[REF] Deprecate kw_only_meth/func decorators

[rmarkello]

Closes #841.

This is a minor update, but it removes a bit of custom code (the nibabel.keywordonly module) forcing methods/functions to have a pre-specified number of positional arguments and instead uses the Python 3.x * argument to achieve the same performance.

I haven't added any additional tests to check that the * argument performs identically to the decorators since test_keywordonly.py was only testing the decorators themselves (and not the functions they wrapped)—and those functions are now removed—but let me know if you'd like to see something explicit!

[codecov]

Codecov Report

Merging #848 into master will decrease coverage by 0.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master     #848      +/-   ##
==========================================
- Coverage   90.09%   90.07%   -0.02%     
==========================================
  Files          98       98              
  Lines       12452    12428      -24     
  Branches     2190     2190              
==========================================
- Hits        11219    11195      -24     
  Misses        883      883              
  Partials      350      350

Impacted Files	Coverage Δ
nibabel/freesurfer/mghformat.py	95.52% <100%> (-0.04%)	⬇️
nibabel/keywordonly.py	100% <100%> (ø)	⬆️
nibabel/dataobj_images.py	95.58% <100%> (-0.19%)	⬇️
nibabel/ecat.py	88.2% <100%> (-0.07%)	⬇️
nibabel/cifti2/cifti2.py	96.67% <100%> (-0.02%)	⬇️
nibabel/arrayproxy.py	100% <100%> (ø)	⬆️
nibabel/parrec.py	91.78% <100%> (-0.08%)	⬇️
nibabel/analyze.py	98.52% <100%> (-0.01%)	⬇️
nibabel/minc2.py	89.87% <100%> (-0.26%)	⬇️
nibabel/minc1.py	90.9% <100%> (-0.11%)	⬇️
... and 2 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update fcc5448...e69ec32. Read the comment docs.

[rmarkello]

Also, in case it needs to be said: this is very low priority and can just hang out until after the 3.0 release! I just had a few minutes and though I'd get it off my to-do list 😅

[effigies]

Since this is part of the API, I'm afraid it will need to go through a deprecation cycle.

Have a look at:

nibabel/nibabel/minc.py

Lines 1 to 10 in fcc5448

	""" Deprecated MINC1 module """
	import warnings
	warnings.warn("We will remove this module from nibabel 3.0; "
	"Please use the 'minc1' module instead",
	DeprecationWarning,
	stacklevel=2)
	from .minc1 import * # noqa

nibabel/nibabel/tests/test_removalschedule.py

Lines 4 to 9 in fcc5448

	MODULE_SCHEDULE = [
	('4.0.0', ['nibabel.trackvis']),
	('3.0.0', ['nibabel.minc', 'nibabel.checkwarns']),
	# Verify that the test will be quiet if the schedule outlives the modules
	('1.0.0', ['nibabel.neverexisted']),
	]

[rmarkello]

Gotcha! Do you have a preference for what version to list for deprecation? I'm assuming 4.0.0 based on what's currently listed.

[effigies]

I think we usually go +2 major versions, so let's say 5.0.

[effigies]

Thanks!