Library provides fallback enum functionality. Fully compatible with pydantic.
It can be helpful to have a fallback value for enums, so that unknown values can be handled gracefully. This is especially useful when working with external data sources that may contain values that are not known at the time of writing the code.
pip install fbenum
fbenum
provides two possible ways to use it:
- As a class mixin, which can be used to enhance any enum class
- As a pydantic validator, which can be used to validate enum values in pydantic models
Class mixin can be used to enhance any enum class with fallback functionality.
from enum import Enum
from fbenum.enum import FallbackEnum
@enum.unique
class MyEnum(enum.StrInt, FallbackEnum):
__unknown_name__ = 'MISSING'
A = 1
B = 2
assert MyEnum(1) == MyEnum.A
assert MyEnum(44).name == 'MISSING'
assert MyEnum(44).value == 44
As a pydantic validator, fbenum
can be used to validate enum values in pydantic models.
In this cases, the enum class no longer needs to be a subclass of FallbackEnum
or somehow
otherwise enhanced.
Two ways of using it are provided:
from enum import Enum
from fbenum.adapter import FallbackAdapter
@enum.unique
class MyEnum(enum.StrInt):
A = 1
B = 2
class MyModel(BaseModel):
value: Annotated[
MyEnum, FallbackAdapter(unknown_name='MISSING'),
]
model = MyModel(value=1)
assert model.value == MyEnum.A
model = MyModel(value=44)
assert model.value.name == 'MISSING'
assert model.value.value == 44
Type wrapper has restricted functionality, as it can only be used to wrap a single enum class, and it does not support any additional arguments. However, it is more concise.
from enum import Enum
from fbenum.adapter import FallbackAdapter
@enum.unique
class MyEnum(enum.StrInt):
A = 1
B = 2
class MyModel(BaseModel):
value: FallbackAdapter[MyEnum]
model = MyModel(value=1)
assert model.value == MyEnum.A
model = MyModel(value=44)
assert model.value.name == 'MISSING'
assert model.value.value == 44
from enum import Enum
from fbenum.adapter import FallbackAdapter
@enum.unique
class UserStatus(enum.StrInt):
"""User status."""
ACTIVE = 1
INACTIVE = 2
class UserRequest(BaseModel):
name: str
email: str
status: FallbackAdapter[UserStatus]
class UserResponse(BaseModel):
name: str
email: str
status: FallbackAdapter[UserStatus]
def create_user(request: UserRequest) -> UserResponse:
"""Create a user."""
# Do some stuff
resp = ...
return UserResponse.validate_model(resp)
def get_user(user_id: int) -> UserResponse:
"""Get a user."""
# Do some stuff
resp = ...
return UserResponse.validate_model(resp)
def get_users() -> list[UserResponse]:
"""Get all users."""
# Do some stuff
resp = ...
return UserResponse.validate_model(resp)
if __name__ == '__main__':
user = create_user(
UserRequest(
name='John Doe',
email='[email protected]',
status=1,
),
)
# on user cretion status always will be known
users = get_users()
# user == [
# UserResponse(
# name='John Doe',
# email='test@emaik',
# status=1,
# ),
# UserResponse(
# name='Jane Doe',
# email='test2@emaik',
# status=3, # unknown status
# ]
# on users_getting status can be unknown, so using fallback wrapper for response
# model will be a good idea to handle unknown values gracefully.