Explanation of the title:

Suppose we have the following setup of pydantic models:

class ParametersA(BaseModel): param_a: str = Field( default=..., allow_mutation=False, exclude=False, description="some description.", extra={ "additional_bool": False } ) param_b: bool = Field( default=False, allow_mutation=True, exclude=False, description="some description.", extra={ "additional_bool": False } ) # many more parameters of all types with additional constaints (le, gt, max_items,...)
class ParametersB(BaseModel): param_c: float = Field( default=-22.0, exclude=False, allow_mutation=True, description="some floaty description", gt=-180.0, le=0.0, extra={ "additional_bool": False, "unit": "degree" } ) param_d: List[str] = Field( default=["auto"], exclude=False, allow_mutation=True, min_items=1, max_items=4, unique_items=True, description="Some listy description.", extra={ "auto_mode_available": False } ) # many more parameters of all types with additional constaints (le, gt, max_items,...)

These two Models are fields in yet another BaseModel:

class FullModelX(BaseModel): parameters_A: ParametersA parameters_B: ParametersB

The question is if it would be possible to create yet another model:

class FullModelY(BaseModel): parameters_A: ParametersA parameters_B: ParametersB

which can reuse the classes ParametersA and ParametersB but with different validation limits and different meta-data? So for example, in FullModelX the default value for param_b of ParametersA might be False, while in FullModelX it should be True. Same goes for any other FieldInfo property such as exclude, gt, le and so on..

So far my only solution is to duplicate ParametersA and ParametersB every time I need some other validation constraints.

1

1 Answer

So in most cases I think inheritance as suggested by the answer from user @Maxim would be the way to go.

In my particular case pretty much all of the model fields change on or the other FieldInfo property, so inheritance would effectively mean rewriting the whole class.

I have found the following solution, which seems to fit my case a bit better. Going by my initial example of changing param_b the code for FullModelY would look something like this:

ParametersAVariant= create_model( 'ParametersAVariant', __base__=ParametersA )
ParametersAVariant.__fields__['param_b'].default=True
class FullModelY(BaseModel): parameters_A: ParametersAVariant parameters_B: ParametersB

Not sure if this make the code more readably than simple code duplication, but at least for the most part the initial model can be reused.

3

Your Answer

Sign up or log in

Sign up using Google Sign up using Facebook Sign up using Email and Password Submit

Post as a guest

Post Your Answer Discard

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.