Pydantic Schemas

A Pydantic BaseModel (or RootModel) subclass may be used as the schema. The registry extracts the JSON Schema via model_json_schema() and, if no default is provided, derives one by instantiating the model with no arguments.

from pydantic import BaseModel, Field
from habarico.contrib.preferences import register_preference

class TablePreference(BaseModel):
    page_size: int = Field(default=25, ge=1)
    dense: bool = False

# Default auto-derived: {"page_size": 25, "dense": False}
register_preference("product.table", schema=TablePreference)

If the model has required fields without defaults, you must pass an explicit default (a dict or a model instance — instances are normalised via model_dump(mode="json")).