Add button to automatically apply RGB presets

[rosteen]

Pretty self explanatory, I hope. Ready for everyone to try to break it. Demo with 6 layers (note that I changed the button text and decided to set stretch to 99% for all layers since I made this recording):

Screen.Recording.2023-11-09.at.1.31.03.PM.mov

[codecov]

Codecov Report

Attention: 4 lines in your changes are missing coverage. Please review.

Comparison is base (dc09c8e) 90.83% compared to head (bb9e0a8) 90.79%.
Report is 1 commits behind head on main.

❗ Current head bb9e0a8 differs from pull request most recent head f8a99b8. Consider uploading reports for the commit f8a99b8 to get more accurate results

Files	Patch %	Lines
...nfigs/default/plugins/plot_options/plot_options.py	93.10%	2 Missing ⚠️
...lt/plugins/plot_options/tests/test_plot_options.py	94.11%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2558      +/-   ##
==========================================
- Coverage   90.83%   90.79%   -0.04%     
==========================================
  Files         160      160              
  Lines       19291    19354      +63     
==========================================
+ Hits        17523    17573      +50     
- Misses       1768     1781      +13     

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

[rosteen]

I'll add a test tomorrow, FWIW.

[kecnry]

Seems to work quite well and gives us a good starting place to test against a bunch of different scenarios and tweak the underlying algorithm/defaults. I'm a bit torn about whether this is too slow that we should set the underlying glue state directly or so fast that we should intentionally slow it down so the user can see the steps being applied 😂 so I guess that means we should just leave it as-is, at least for now.

This could definitely use test coverage before merge as well, though!

[pllim]

OK looks like expected now, I think. 96.96% of diff hit (target 90.00%)

[kecnry]

I'm a bit worried about two different if-else statements that need to be synced to each other (if we ever change the length of the supported number of presets, we need to update the list above, this if-else, and the if-else in the for loop. Maybe we can create the list of colors before the for-loop and then just access and set them within the for loop?

[rosteen]

Yeah, I don't love all the conditionals here either. Creating a matplotlib colormap is cheap enough I think I'll just do it anytime there are more than 2 visible layers in that first if statement. Actually, instead of hard-coding numbers, I can check against the length of preset_colors.

[kecnry]

what if we had (before the for-loop):

if n_visible > len(preset_colors):
    preset_colors = [matplotlib.colors.to_hex(cmap(i), keep_alpha=True) for i in range(n_visible)]

and then in the for-loop, just access preset_colors[i] (or better yet, loop over preset colors as well)?

[rosteen]

Can't quite just loop over the preset colors, since we also need to get the right index from the layer list in the loop. I made the other changes though.

[kecnry]

Could loop over both with zip (for visible_layer, preset_color in zip(visible_layers, preset_colors)) if we don't need to access much else, but this works too

[kecnry]

Thanks!

[javerbukh]

LGTM!

[camipacifici]

Sorry, coming here late. Can you please invert the color scheme? An intuitive workflow would be:

• I load images in order of wavelength (F115W, F150W, F200W)
• These go from low wavelength to high wavelength
• So the colors should go from blue to red.

Currently it goes from red to blue.
Thank you!

[pllim]

Since this PR is merged, do we need a new ticket?

[rosteen]

Since this PR is merged, do we need a new ticket?

Nah, that's fast enough I'll just do it without one.