jsonable_encoder does not use custom_encoder if Config.json_encoders is not defined.

[matrixise]

Describe the bug

With factory_boy and faker, I wanted to define a Factory, here is the code...

from datetime import datetime

import factory
import faker.utils.datetime_safe
from fastapi.encoders import jsonable_encoder
from pydantic import BaseModel


class MyModel(BaseModel):
    dt_field: datetime


class MyModelFactory(factory.Factory):
    dt_field = factory.Faker("date_time")  # should return a datetime

    class Meta:
        model = MyModel


inst = MyModelFactory.build()

assert isinstance(inst, MyModel)
assert isinstance(inst.dt_field, faker.utils.datetime_safe.datetime)
assert isinstance(inst.dt_field, datetime)

inst_encoded = jsonable_encoder(inst)
assert isinstance(inst_encoded["dt_field"], dict)
assert inst_encoded["dt_field"] != inst.dt_field.isoformat()

inst_encoded = jsonable_encoder(
    inst, custom_encoder={faker.utils.datetime_safe.datetime: lambda o: o.isoformat()}
)
assert isinstance(inst_encoded["dt_field"], str)
assert inst_encoded["dt_field"] == inst.dt_field.isoformat()

To Reproduce
Use my script

Expected behavior
custom_encoder is not used...

Environment:

• OS: Linux
• FastAPI Version [e.g. 0.3.0], get it with:

import fastapi
print(fastapi.__version__)

FastAPI: 0.42.0

• Python version, get it with:

python --version

Python: 3.8.0

Additional context
Add any other context about the problem here.

[chbndrhnns]

Can we assume this issue is fixed with the merged changed in 0.47.1?

[tiangolo]

Thanks for the report @matrixise , I think this should be fixed in the latest FastAPI versions.

Could you confirm if it works for you and if so, close the issue?

[matrixise]

I will test on tomorrow and will keep you in touch.

[tiangolo]

Did you confirm if it works? May we close this issue?

[github-actions]

Assuming the original issue was solved, it will be automatically closed now. But feel free to add more comments or create new issues.

[github-actions]

Assuming the original need was handled, this will be automatically closed now. But feel free to add more comments or create new issues or PRs.