uses pytest, with hypothesis for property-based testing and pytest-cov for coverage.

1. High-Level Unit Tests & Coverage

Run the core suite; the 92% coverage gate in pyproject.toml is enforced automatically:

pytest tests/ --cov=philanthropy

2. Scikit-Learn API Compliance

Confirm every transformer and estimator adheres to the scikit-learn API:

pytest tests/test_sklearn_compliance.py -v

3. Property-Based Testing

Check the transformers against randomized data:

pytest tests/test_properties.py -v --hypothesis-seed=12345

4. Running Doctests

Check that the code examples in docstrings still run:

pytest philanthropy --doctest-modules