rename selem to footprint everywhere

[grlee77]

Description

In the discussion in #4154 it was decided that footprint should be less ambiguous than selem for use in all rank filters and morphology functions.

This PR makes that change

Checklist

• Docstrings for all functions
• Gallery example in ./doc/examples (new features only)
• Benchmark in ./benchmarks, if your changes aren't covered by an
  existing benchmark
• Unit tests
• Clean style in the spirit of PEP8

For reviewers

• Check that the PR title is short, concise, and will make sense 1 year
  later.
• Check that new functions are imported in corresponding __init__.py.
• Check that new features, API changes, and deprecations are mentioned in
  doc/release/release_dev.rst.

[pep8speaks]

Hello @grlee77! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2021-06-28 13:56:00 UTC

[jni]

Awesome. 🙏 I've left one minor comment about the future of the API, but happy to merge this as-is and leave that discussion for later.

[jni]

Are you consciously leaving structure -> footprint to a later PR?

[grlee77]

No, but all instances of structure as a kwarg seem to be when calling scipy.ndimage functions. I doubt there will be much enthusiasm for renaming those upstream (at least there was resistance to renaming the first argument from input to image in the past).

[jni]

Oh my apologies, I thought this was measure.label. 🤦

[jni]

Lovely.

I wonder though whether we should not instead encourage directly using from skimage.morphology import footprints; footprint = footprints.square(5). To me this addresses the point in #5439 about grouping functions according to "purpose and equivalence classes". morphology.square() is also less descriptive than footprints.square.

Perhaps something to add to a dedicated issue for discussion (tracked from #5439) rather than hold up this PR.

[jni]

(But, a quick fix here would be to remain agnostic in the prescription: instead of "should be imported from skimage.morphology instead", we could write "skimage.morphology.selem has been moved to skimage.morphology.footprints.") (And yes I have a mild preference for "footprints" over "footprint", because the latter is easily confused with a variable/kwarg name.)

[grlee77]

I was not sure if this selem/footprint module within filters, morphology, etc. was really intended to be public or not. Specifically, these functions are grouped in with all other morphology functions in the docs.

I am fine with renaming to footprints.py and importing from there, but wonder if we can perhaps follow up later to make these logical groupings (binary vs. gray vs. footprints) clearer in the docs.

[jni]

Yes, following things up later is totally fine, but that's why I'm saying, let's not be prescriptive in the deprecation message, because we might change our mind during those discussions.

[jni]

Gotta love this function 😂

[jni]

Note: if there are no further updates I'll merge this by the end of the week.

[grlee77]

Thanks for reviewing @jni. I made the suggested rename and warning changes. As far as I could tell, the uses of structure arise of scipy.ndimage functions, not ours.

Two other things were also updated since you reviewed:
1.) renamed test_selem.py -> test_footprints.py
2.) There were still tons of mentions of "structuring element" in the docstrings! These now refer to footprint

[jni]

Great. Totally happy with the changes now. I'll merge this in <24h if no one has stepped in to review+merge before then.

[rfezzani]

Thank you @grlee77, merging 😉