Type annotation requires UDFs to return a type instead of an array

[progval]

Describe the bug
Here:

datafusion-python/python/datafusion/user_defined.py

Line 36 in e0284c6

_R = TypeVar("_R", bound=pa.DataType)

we bind _R to be a subtype of pyarrow.DataType

And in these two places we require that functions return a value of type _R (ie. they should return a pyarrow datatype):

datafusion-python/python/datafusion/user_defined.py

Lines 123 to 130 in e0284c6

	def __init__(
	self,
	name: str,
	func: Callable[..., _R],
	input_fields: list[pa.Field],
	return_field: _R,
	volatility: Volatility | str,
	) -> None:

datafusion-python/python/datafusion/user_defined.py

Lines 166 to 174 in e0284c6

	@overload
	@staticmethod
	def udf(
	func: Callable[..., _R],
	input_fields: Sequence[pa.DataType | pa.Field] | pa.DataType | pa.Field,
	return_field: pa.DataType | pa.Field,
	volatility: Volatility | str,
	name: str | None = None,
	) -> ScalarUDF: ...

which seems incorrect, as functions should return arrays that contain values of that type

To Reproduce

datafusion-python ((53.0.0)) $ mypy examples/python-udf.py
examples/python-udf.py:27: error: Value of type variable "_R" of function cannot be "Array[Any]" [type-var]
Found 1 error in 1 file (checked 1 source file)

Expected behavior
mypy should pass

Additional context