Pydantic preprocessing field value

Question

I have data in which dates are represented as ISO strings. Example:

{"name": "John", "last_active": "2021-10-13T17:16:49-04:00"}

My Pydantic model:

from datetime import datetime

from pydantic import BaseModel

class User(BaseModel):
    name: str
    last_active: datetime

I want to achieve automatic date conversion when I do: User(**data). Pydantic actually parses datetime object from ISO string automatically, but I also need to convert these dates to UTC timezone, like:

datetime.datetime.fromisoformat(iso_str).replace(
    tzinfo=datetime.timezone.utc
)

What is the clean way of doing this with pydantic? More generally, do pydantic provide hook for preprocessing the passed data?

Answer 1

What you are looking for is validators. Here is the documentation for Pydantic Field Validators.

It's possible to write a validator that uses mode='before' for validating value before passing it to the model constructor.

from datetime import datetime

from pydantic import BaseModel, field_validator

class User(BaseModel):
    name: str
    last_active: datetime

    @field_validator('last_active', mode='before')
    def force_utc(cls, value: datetime) -> datetime:
        return datetime.fromisoformat(iso_str).replace(tzinfo=datetime.timezone.utc)

It's also possible to use typing.Annotated for defining validators.

from datetime import datetime
from typing import Annotated

from pydantic import BaseModel, BeforeValidator


def force_utc(cls, value: datetime) -> datetime:
    return datetime.fromisoformat(iso_str).replace(tzinfo=datetime.timezone.utc)


class User(BaseModel):
    name: str
    last_active: Annotated[datetime, BeforeValidator(force_utc)]