Use a class as a nested model, still access inheriting classes fields?

[AdrianDeAnda]

Trying to do a migration from Pydantic v1 -> v2 on a large OOP codebase however we ran in to an issue where if a class has a nested model and we pass a value that is a class where the parent class is defined as a type we don't get the fields of the inheriting class when doing model_dump like in v1 unless we specify all the classes. Example:

Pydantic v1 behavior:

Pydantic v2 behavior, no change to types:

Pydantic v2 behavior, change to types:

Is there a configuration to keep the v1 behavior on v2 without specifying all the possible types of the classes with inheritance?

[uriyyo]

You can try to use SerializeAsAny:

from pydantic import BaseModel, SerializeAsAny


class Test1(BaseModel):
    age: int


class Test2(Test1):
    name: str


class Test3(BaseModel):
    what_test: SerializeAsAny[Test1]


test_model = Test2(age=10, name="test")
model = Test3(what_test=test_model)

print(model.model_dump())

{'what_test': {'age': 10, 'name': 'test'}}

[AdrianDeAnda]

Thanks, that is exactly what I was looking for.