Add peewee playhouse.flask_utils stubs

[pylipp]

@AlexWaygood @mikeziminio seems like you have worked with peewee in the past, mind reviewing?

Source: https://github.com/coleifer/peewee/blob/master/playhouse/flask_utils.py

[Avasam]

Would it be better to maybe add playhouse\.\w+? to stubs/peewee/@tests/stubtest_allowlist.txt to ignore missing modules in playhouse rather than ignore_missing_stub = true in METADATA.toml ? (which won't reveal missing new names added in updates like #11373 ).

I'd keep partial_stub = true though, with a comment explaining that we don't provide all playhouse modules (one of the rare cases where you'd have partial_stub w/o ignore_missing_stub)

[Avasam]

Suggested a few improvements. Otherwise these changes look good.

[Avasam]

I really get the feeling this is a stubtest false-positive because of the property Model.
It looks like stubtest is seeing model_class=Model and thinks its assigning the value of the Model property (as if it was called, likely because mypy special cases properties) rather than the Model class.

@hauntsaninja thoughts?

[Avasam]

@pylipp In the meantime, I think you can simply avoid typing the model_class param in __init__ to get your PR though.
It would have been nice to get the class fully typed from the get-go, but our peewee stubs are marked as incomplete anyway.

[pylipp]

@Avasam thanks for sticking with this PR :) I'm learning a lot about writing stubs from your helpful explanations, really grateful for that and the welcoming culture.
I tried your latest suggestion in 0f2bc77, let's see how it goes

[github-actions]

According to mypy_primer, this change has no effect on the checked open source code. 🤖🎉

[Avasam]

Thanks!

[pylipp]

@Avasam I was hoping to fix a name-defined error I saw in my code base but it still persists.
I added an MWE here. Do you mind taking a look?
I suspect it could be because of the "name clash" with Model you mentioned above.