pydantic-collections

The pydantic-collections package provides BaseCollectionModel class that allows you to manipulate collections of pydantic models (and any other types supported by pydantic).

Requirements

• Python >= 3.7
• pydantic >= 1.8.2

Installation

pip install pydantic-collections

Usage

Basic usage

from datetime import datetime

from pydantic import BaseModel
from pydantic_collections import BaseCollectionModel


class User(BaseModel):
    id: int
    name: str
    birth_date: datetime


class UserCollection(BaseCollectionModel[User]):
    pass


 user_data = [
        {'id': 1, 'name': 'Bender', 'birth_date': '2010-04-01T12:59:59'},
        {'id': 2, 'name': 'Balaganov', 'birth_date': '2020-04-01T12:59:59'},
    ]

users = UserCollection(user_data)
print(users)
#> UserCollection([User(id=1, name='Bender', birth_date=datetime.datetime(2010, 4, 1, 12, 59, 59)), User(id=2, name='Balaganov', birth_date=datetime.datetime(2020, 4, 1, 12, 59, 59))])
print(users.dict())
#> [{'id': 1, 'name': 'Bender', 'birth_date': datetime.datetime(2010, 4, 1, 12, 59, 59)}, {'id': 2, 'name': 'Balaganov', 'birth_date': datetime.datetime(2020, 4, 1, 12, 59, 59)}]
print(users.json())
#> [{"id": 1, "name": "Bender", "birth_date": "2010-04-01T12:59:59"}, {"id": 2, "name": "Balaganov", "birth_date": "2020-04-01T12:59:59"}]

Strict assignment validation

By default BaseCollectionModel has a strict assignment check

...
users = UserCollection()
users.append(User(id=1, name='Bender', birth_date=datetime.utcnow()))  # OK
users.append({'id': 1, 'name': 'Bender', 'birth_date': '2010-04-01T12:59:59'})
#> pydantic.error_wrappers.ValidationError: 1 validation error for UserCollection
#> __root__ -> 2
#>  instance of User expected (type=type_error.arbitrary_type; expected_arbitrary_type=User)

This behavior can be changed via Model Config

...
class UserCollection(BaseCollectionModel[User]):
    class Config:
        validate_assignment_strict = False
        
users = UserCollection()
users.append({'id': 1, 'name': 'Bender', 'birth_date': '2010-04-01T12:59:59'})  # OK
assert users[0].__class__ is User
assert users[0].id == 1

Using as a model field

BaseCollectionModel is a subclass of BaseModel, so you can use it as a model field

...
class UserContainer(BaseModel):
    users: UserCollection = []
        
data = {
    'users': [
        {'id': 1, 'name': 'Bender', 'birth_date': '2010-04-01T12:59:59'},
        {'id': 2, 'name': 'Balaganov', 'birth_date': '2020-04-01T12:59:59'},
    ]
}

container = UserContainer(**data)
container.users.append(User(...))
...