Pydantic models should not trigger too-few-public-methods

[chaoflow]

Is your feature request related to a problem? Please describe

Pylint triggers too-few-public-methods for pydantic models.

They are very similar to dataclasses which got special treatment in #3025

Describe the solution you'd like

Pylint should not emit too-few-public-methods for pydantic models or provide a mechanism so pydantic can declare this exception.

Additional context

[PCManticore]

Thanks, at this point we can probably add an option to skip too-few-public-methods for certain data style classes, such as namedtuple, pydantic and dataclasses.

[chaoflow]

That sounds great. Note that pydantic comes also with a dataclass implementation that supports validation. It might make sense to look into supporting this as well.

[NickKush]

Any updates on this issue? 🙂

[Olegt0rr]

Please set "help wanted" tag to search for someone who can fix this issue

[miketheman]

I'm game to take a crack at this, but I'm uncertain if it's acceptable to add the pydantic package to the requirements_test.txt file for this test? Is there another place I should look for how external test deps are expressed?

[Pierre-Sassoulas]

We're trying not to add packages to the tests requirements right now. I think at this point it could be a configurable option to add classes that should be ignored by this check. This way we could test with any classes even one we created specifically for the test and easily add other similar classes.

[miketheman]

... it could be a configurable option to add classes that should be ignored by this check

I like this - something like this?

[DESIGN]
exclude-too-few-public-methods=pydantic.*

?

[Pierre-Sassoulas]

Looks great ! How would we handle multiple values ? filter separated by comma, a regex ?

[miketheman]

I guess following the existing patterns, it'd be a comma-separated list of module names/prefixes.