melevir / flake8-functions-names Goto Github PK

Magic/dunder methods such as __exit__(self, exc_type, exc_val, exc_tb) -> bool and __contains__(self, x) -> bool raise a FNE005 that is impossible to fix, and I don't want to add #noqa lines to these methods.

... even if it also has a pure word. For example, test_root_category_create should not require pureness.

from typing import TextIO

from overrides import overrides

class MyTextIO(TextIO):
    ...

    @overrides
    def writable(self) -> bool:
        return True

It would be very cool if there was an option to ignore overridden methods. There isn't much I can do about the names given to 3rd party interface methods, let alone Python stdlib.

For now, it will fail to give representation to a decorator like @foo.bar.baz, for example.

The validator should fail in such cases.

Small code example:

def __eq__(self, other: Any) -> bool:
    return str(self) == str(other)

The plugin reports FNE005 Return type of the function is bool, but the name doesn't show it.
Since these are built-in methods, there's no way to rename them. I think in such cases this error shouldn't be reported. Applies also to other methods like __lt__(), __contains__() etc.

Exceptions seem like a problem since pure functions can't raise exceptions, but most functions with data-processing verbs can.

We need a decent resolution for the problem.

Perhaps adding a rule to validate that async functions have some naming convention will be useful. For example have a* prefix. This way the caller can always determine if the function is async or not just by the name

FNE005 Return type of the function is bool, but the name doesn't shows it

class User:
    @property
    def profile_picture(self):
        return "img.jpg"

Right now the above gives the error: FNE002 The method has a @property decorator, but has a verb in it's name (profile) [flake8-functions-names] - which I think is not really warranted in this case.

I mean it is easy enough to disable for specific cases, but would it make sense to remove such words that can act as both nouns and verbs from the list? Or was it a deliberate decision anyways to include such examples as well?

I am using python 3.10 and my TypeGuard functions are triggering FNE001. Here's an example that fails:

from typing import TypeGuard


class Point:
    x: int
    y: int


class NEQuadrantPoint(Point):
    pass


def is_ne_quadrant(pnt: Point) -> TypeGuard[NEQuadrantPoint]:
    return pnt.x > 0 and pnt.y > 0

Details can be found here: https://packaging.python.org/guides/publishing-package-distribution-releases-using-github-actions-ci-cd-workflows/

@pytest.fixture
async def is_endpoint_raises_500_factory(client, webhooks_payload_factory):
    async def is_endpoint_raises_500(route, payload_kwargs):
        payload = webhooks_payload_factory(**payload_kwargs)
        resp = await client.post(route, json=payload)
        return resp.status == 500
    return is_endpoint_raises_500

Gives double FNE001 error.