-
Notifications
You must be signed in to change notification settings - Fork 14.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
forbid extra fields in BaseModel #44306
base: main
Are you sure you want to change the base?
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes I think it's a good idea.
If a user PATCH "satte"
instead of "state"
this will give him an explicit message instead of accepting the PATCH request and basically ignoring the field, resulting in a silent error for the user.
It's worth fixing test and getting it ready to merge 👍
Note: Add this to the breaking change of the API tracked here: (We do not create the newsfragments yet, format is not settled for those). |
…AIP-84/model_conf_extra_forbid
7cf2095
to
9d36d40
Compare
priority_weight_strategies: list[Any] | ||
admin_views: list[Any] | ||
menu_links: list[Any] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm not sure what types to add here.
I tried to add list[type[PriorityWeightStrategy]] to priority_weight_strategies
. But, its throwing following error when generating openapi spec.
raceback (most recent call last):
File "/opt/airflow/scripts/in_container/run_update_fastapi_api_spec.py", line 38, in <module>
get_openapi(
File "/usr/local/lib/python3.9/site-packages/fastapi/openapi/utils.py", line 493, in get_openapi
field_mapping, definitions = get_definitions(
File "/usr/local/lib/python3.9/site-packages/fastapi/_compat.py", line 231, in get_definitions
field_mapping, definitions = schema_generator.generate_definitions(
File "/usr/local/lib/python3.9/site-packages/pydantic/json_schema.py", line 361, in generate_definitions
self.generate_inner(schema)
File "/usr/local/lib/python3.9/site-packages/pydantic/json_schema.py", line 441, in generate_inner
if 'ref' in schema:
File "/usr/local/lib/python3.9/_collections_abc.py", line 769, in __contains__
self[key]
File "/usr/local/lib/python3.9/site-packages/pydantic/_internal/_mock_val_ser.py", line 41, in __getitem__
return self._get_built().__getitem__(key)
File "/usr/local/lib/python3.9/site-packages/pydantic/_internal/_mock_val_ser.py", line 58, in _get_built
raise PydanticUserError(self._error_message, code=self._code)
pydantic.errors.PydanticUserError: `TypeAdapter[typing.Annotated[airflow.api_fastapi.core_api.datamodels.plugins.PluginCollectionResponse, FieldInfo(annotation=PluginCollectionResponse, required=True)]]` is not fully defined; you should define `typing.Annotated[airflow.api_fastapi.core_api.datamodels.plugins.PluginCollectionResponse, FieldInfo(annotation=PluginCollectionResponse, required=True)]` and all referenced types, then call `.rebuild()` on the instance.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes I think Generic inside Generic are not well supported T[K[V]]
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This new fields shouldn't be returned by the API, they are not needed at the moment.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We can either pop them manually before using the model (not great because we would need to extend that to every endpoint constructing a XXXXResponse object that has a richer dict as a source), or allow extra for XXXXResponse
and only apply it to XXXXBody
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Taking a closer look, I think that there are two cases:
- Response models, those are instantiated by the server, there we want to be able to provide 'more' for the model to select only what should be returned. (This way we don't need to manually pop keys or anything else every time we are constructing a response)
- Payload / Body datamodels, those are constructed by the client and send to the application, they should be strict and not allow unknown field.
I would recommend creating a different baseclass for Bodies. BaseBody
or something else, with the model_config = ConfigDict(from_attributes=True, extra="forbid")
and use that for all body models.
priority_weight_strategies: list[Any] | ||
admin_views: list[Any] | ||
menu_links: list[Any] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yes I think Generic inside Generic are not well supported T[K[V]]
priority_weight_strategies: list[Any] | ||
admin_views: list[Any] | ||
menu_links: list[Any] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This new fields shouldn't be returned by the API, they are not needed at the moment.
priority_weight_strategies: list[Any] | ||
admin_views: list[Any] | ||
menu_links: list[Any] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We can either pop them manually before using the model (not great because we would need to extend that to every endpoint constructing a XXXXResponse object that has a richer dict as a source), or allow extra for XXXXResponse
and only apply it to XXXXBody
Note: As PR #45312 has been merged, the code formatting rules have changed for new UI. Please rebase and re-run pre-commit checks to ensure that formatting in folder airflow/ui is adjusted. |
As of now, we allow any fields to be in the Data Models. This PR ensures that requests have the exact required fields.
As of now, if we pass the wrong fields, the data models simply ignore them.
@pierrejeambrun , do you think this is a good idea?