Prevent response from "re-validating" before being returned via API

[robswc]

Hey! Have been using django-ninja for awhile and love what it's been able to do for our org!

One small problem I keep running into is when I provide context, via the model_validate method, to create the models, it works perfectly fine. However, when I return the model, it gets "re-validated" before being returned via the API, and it essentially wipes the custom context that the resolvers rely on.

things =  [
    Thing.model_validate(model_to_dict(t),
                                          context={
                                              'custom': 'stuff',
                                          }) for t in things
]

Is there a way to approach this problem?

Thanks!

[robswc]

At the moment, I just have:

response={
    200: List[Any]}

As a quick and dirty fix, but it removes the openapi schema info.

[jannekanerva]

Thanks, this worked for me also. Have you found any other way of preventing the re-validation?

[robswc]

I don't know if it is a good method, but I was able to override:

_run_root_validator = custom_validator

With a custom validator, for the schema that I needed to avoid double-validation on.

@model_validator(mode="wrap")
def custom_validator(cls, values: Any, handler: ModelWrapValidatorHandler[S], info: ValidationInfo) -> Any:

    if info.context:
       # do stuff with context here
       
    forbids_extra = cls.model_config.get("extra") == "forbid"
    should_validate_assignment = cls.model_config.get("validate_assignment", False)
    if forbids_extra or should_validate_assignment:
        handler(values)

    values = DjangoGetter(values, cls, info.context)
    return handler(values)