error[invalid-type-form]: Function calls are not allowed in type expressions with Pydantic conlist

[alex-goswag]

Summary

I can use Field(min_length=1) as a work around or maybe that is the better way to do it but I guess conlist should still be supported by ty ?

Example

# !pip install pydantic==2.11.4
# !pip install ty==0.0.1a1

from pydantic import BaseModel, Field, conlist


class OrderLine(BaseModel):
    product_id: int
    quantity: int


class Order(BaseModel):
    order_lines: conlist(item_type=OrderLine, min_length=1)
    field_order_lines: list[OrderLine] = Field(
        min_length=1,
    )


if __name__ == "__main__":
    order = Order(
        order_lines=[OrderLine(product_id=1, quantity=10)],
        field_order_lines=[OrderLine(product_id=1, quantity=10)],
    )
    print(order)

Ty

WARN ty is pre-release software and not ready for production use. Expect to encounter bugs, missing features, and fatal errors.
error[invalid-type-form]: Function calls are not allowed in type expressions
  --> py.py:13:18
   |
12 | class Order(BaseModel):
13 |     order_lines: conlist(item_type=OrderLine, min_length=1)
   |                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
14 |     field_order_lines: list[OrderLine] = Field(
15 |         min_length=1,
   |
info: rule `invalid-type-form` is enabled by default

Found 1 diagnostic

Mypy

pip install mypy==1.15.0

py.py:13: error: Invalid type comment or annotation  [valid-type]
py.py:13: note: Suggestion: use conlist[...] instead of conlist(...)
Found 1 error in 1 file (checked 1 source file)

Version

ty 0.0.1-alpha.1 (12f466e 2025-05-13)

[sharkdp]

Thank you very much for reporting this.

Call expressions are indeed not valid in type annotations, so mypy and ty are correct here. As @AlexWaygood suggested, we could probably add a hint to the invalid-type-form diagnostic to help users.

Apart from that, it's not completely clear how we could improve the behavior here?

[carljm]

Is the use of conlist as a function in type annotations something that is recommended and supported by pydantic? This seems like a problem that should be addressed in pydantic, because it's very clear in the typing specification that calls in annotations are not supported, and other type checkers also prohibit it.

Happy to continue the discussion here and reopen if a concrete improvement suggestion arises from discussion, but in order to keep the open issues actionable, I'm going to close this as "not planned" for now.

[sharkdp]

As @AlexWaygood suggested, we could probably add a hint to the invalid-type-form diagnostic to help users.

That has been added in astral-sh/ruff#18104.

[jc-louis]

Another case which is recommended by pydantic

from pydantic import create_model, BaseModel


DynamicFoobarModel = create_model('DynamicFoobarModel', foo=str, bar=(int, 123))

class StaticFoobarModel(BaseModel):
    foo: str
    bar: int = 123

# DynamicFoobarModel and StaticFoobarModel are equivalent
# https://docs.pydantic.dev/latest/concepts/models/#dynamic-model-creation

type mytypestatic = list[StaticFoobarModel] # OK
type mytypedynamic = list[DynamicFoobarModel] # Variable of type `type[BaseModel]` is not allowed in a type expression


class MyNestedDynamic(BaseModel):
    inner: DynamicFoobarModel # Variable of type `type[BaseModel]` is not allowed in a type expression

class MyNestedStatic(BaseModel):
    inner: StaticFoobarModel

# Both work fine
print(MyNestedDynamic.model_json_schema())
print(MyNestedStatic.model_json_schema())

ty output (which is right according to spec)

error[invalid-type-form] Variable of type `type[BaseModel]` is not allowed in a type expression
error[invalid-type-form] Variable of type `type[BaseModel]` is not allowed in a type expression

[MichaReiser]

For others interested. Here's where this is recommended in the pydantic docs.

[MichaReiser]

It seems the con functions are discouraged, see pydantic/pydantic#9011 (comment)

[OliverFarren]

I found this a little tricky to piece together myself for handling constrained lists.

This did not work for me.

This worked for me and mypy

from typing import Annotated
from pydantic import BaseModel, conlist


TEXT_EMBEDDING_ADA_002_DIMENSIONALITY = 1536

class Model(BaseModel):
    embeddings: Annotated[
        list[float] | None,
        conlist(
            float,
            min_length=TEXT_EMBEDDING_ADA_002_DIMENSIONALITY,
            max_length=TEXT_EMBEDDING_ADA_002_DIMENSIONALITY,
        ),
    ] = None