python-testing

Define a function decorated with @pytest.fixture that performs setup, uses yield to provide the resource to the test, and places teardown code after the yield. Fixture scopes like function, module, or session control how often setup runs.

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. Assert call counts with assert_called_once or assert_awaited_once for async functions.

Yes, pytest-asyncio enables async test support through the @pytest.mark.asyncio decorator. Async fixtures can also be defined with async def and yield to provide resources like async HTTP clients.

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

Target 80%+ overall code coverage with 100% on critical paths. Measure it with pytest --cov=mypackage --cov-report=term-missing to see uncovered lines, and generate HTML reports for detailed inspection.

Shared state makes tests order-dependent and causes flaky failures that are hard to debug. Each test should be independent, using fixtures to create fresh resources and tmp_path for isolated temporary files.