Add the 6 color and the 8 color Petroff color cycles from@mpetroff'sMatthew A. Petroff, Accessible Color Sequences for Data Visualization. The 10 color cycle was added in PR#27851 which landed in Matplotlib v3.10.0.

PR checklist

"closes #0000" is in the body of the PR description tolink the related issue
new and changed code istested
[N/A]Plotting related features are demonstrated in anexample
- Thepetroff10 color cycle already has an example inhttps://matplotlib.org/stable/gallery/style_sheets/petroff10.html, so to add ones for thepetroff6 andpetroff8 might be redundant.
New Features andAPI Changes are noted with adirective and release note
Documentation complies withgeneral anddocstring guidelines

matthewfeickert added2 commits

May 18, 2025 01:07

ENH: Add Petroff 6 and 8 color cycle style sheets

0d8449f

* Add the 6 color and the 8 color Petroff color cycles from  Matthew A. Petroff, Accessible Color Sequences for Data Visualizationhttps://arxiv.org/abs/2107.02270.  The 10 color cycle was added in PR 27851 which landed in Matplotlib v3.10.0.

TST: Add 6 and 8 color Petroff color sequences to tests

324e7f9

* Verify they appear in plt.color_sequences.

github-actionsbot added topic: color/color & colormaps topic: styles labels

May 18, 2025

DOC: Add 'What's New' entry for 6 and 8 Petorff color cycles

b89e4ce

* Note the addition of 'petroff6' and 'petroff8' and give an example  of how to load them.

matthewfeickert force-pushed theenh/add-petroff6-and-petroff8 branch from5127e01 tob89e4ceCompare

May 18, 2025 07:58

matthewfeickert marked this pull request as ready for review

May 18, 2025 07:59

Copy link

ContributorAuthor

matthewfeickert commentedMay 18, 2025•
edited
Loading

This should also become part of Issue#28750. This PR is just implementing the same components as PR#27851 did forpetroff10, and so could be viewed as the bare minimum. If this isn't desirable without solving Issue#28750 in the same PR, that's fine, but I'll probably need additional guidance and won't have time to do something of that level of detail until June 2025.

Copy link

ContributorAuthor

matthewfeickert commentedMay 18, 2025•
edited
Loading

Tagging@timhoffm and@rcomer for review given discussion in Issue#30060 and tagging@tacaswell given review of PR#27851.

@mpetroff, your review is of course welcome too.

matthewfeickert commented

May 18, 2025

View reviewed changes

lib/matplotlib/_cm.py

Copy link

ContributorAuthor

matthewfeickertMay 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others.Learn more.

The RGB values were determiend from converting the hex usinghttps://www.rapidtables.com/convert/color/hex-to-rgb.html. If a more specific method is desired, I'm happy to double check the RGB values with it.

Copy link

Member

oscargusMay 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others.Learn more.

I think this should do, since screens have much more difference between them than a single RGB color step anyway.

(Since I didn't know which was the original format, I did some conversions, thinking that the RGB was the approximation of the 0-1-range. And then read the paper to see that it was the other way around. It seems to match exactly.)

Copy link

Contributor

mpetroffMay 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others.Learn more.

This code will exactly convert thetop-cycles.json file to the format here:

#!/usr/bin/env python3importjsonwithopen("top-cycles.json")asinfile:cycles=json.load(infile)forkincycles:print(f"_petroff{k}_data = (")forcincycles[k]:cf= [str(int(c[i*2 : (i+1)*2],16)/255)foriinrange(3)]print(f"    ({cf[0]+',':20}{cf[1]+',':20}{cf[2]+'),':22}  #{c}")print(")\n")

oscargus approved these changes

May 18, 2025

View reviewed changes

Copy link

Member

rcomer commentedMay 18, 2025

Thanks for following this up@matthewfeickert! Could you also add the new sequences to the list inthis example? It might also be good to turnthis one into a comparison of the three Petroff style sheets, but that could be a follow up.

jklymak mentioned this pull request

May 18, 2025

[Doc]: new color documentation should go into user docs#30071

Open

Copy link

ContributorAuthor

matthewfeickert commentedMay 19, 2025

Could you also add the new sequences to the list inthis example? It might also be good to turnthis one into a comparison of the three Petroff style sheets, but that could be a follow up.

@rcomer For sure. I am teaching at a workshop for the next two days so I might be delayed in doing this until very late tonight, but I'll ping you for review once I've done so.

I'll also update the values (if different) using@mpetroff's code snippet from#30065 (comment).

Copy link

Contributor

mpetroff commentedMay 19, 2025

I'll also update the values (if different) using@mpetroff's code snippet from#30065 (comment).

They're identical, so there's no update needed.

tacaswell approved these changes

May 19, 2025

View reviewed changes

tacaswell added this to thev3.11.0 milestone

May 19, 2025

Copy link

Member

tacaswell commentedMay 19, 2025

The test drop is due to at least one missing CI upload.

tacaswell merged commit1c881a3 intomatplotlib:main

May 19, 2025

38 of 39 checks passed

matthewfeickert deleted the enh/add-petroff6-and-petroff8 branch

May 20, 2025 04:47

Copy link

ContributorAuthor

matthewfeickert commentedMay 20, 2025

Thanks for the reviews everyone and for the fast merge@tacaswell.

Could you also add the new sequences to the list inthis example? It might also be good to turnthis one into a comparison of the three Petroff style sheets, but that could be a follow up.

@rcomer I'll follow up in another PR on this so that it doesn't get dropped.