Skip to content

Testing

This document describes the testing strategy and procedures for synth-pdb.

Testing Framework

We use pytest as our primary testing framework. Our test suite is designed to ensure the reliability and correctness of protein structure generation, biophysical calculations, and NMR predictions.

Test Organization

The tests/ directory is organized into several subdirectories based on the type of tests:

  • unit/: Tests for individual functions and classes in isolation.
  • functional/: Tests for end-to-end functionality of the CLI and core modules.
  • integration/: Tests that verify the interaction between different modules.
  • geometry/: Specialized tests for coordinate transformations and NeRF calculations.

Running Tests

To run the full test suite, execute the following command from the project root:

pytest

To run a specific test file:

pytest tests/test_generator.py

To run tests with a specific marker (if available):

pytest -m "not slow"

Coverage

We use pytest-cov to measure code coverage. Aim for at least 80% coverage on all core modules.

To run tests with a coverage report:

pytest --cov=synth_pdb --cov-report=term-missing

This will generate a summary in the terminal showing the coverage percentage for each file.

Continuous Integration (CI)

Our CI pipeline is managed through GitHub Actions. Every push and pull request to the main branch triggers a suite of automated tests across different operating systems (Linux, macOS) and Python versions (3.10, 3.11, 3.12).

The CI workflow (.github/workflows/test.yml) includes: - Static Analysis: Linting with ruff and type checking with mypy. - Test Execution: Running the full pytest suite. - Coverage Upload: Uploading coverage reports to Codecov for tracking.

Writing New Tests

When adding new features or fixing bugs, please include corresponding tests: 1. For Bugs: Create a regression test that fails without the fix and passes with it. 2. For Features: Add unit tests for the new functionality and integration tests for its usage within the system.

Please ensure all tests are well-documented and follow the existing naming conventions (test_*.py).