fix opacity adjustment and set it to low non-zero value

[zoccoler]

This should fix #279 and could potentially solve #271 .

This lets the selected layer barely visible (opacity = 0.2), which would not lead (I think) to the user thinking the original labels layer was broken, but still hide it quite a bit so that the clustering result does not blend much with the labels colors.

The best option in my opinion is to interactively hide and show it, which is already written in the documentation under Manual clustering.

Diving deep into git blame, it seems this was set by @thorstenwagner here, which then was later updated to setting opacity to 0. @thorstenwagner, would it be OK for your use cases to let a small opacity like 0.2?

Best,
Marcelo

[zoccoler]

The changes are minimal, I added both @stefanhahmann and @thorstenwagner as reviewers because you are somehow involved in the issues.

Let me know if you agree with the changes.

[codecov]

Codecov Report

Attention: 1 lines in your changes are missing coverage. Please review.

Comparison is base (2b47b7f) 76.52% compared to head (d74d308) 76.52%.

Files	Patch %	Lines
napari_clusters_plotter/_plotter.py	0.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #280   +/-   ##
=======================================
  Coverage   76.52%   76.52%           
=======================================
  Files          16       16           
  Lines        1853     1853           
=======================================
  Hits         1418     1418           
  Misses        435      435           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Cryaaa]

I would generally agree to the changes but it would be good to hear from the others! Apart from that feel free to merge

[stefanhahmann]

I think this PR is a good idea. I would even increase the opacity to 0.25 or 0.3, but this is subjective.
In my view, this would solve #271 indeed. In fact, this PR implements one of the described solutions of #271.

[stefanhahmann]

I tested a bit. I would prefer 0.25 or even 0.3

[Cryaaa]

I would like to not increase it too much so maybe we can stick to 0.25, so that the experience does not change too dramatically?

[stefanhahmann]

0.25 would be fine for. (I could also live with 0.2). The slight increase was just a subjective suggestion.

[Cryaaa]

@zoccoler,
I have one question before merging: does the visbility actually change the speed of the plugin? I could imagine that by making the layer visible napari would need some more time to render the image or is this not the case? Unfortunately I do not have much time for testing this week but will reply once I have had the chance to.

[stefanhahmann]

@Cryaaa I could imagine that by making the layer visible napari would need some more time to render the image or is this not the case?

I just tested this a bit on my machine and it did not seem to make a difference, even though I would agree, that one might expect a difference.

[Cryaaa]

@Cryaaa I could imagine that by making the layer visible napari would need some more time to render the image or is this not the case?

I just tested this a bit on my machine and it did not seem to make a difference, even though I would agree, that one might expect a difference.

Thanks for checking this so soon! I am fine with merging this from my end maybe @thorstenwagner can chime in over the next few days if he is also fine with this? otherwise I'll merge it.

[thorstenwagner]

I totally forgot to check that. Will do it today.

[thorstenwagner]

Hi, from my side everything still works as it should be. Probably because I set the opacity to zero when I load the the plugin:
https://github.com/MPI-Dortmund/napari-tomotwin/blob/main/src/napari_tomotwin/load_umap.py#L79

[Cryaaa]

Perfect! Then I'll merge this fixing #279