`InnerShadowEffect` `required` list is missing the `type` property

[DavidVanKampen-eaton]

The InnerShadowEffect does not have type listed as a required property, which can create a problem when you run this schema through the datamodel-code-generator tool. Documentation for that tool can be found here - https://datamodel-code-generator.koxudaxi.dev/.

generated code without type set as required:

class InnerShadowEffect(BaseShadowEffect):
    type: Literal["INNER_SHADOW"] | None = Field(
        None,
        description="A string literal representing the effect's type. Always check the type before reading other properties.",
    )

generated code with type set as required:

class InnerShadowEffect(BaseShadowEffect):
    type: Literal["INNER_SHADOW"] = Field(
        ...,
        description="A string literal representing the effect's type. Always check the type before reading other properties.",
    )

The key differences - no Union operator with None (| None), and the default is ... instead of None.

The code without the fix cannot be imported - it produces the following error:

pydantic.errors.PydanticUserError: Model 'InnerShadowEffect' needs field 'type' to be of type `Literal`

With the fix, the import works fine.

type should be required, because it is used as the discriminator here:

discriminator:
        propertyName: type
        mapping:
          DROP_SHADOW: "#/components/schemas/DropShadowEffect"
          INNER_SHADOW: "#/components/schemas/InnerShadowEffect"
          LAYER_BLUR: "#/components/schemas/BlurEffect"
          BACKGROUND_BLUR: "#/components/schemas/BlurEffect"
          TEXTURE: "#/components/schemas/TextureEffect"
          NOISE: "#/components/schemas/NoiseEffect"

It appears all other items in that mapping list DO have type set as required.

Reproducing the Error

Here is a pyproject.toml for running the generator:

[tool.datamodel-codegen]
url = "https://github.com/figma/rest-api-spec/raw/refs/tags/v0.35.0/openapi/openapi.yaml"
input-file-type = "openapi"
output = "models.py"
output-model-type = "pydantic_v2.BaseModel"
formatters = ["isort", "ruff-format"]
enum-field-as-literal = "all"

Note that the enum-field-as-literal flag set to all.

With the above, run datamodel-codegen, then just run import models in a python REPL to see if an error occurs or not.