Fix ImportError in plantcv caused by change in skimage.feature function name

[fbarbe00]

Currently, when installing plantcv, the following error is raised: ImportError: cannot import name 'greycomatrix' from 'skimage.feature'
This is because skimage.feature changed their function name from greycomatrix to graycomatrix (see https://stackoverflow.com/questions/67537650/cannot-import-name-graycomatrix-from-skimage-feature ) from version 0.19. This commit changes the requirements.txt file and the function name.

[HaleySchuhl]

@fbarbe00 thanks very much for opening this PR and sorry for the delay in response.

[HaleySchuhl]

Ugh..Looks like the renaming issues also trickle down past the imports part.

Now tests fail on

• plantcv/plantcv/closing.py:26: in closing --> E TypeError: binary_closing() got an unexpected keyword argument 'selem'
• The watershed_segmentation function --> E TypeError: peak_local_max() got an unexpected keyword argument 'indices'

If you'd prefer, I can dig into their updated parameter names, push commits to this PR, and hopefully resolve those remaining bugs. Thanks again @fbarbe00 for bringing this to our attention. @nfahlgren can we add them as a contributor so tests automatically run next time?

Update:
I pushed two commits to address the updates to parameter names from the two functions I mentioned. However, the watershedd_segmentation function still isn't un-broken since the output of skimage.feature.peak_local_max is now returns The coordinates of the peaks and doesn't seem to give us the option any longer to change indices (If False, the output will be a boolean array shaped as image.shape with peaks present at True elements).

https://scikit-image.org/docs/0.20.x/api/skimage.feature.html#peak-local-max

vs.

https://scikit-image.org/docs/0.18.x/api/skimage.feature.html#peak-local-max

[fbarbe00]

Thanks a lot for your reply! Is there anything I can still try to help with?

[fbarbe00]

Btw, this solves Issue #1002

[HaleySchuhl]

Yes @fbarbe00 If you're able to figure out how to get an array-like output from the latest version of skimage.feature.peak_local_maxskimage.feature.peak_local_max . TThe that functionality isn't recoverable I might suggest pinning the version down to the older functionality.

[fbarbe00]

Hi @HaleySchuhl , I believe my last commit fixes it! I just looked at what the peak_local_max function did in v0.18 (https://github.com/scikit-image/scikit-image/blob/v0.18.x/skimage/feature/peak.py) , and looks like indices=False only affected two lines at the end. These lines have now been added to plantcv.

Note that I do not have a deep understanding of what this code actually does, but this does seem to fix this function and work as before.

[codecov]

Codecov Report

Merging #998 (2b16a1d) into main (df07ef6) will decrease coverage by 0.02%.
The diff coverage is 100.00%.

❗ Current head 2b16a1d differs from pull request most recent head 61f4fca. Consider uploading reports for the commit 61f4fca to get more accurate results

@@             Coverage Diff             @@
##              main     #998      +/-   ##
===========================================
- Coverage   100.00%   99.98%   -0.02%     
===========================================
  Files          129      141      +12     
  Lines         5803     6004     +201     
===========================================
+ Hits          5803     6003     +200     
- Misses           0        1       +1     

Flag	Coverage Δ
unittests	99.98% <100.00%> (-0.02%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
plantcv/plantcv/closing.py	100.00% <100.00%> (ø)
plantcv/plantcv/threshold/threshold_methods.py	100.00% <100.00%> (ø)
plantcv/plantcv/watershed.py	100.00% <100.00%> (ø)

... and 14 files with indirect coverage changes

[HaleySchuhl]

Awesome! Tests seem to pass but now there's an issue in the Python 3.7 build on codecov for some reason? ErrorDetail(string='Unable to locate build via Github Actions API. Please upload with the Codecov repository upload token to resolve issue.', code='not_found')

[nfahlgren]

I think codecov stopped working with Python 3.7, which was the oldest version we were testing. I bumped up the testing environments to 3.8, 3.9, and 3.10 since 3.11 is out now.