feat: add ignore_validators flag

ADR-007 commented 8 months ago

Take Annotated into account when ignore_validators=False

I believe this field ignore_validators should be False by default. This is a for this issue

class SomethingListWithAnnotatedMinItems(PartialModelMixin, pydantic.BaseModel):
    items: Annotated[list[Something], MinLen(1)]

Before:
# SomethingListWithAnnotatedMinItems.as_partial() -> items: Optional[ist[Something]]

Now:
# SomethingListWithAnnotatedMinItems.as_partial() -> items: Optional[Annotated[list[Something], MinLen(1)]]

ddanier commented 7 months ago

My initial feeling would be this is a great addition. Need to play with this a little bit :)

ddanier commented 6 months ago

@ADR-007 Sorry for the delayed reply.

I see two major issues with the PR:

This only works for pydantic v2, while the library still supports v1 as well. Also annotated-types is not installed with pydantic v1, which would break the tests completely.
You are missing support for field validators (https://docs.pydantic.dev/latest/concepts/validators/#field-validators) and model validators (https://docs.pydantic.dev/latest/concepts/validators/#model-validators). This means the PR only handles part of the validation.

I would argue that having this PR merged would only increase the confusion around validators at the moment. We need to have a complete solution handling those.

Would you be willing to extend the PR to also contain this?

ADR-007 commented 6 months ago

Thank you for that detailed code review!

It seems to be much easier in my case to support Pydantic v2 only, as I described it here. We can close this PR :)

team23 / pydantic-partial

feat: add ignore_validators flag #28