Add Support for Pydantic Discriminated Union

[wholmen]

Since Pydantic 1.9, pydantic has supported the use of discriminated unions as a way to force what union type Pydantic evaluates an object by. Pydantic models with discriminated unions raise an error, at least because the typing.Literal is not supported

Feature Request Type

• [ ] Core functionality
• [x] Alteration (enhancement/optimization) of existing feature(s)
• [ ] New behavior

Description

The following code will raise the error when it's converted to a GraphQL model: TypeError: fields cannot be resolved. Unexpected type 'typing.Literal"

class Foo(BaseModel):
    class_type: Literal["foo"] = "foo"
    foo: str

class Bar(BaseModel):
    class_type: Literal["bar"] = "bar"
    bar: str

class Model(BaseModel):
    foobar: Union[Foo, Bar] = Field(discriminator="class_type")

The point of this, is to tell pydantic explicitly if the foobar variable should be parsed as a Bar object or a Foo object. The default behavior without discriminated unions is to try Foo first, which will always succeed in this case.

Upvote & Fund

• We're using Polar.sh so you can upvote and help fund this issue.
• We receive the funding once the issue is completed & confirmed by you.
• Thank you in advance for helping prioritize & fund our backlog.

[huyz]

I'm also running into this when generating Pydantic models using https://github.com/jhnnsrs/turms