Testing

Library	Use
pytest	Test framework: fixtures, parametrization, plugins, assertion rewriting.
pytest-cov / coverage	Coverage measurement and reporting, including branch coverage.
Hypothesis	Property-based testing; generates inputs and shrinks failing cases to a minimal example.
pytest-asyncio / anyio pytest plugin	Async test support.
Testcontainers	Runs real service dependencies in containers for integration tests.
syrupy	Snapshot testing for structured or generated output.
time-machine, freezegun	Deterministic control of the current time.
pytest-benchmark	Timing and regression detection for performance-sensitive functions.
respx, responses, `httpx.MockTransport`	HTTP-level mocking without patching internals.

Common practices:

Parametrize instead of looping inside a test, so each case is reported individually.
Use tmp_path and other builtin fixtures rather than writing to fixed paths.
Property-based tests for round trips and invariants: encode/decode, serialize/parse, transform/inverse, sums that must be preserved.

@given(st.lists(st.integers()))
def test_sort_is_a_permutation(xs):
    assert sorted(sorted(xs)) == sorted(xs)
    assert Counter(sorted(xs)) == Counter(xs)

Test against real dependencies where feasible; mocking an ORM or driver tests the mock rather than the query.
Coverage is a threshold, not an objective; branch coverage is more informative than line coverage.