Pytest: Deselecting tests

Question

With pytest, one can mark tests using a decorator

@pytest.mark.slow
def some_slow_test():
    pass

Then, from the command line, one can tell py

Answer 1

Additionally, with the recent addition of the "-m" command line option you should be able to write:

py.test -m "not (slow or long)"

IOW, the "-m" option accepts an expression which can make use of markers as boolean values (if a marker does not exist on a test function it's value is False, if it exists, it is True).

Answer 2

Looking through the pytest code (mark.py) and further experimentation shows the following seems to work:

pytest -k "-slow -long"

(Using the --collect-only option speeds up experimentation)

Answer 3

It's also possible to stack the mark decorators.

@pytest.mark.slow
@pytest.mark.main
def test_myfunction():
    pass

I then called py.test -m "slow and main" and only the tests with both decorators were called.

py.test -m "not (slow and main)" resulted in the other tests running

Answer 4

If you are trying to run the tests from inside a python file, that is, you run your tests by calling

$ python testfile.py

which has contents

import pytest

pytest.main()

and you want to know how to pass the CLI flag in to pytest.main, the answer is:

pytest.main(["-m", "not slow"])

PS - yes, there are legitimate reasons to call tests this way. Pray you never have to learn them.

Answer 5

Is your test correctly written? Normally, tests start with test_? But anyway, it depended on what you try to filter, you can filter those tests by name using -k "not slow and not long" or by tag using -m "not slow and not long".

Run tests by keyword expressions

-k: This will run tests which contain names that match the given string expression (case-insensitive), which can include Python operators that use filenames, class names and function names as variables.

Run tests by marker expressions

-m: Will run all tests which are decorated with the @pytest.mark.slow decorator.

More documentation info can be found on:

• https://docs.pytest.org/en/latest/usage.html
• Or do: pytest --help