Pypalette / arbitrary cmaps

[TyberiusPrime]

I recently ran across https://y-sunflower.github.io/pypalettes/
and they have some nice color maps, for which they
provide list of colors (load_palette()) and cmaps (load_cmap).

One can use the palettes using:
scale_fill_manual(pypalettes.load_pallete("Lime") for the discrete case
and scale_color_gradientn(pypalettes.load_pallete("Lime")
for the continious one.

What one can't use is 'the obvious choice for cmaps:
scale_*_cmap (or scale_*_cmap_d), since these only work with the named color palettes in matplotlib.cm.

Now I'm not certain it's worth the effort to support arbitrary cmaps (or even feasible, plotnine might need the name, haven't dug into mizanis handling here),
but I want to leave the workaround in this discussion, for the next searcher.

(Sidenote: The plotnine docs suggests to query 'matplotlib.cm.cmap_d.keys()' but at least on my matplotlib cmap_d doesn't exist).

So long,
TP

[has2k1]

Mizani needs some big work on colors and there is a lot to improve.

1. Palettes need to catchup with the "latest" colour scale and interpolation science. i.e. oklch ...
2. Harmonise mizani's handling of colors with the rest of the echo system. This is where pypalettes would come in.
3. Userspace color manipulation methods. Taking a leaf from the colour functions in SCSS/SASS. These can be really handy.
4. May be faster colour math

[TyberiusPrime]

One giant task at a time! So psyched about the new composing of plots!
Can we throw in 'multiple color/fill scales'? It's the one aspect where I'd be ok with multiple scales
in one plot.