Testing

Testing Guidelines

This page describes testing recommendations for the entropic_measurement library.

Types of tests

• Unit tests: Each function (e.g. shannon_entropy) tested in isolation
• Integration tests: Verifies end-to-end workflows (reading, config, calculation)
• Performance tests: Benchmarking time, memory usage for large data

Recommended structure

• Place tests in a tests/ directory
• Use pytest (or unittest)
  • e.g., pytest tests/

Example unit test

from entropic_measurement import shannon_entropy

def test_shannon_entropy_binary():
    data = [0.5, 0.5]
    assert shannon_entropy(data) == 1.0

Tips

• Include edge cases (zero probability, degenerate distributions, etc.)
• Add compatibility tests for multiple Python versions if needed
• Test with both real and synthetic data samples
• Use simple test datasets for sanity checks

Coverage & Automation

• Use a tool like coverage to check test coverage (e.g., coverage run -m pytest)
• Automate tests (e.g., with GitHub Actions or another CI)