the following validator checks if there are any lingering ... values in coming data. This value is produced when data is generated by a schema initializer. a field with the is value implies that the field is required and does not have a default.
The nice thing about this validator is that checks all fields @field_validator("*") and does this validation for all the fields of all models that inherit from our ModelConfig, But this pro is also its con, it overrides/ignores all child field validator since they have already been executed by parent.
The solution is to change it from a field_validator to a model_validator a model can run multiple model_validator thus our override/ignore problem can be avoided.
Example Code Snippet Solution:
@model_validator(mode="before")
@classmethod
def check_for_ellipses(cls, data: Any) -> Any:
def any_ellipses(data: Any):
return any(
any_ellipses(value) if isinstance(value, dict) else value == "..."
for value in data.values()
)
if any_ellipses(data):
raise ValueError(
"Please replace any and/or all `...`, these field are required"
)
return data
https://github.com/equinor/everest-models/blob/b0539dc26a0f83025c3e34793b6701515b8abf0f/src/everest_models/jobs/shared/models/base_config/base.py#L24
the following validator checks if there are any lingering
...values in coming data. This value is produced when data is generated by a schema initializer. a field with the is value implies that the field is required and does not have a default.The nice thing about this validator is that checks all fields
@field_validator("*")and does this validation for all the fields of all models that inherit from ourModelConfig, But this pro is also its con, it overrides/ignores all child field validator since they have already been executed by parent.The solution is to change it from a
field_validatorto amodel_validatora model can run multiplemodel_validatorthus our override/ignore problem can be avoided.Example Code Snippet Solution: