`func` as function input variable name

[htwangtw]

func as function input variable name creates conflict with mark.task(func)

Pydra version: 0.14
Minimal example to recreate the bug:

import nibabel as nb
import numpy as np
import pydra


@pydra.mark.task
def get_func_affine(func: str)-> np.array:
    aff_orig = nb.load(func).affine[:, -1]
    return aff_orig

wf = pydra.engine.Workflow(name="wf1",
                           input_spec=["func"])
wf.add(get_func_affine(name="get_func_affine",
                 func=wf.lzin.func))
wf.set_output([("out", wf.get_func_affine.lzout.out)])

Error:

TypeError                                 Traceback (most recent call last)
<ipython-input-1-6fa711100f80> in <module>
     10                            input_spec=["func"])
     11 wf.add(get_func_affine(name="get_func_affine",
---> 12                  func=wf.lzin.func))
     13 wf.set_output([("out", wf.get_func_affine.lzout.out)])

~/git_repos/pydra_playground/pydra/pydra/mark/functions.py in decorate(**kwargs)
     42 
     43     def decorate(**kwargs):
---> 44         return FunctionTask(func=func, **kwargs)
     45 
     46     return decorate

TypeError: type object got multiple values for keyword argument 'func'

[effigies]

The minimal reproduction is:

@pydra.mark.task
def f(func):
    print(func)

task = f(func="some_val")

This feels like it might be related to #295 as well.

[tclose]

this type of problem will persist, but for args called function instead. Maybe they should be automatically translated to function_ in the task class builder

[djarecka]

yes, that why we used _func at the end.

[tclose]

function_ is the one that makes the most sense