Using color maps in plot_khat

[agustinaarroyuelo]

Description

plot_khat can take a c argument to use with a color map

Checklist

• Follows official PR format
• Includes a sample plot to visually illustrate the changes (only for plot-related functions)
• Code style correct (follows pylint and black guidelines)
• Changes are listed in changelog

[OriolAbril]

Looking now at the code and the comments in slack, does az.plot_khat(loo, color=values) work? I think it should work already with the current rgba_c generation and use a colormap. We may have to handle the presence of the c kwarg better though.

[agustinaarroyuelo]

You are right az.plot_khat(loo, color=values) works. Is this PR enough for handling the c kwarg?

[OriolAbril]

I would have the color argument of plot_khat take preference over the c passed as kwarg, I think we can do that by adding these lines after the dealiaser, around line 78:

c_kwarg = kwargs.pop("c", None)
if c_kwarg is not None:
    warnings.warn("Found kwarg named `c` which will be ignored, use `color` instead")

But I would also understand wanting to pass c to avoid all the preprocessing done to color, ,we could also use the check to instead of printing a warning, use c directly and avoid all the color preprocessing crazyness.

[OriolAbril]

Oh, the preprocessing is also used to create a legend based on the colors shown if desired in case the variable that gets color encoded is discrete instead of continuous.

[OriolAbril]

This has to come after the except. At this point color can be a list of strings or a nx3 array of rgb colors. I would also suggest using matplotlib normalizer function in case someone wanted to use a different one

cmap_name = kwargs.get("cmap", plt.rcParams["image.cmap"])
cmap = getattr(cm, cmap_name)
norm_fun = kwargs.get("norm", mpl.colors.Normalize(color.min(), color.max()))
rgba_c = cmap(norm_fun(color))

[OriolAbril]

I would not add that. I think c should always behave exactly like the argument in matplotlb.scatter if they want any of the other features, users should use color

[codecov]

Codecov Report

Merging #1592 (ed4309e) into main (c562fad) will decrease coverage by 0.03%.
The diff coverage is 86.11%.

@@            Coverage Diff             @@
##             main    #1592      +/-   ##
==========================================
- Coverage   90.05%   90.02%   -0.04%     
==========================================
  Files         108      108              
  Lines       11584    11594      +10     
==========================================
+ Hits        10432    10437       +5     
- Misses       1152     1157       +5     

Impacted Files	Coverage Δ
arviz/plots/backends/matplotlib/khatplot.py	73.07% <83.87%> (-1.93%)	⬇️
arviz/data/base.py	97.87% <100.00%> (ø)
arviz/data/io_pymc3.py	90.68% <100.00%> (ø)
arviz/plots/backends/bokeh/energyplot.py	93.44% <100.00%> (ø)
arviz/plots/backends/matplotlib/energyplot.py	90.00% <100.00%> (ø)
arviz/plots/backends/matplotlib/ppcplot.py	98.07% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c562fad...be47816. Read the comment docs.