Order of workflow inputs is sometimes incorrect

[jvdzwaan]

If you run:

import nlppln
from nlppln import WorkflowGenerator

wf = WorkflowGenerator()

dir_in, skip = wf.add_inputs(dir_in='Directory', skip='string?')

dir_in has the value of skip and skip the value of dir_in

[jvdzwaan]

This is addressed in the documentation of WorkflowGenerator:

addd_inputs method returns a list of strings containing the names that
can be used to connect these input parameters to step input parameter
names. (Please note that because **kwargs are unordered, the list of input
names may not be in the same order as the **kwargs. When a workflow has
multiple inputs, it is probably safer to call add_inputs for every
parameter separately.)

Perhaps it should also be mentioned in the README.

[LourensVeen]

Maybe it should return a dictionary? Then you could write
inputs = wf.add_inputs(dir_in='Directory', skip='string?')
and use inputs['dir_in'] to refer to that input?

[bouweandela]

Solution: Janneke decided that the method should be changed to add_input.