Fields and validation¶

Mongoz field factories create Pydantic field definitions with MongoDB-oriented metadata. Python annotations remain the source of the model's public value types.

from datetime import datetime
from mongoz import Boolean, DateTime, Document, Integer, String

class Event(Document):
    title: str = String(min_length=1, max_length=120)
    capacity: int = Integer(minimum=1)
    published: bool = Boolean(default=False)
    created_at: datetime = DateTime(auto_now_add=True)

Validation occurs during model construction and modeled persistence. Field validators and Pydantic model validators are not bypassed merely because data originated in MongoDB. Invalid modeled data raises the applicable Pydantic or Mongoz-owned error.

Common field options¶

Option	Purpose
`default`	Supply a value when the caller omits the field. A callable is invoked per instance.
`null`	Allow a stored `None` where supported.
`read_only`	Document model intent; it is not an authorization boundary.
`index`, `unique`	Contribute index metadata; creation still requires explicit reconciliation.
`alias`	Map a public/model field to a stored key.
`min_length`, `max_length`	Bound strings, arrays, or binary values where supported.
`minimum`, `maximum`	Bound numeric values.

Applications must use dedicated input models or allowlists for request authorization. A field marked read_only does not stop an untrusted caller from supplying a key to a raw query or native driver operation.

See the complete field reference and BSON and serialization.