[BUG] map_task sensitive to argument order

[j-hartshorn]

Describe the bug

Using map_task with a partially applied function will raise an exception depending on the order of srguments in the original function definition.

Expected behavior

Argument order ought not to matter with a partially applied function being mapped. This would align with Python syntax.

Additional context to reproduce

Workflow python file workflows/example.py:

from functools import partial

from flytekit import map_task, task, workflow


@task()
def create_names() -> list[str]:
    return ["a", "bb", "ccc"]


@task()
def get_number() -> int:
    return 3


@task()
def multiply_name(
    number: int,
    name: str,
) -> str:
    return " ".join([name] * number)


@workflow
def process_all() -> list[str]:
    name_list = create_names()
    number = get_number()
    multiply_name_partial = partial(
        multiply_name,
        number=number,
    )
    result = map_task(multiply_name_partial)(name=name_list)
    return result

Python 3.11.5

requirements.in:

black==23.9.1
dataclasses-json==0.6.1
flytekit==1.8.0
marshmallow-enum==1.5.1
ipykernel==6.25.2
isort==5.12.0

To reproduce: pyflyte run workflows/example.py process_all

It will work of you change the argument order in the multiply_nametask to this:

@task()
def multiply_name(
    name: str,
    number: int,
) -> str:
    return " ".join([name] * number)

Screenshots

No response

Are you sure this issue hasn't been raised already?

• Yes

Have you read the Code of Conduct?

• Yes

[welcome]

Thank you for opening your first issue here! 🛠

[soumya-dash3]

Hello @pingsutw, can you please assign me this issue?

[soumya-dash3]

Sorry for the inconvenience @pingsutw , you may assign this to someone else.

[chaohengstudent]

Hi @pingsutw, can you assign this issue to me, thanks.