Limitations of the kwargs_to_strings decorator not handling positional arguments

[seisman]

Here are some limitations of the kwargs_to_strings decorator:

1. Doesn't work for the default values of parameters:

#2208 (comment)

Please note that we can't use decorator @kwargs_to_strings(offset="sequence") here, because the decorator can only deal with keyword arguments and doesn't work for the default values. I.e., it doesn't work for Figure.timestamp() but works for Figure.timestamp(offset=["-54p", "-54p"]), although offset defaults to ["-54p", "-54p"].

@kwargs_to_strings(offset="sequence")
def timestamp(
    self,
    text=None,
    label=None,
    justification="BL",
    offset=("-54p", "-54p"),
    font="Helvetica",
    timefmt="%Y %b %d %H:%M:%S",
):

2. Doesn't work if the parameter name is not given

>>> from pygmt import which
>>> # doesn't work
>>> which(["@hotspots.txt", "@earth_relief_01d_g"])
gmtwhich [ERROR]: File ['@hotspots.txt', not found!
gmtwhich [ERROR]: File '@earth_relief_01d_g'] not found!
>>> # works
>>> which(fname=["@hotspots.txt", "@earth_relief_01d_g"])

[weiji14]

This issue would also apply to fig.text's textfiles parameter that uses sequence_space added in #325.

pygmt/pygmt/src/text.py

Lines 41 to 50 in c9d6147

	textfiles="sequence_space",
	angle="sequence_comma",
	font="sequence_comma",
	justify="sequence_comma",
	c="sequence_comma",
	p="sequence",
	)
	def text_(
	self,
	textfiles=None,

Should we remove sequence_space entirely since it is only used in pygmt.which and fig.text? And use the if is_nonstr_iter(fname): fname = " ".join(fname) patch in #2726?

[seisman]

And use the if is_nonstr_iter(fname): fname = " ".join(fname) patch in #2726?

No, it won't work for Figure.text, because textfiles can be a 2-D array, e.g., textfiles=[[1, 2, "A"], [3, 4, "B"]].

[weiji14]

And use the if is_nonstr_iter(fname): fname = " ".join(fname) patch in #2726?

No, it won't work for Figure.text, because textfiles can be a 2-D array, e.g., textfiles=[[1, 2, "A"], [3, 4, "B"]].

Ah ok, you're right. Let's keep sequence_space then. So #2726 will fix this issue entirely? Can we change Addresses #2726 to Fixes #2726 or was there something else?

[seisman]

So #2726 will fix this issue entirely?

No, see the first point:

Doesn't work for the default values of parameters: