python-testing

Define a function decorated with @pytest.fixture that performs setup, then uses yield to hand the resource to the test, with teardown code after the yield. Fixture scope can be set to function, module, or session to control how often it runs.

Use @pytest.mark.parametrize with a list of input and expected value tuples above the test function. The test runs once per parameter set, and you can add ids for readable test names in output.

Yes, pytest supports async tests through the pytest-asyncio plugin. Mark tests with @pytest.mark.asyncio and use async fixtures to test coroutines, and use assert_awaited_once when mocking async calls.

Use unittest.mock.patch as a decorator targeting the import path of the external call, then set return_value or side_effect on the mock. Autospec helps catch API misuse by validating calls against the real object's signature.

Marking slow tests with @pytest.mark.slow lets you skip them during fast feedback loops using pytest -m "not slow". Markers must be registered in pytest.ini or pyproject.toml to avoid warnings with --strict-markers.

This Skill recommends 80%+ overall code coverage with 100% on critical paths, measured via pytest --cov with term-missing and HTML reports. Coverage alone does not guarantee quality, so combine it with behavior-focused assertions.