Pydantic¶

Pydantic offers build-in data validation and type checking, suitable for web APIs and external data handling.

2.0 update list¶

https://2pointers.medium.com/an-introduction-to-pydantic-v2-alpha-pre-release-a-massive-improvement-over-previous-version-748a1f1118ba

`model` field name¶

have reserved every field that starts with model_ for themselves.

option 1: using alias

from pydantic import BaseModel, Field

class MyModel(BaseModel):
    model_: str = Field(..., alias='model')

option 2: disable reserved words

from pydantic import BaseModel, ConfigDict

class MyModel(BaseModel):
    model_config = ConfigDict(protected_namespaces=())
    model: str = Field(..., title='The name of the model to use')

from_attributes¶

In Pydantic, the from_attributes setting in the Config class is a relatively new feature (introduced in Pydantic v2) that allows you to create a Pydantic model instance directly from attributes (i.e., keyword arguments) instead of the usual dictionary-based initialization.

By default, Pydantic models expect input data to be passed as a dictionary (or something that can be converted to one), but with from_attributes = True, you can initialize the model using individual attributes directly as you would with a regular class.

Example¶

from pydantic import BaseModel

class MyPydanticModel(BaseModel):
    id: int
    name: str

    class Config:
        from_attributes = True

Without `from_attributes` (default behavior)¶

Normally, you would instantiate the model like this:

data = {"id": 1, "name": "Alice"}
model_instance = MyPydanticModel(**data)

This works because Pydantic automatically parses the dictionary and assigns the values to the model's fields.

With `from_attributes = True`¶

With from_attributes enabled, you can instantiate the model directly with attributes:

model_instance = MyPydanticModel(id=1, name="Alice")

This works as if id and name were regular class attributes, but Pydantic will still validate them according to the model's type annotations and run any other model validation logic defined in the model.

Why would you use `from_attributes`?¶

This feature can be helpful in scenarios where:

You have objects or data sources that already provide values as individual attributes, and you want to avoid packing them into a dictionary first.
You want to simplify the instantiation process when creating models programmatically.

In this case, the from_attributes setting makes the instantiation of MyPydanticModel feel like working with a normal Python class, where attributes can be passed directly instead of a dictionary.

Summary¶

from_attributes = True allows Pydantic models to be initialized using keyword arguments (like a regular Python class) rather than needing to pass a dictionary.
It's useful when you already have data in the form of attributes, and it can simplify your model instantiation code.

Pydantic¶

2.0 update list¶

model field name¶