Running and selecting tests

In a suite of thousands of tests, running exactly the ones you're working on is what keeps the feedback loop fast.

The pytest command line is your daily tool:

Node ids select precisely: pytest tests/test_cart.py, pytest tests/test_cart.py::test_total, pytest tests/test_cart.py::TestDiscounts::test_zero_tax.
-k "expr" selects by name substring with and, or, not: pytest -k "cart and not slow".
-m marker selects by marker (see markers).
-x stops at the first failure; --maxfail=3 after three.
--lf (last failed) reruns only the tests that failed last time; --ff runs them first, then the rest. pytest remembers results in the .pytest_cache directory.
-v shows each test; -q is quieter; -r chooses the short summary (-ra lists everything that didn't pass).
-s disables output capturing so print() shows up immediately; normally output is shown only for failing tests.
--pdb drops into the debugger on failure; -l shows local variables in tracebacks.
--durations=10 lists the slowest tests.

pytest --collect-only -q lists what would run without running it, handy to check that discovery and -k expressions do what you expect.

A typical loop while fixing a bug: pytest -x --lf, then the full suite before committing.

Example

# Run one file, one test, one method
pytest tests/test_cart.py
pytest "tests/test_cart.py::test_total_includes_tax"
pytest "tests/test_cart.py::TestDiscounts::test_zero_tax"

# Select by name expression, stop at first failure
pytest -k "cart and not slow" -x

# Rerun only last failures, show prints, debug on failure
pytest --lf -s --pdb

# See what would run, and the 5 slowest tests
pytest --collect-only -q
pytest --durations=5

Remember

Node ids (file::class::test) select exact tests
-k matches names with and/or/not; -m selects markers
-x stops early, --lf reruns failures, -s shows print output
--collect-only -q previews selection

Watch out for

Debugging with print() and not seeing any output because capture is on (use -s or look at the failure report).

Check yourself

You fixed two failing tests out of 800. What reruns just those two?

Official docs: docs.pytest.org/en/stable/how-to/usage.html

Example

Remember

Watch out for

Check yourself

Builds on

Unlocks