Skip to content

Usage with Pydantic

Defining models with BSON Fields

You might need to define pure Pydantic models which include BSON fields. To that end, you can use the BaseBSONModel as the base class of your Pydantic models. This class adds the JSON encoders required to handle the BSON fields.

Also, you will have to use the bson equivalent types defined in the odmantic.bson module. Those types, add a validation logic to the native types from the bson module.

Custom json_encoders with BaseBSONModel

If you want to specify additional json encoders, with a Pydantic model containing BSON fields, you will need to pass as well the ODMantic encoders (BSON_TYPES_ENCODERS).

Custom encoders example 
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
from datetime import datetime

from odmantic.bson import BSON_TYPES_ENCODERS, BaseBSONModel, ObjectId


class M(BaseBSONModel):
    id: ObjectId
    date: datetime

    model_config = {
        "json_encoders": {
            **BSON_TYPES_ENCODERS,
            datetime: lambda dt: dt.year,
        }
    }


print(M(id=ObjectId(), date=datetime.utcnow()).model_dump_json())
#> {"id": "5fa3378c8fde3766574d874d", "date": 2020}

An issue that would simplify this behavior has been opened: pydantic#2024

Accessing the underlying pydantic model

Each ODMantic Model contain a pure version of the pydantic model used to build the ODMantic Model. This Pydantic model can be accessed in the __pydantic_model__ class attribute of the ODMantic Model/EmbeddedModel.