SLF001 allowed for tests?

[doolio]

I should preface this by saying I'm still learning. Say I have the following pattern:

from dataclasses import dataclass, field


@dataclass
class A:
    _foo: str | None = field(default=None, init=False)

    @property
    def foo(self) -> str | None:
        if self._foo is None:
            get_foo()
        return self._foo

    @foo.setter
    def foo(self, value: str | None) -> None:
        self._foo = value


@pytest.mark.parametrize(
    ("value", "expected_value"),
    [
        pytest.param("123456789", "123456789"),
        pytest.param(None, None),
    ],
)
def test_setting_foo(value: str | None, expected_value: str | None) -> None:
    a = A()
    a.foo = value

    if value is not None:
        assert a.foo == expected_value
    else:
        # Verify internal state directly.
        assert a._foo is expected_value   # <- SLF001 error on this line

I want to be able to test setting the foo property to None i.e. that it is reset as well as it being set to a non-None value. To be able to check that it has been set to None I assertthat the internal instance variable is set to None. However, ruff then raises the SLF001 error. I could use a noqa: SLF001 to suppress the error but am wondering if there is a best practice to achieve what I want. Thanks for your time and ruff.

[ntBre]

I'm not sure I have a good suggestion for this pattern in general, but another option other than noqa would be per-file-ignores, if all of your tests are in a separate file.

[doolio]

Thank you. Yes, that is a nice alternative to suppress the error. I won't mark as answered (at least not yet) as I'd like to see if others more experienced than I will chime in to suggest perhaps a better pattern to what I am trying to achieve. Thanks again.