pydantic-strict-partial

About

Create partial models based on the original Pydantic models.

This makes all the fields optional. This doesn't make them nullable and doesn't disable validation. The only thing it does is provide default values for those fields (None by default), so you can use model.model_dump(exclude_unset=True) command to receive specified values only.

The most common use case is a PATCH request on FastAPI endpoints where you want to allow partial updates.

Installation

pydantic-strict-partial compatible with Python 3.10+ and Pydantic 2.1+.

Using pip

pip install pydantic-strict-partial

Using poetry

poetry add pydantic-strict-partial

Usage

from typing import Annotated

from annotated_types import Ge
from pydantic import BaseModel

from pydantic_strict_partial import create_partial_model


class UserSchema(BaseModel):
    name: str
    nickname: str | None
    age: Annotated[int, Ge(18)]


UserPartialUpdateSchema = create_partial_model(UserSchema)

assert UserPartialUpdateSchema(age=20).model_dump(exclude_unset=True) == {
    'age': 20
}

UserPartialUpdateSchema(name=None)  # raises ValidationError
UserPartialUpdateSchema(age=17)  # raises ValidationError

There is also possible to specify a limited list of fields to be partial:

UserPartialUpdateSchema = create_partial_model(UserSchema, 'name', 'nickname')

Known limitations

MyPy: "is not valid as a type" error

You may be faced with Variable "UserPartialUpdateSchema" is not valid as a type error. There is no good solution for that. But the next approach can be used as a workaround:

class UserPartialUpdateSchema(create_partial_model(UserSchema)):  # type: ignore[misc]
    pass

Alternatives

pydantic-partial - it makes all fields nullable and disables all validators, which is not suitable for payload validation on PATCH endpoints.