Plotting seaborn.FacetGrid compatibility

[remrama]

This addresses a feature request in #306 to offer more flexibility in pingouin's plotting module. The changes were targeted at making some of the functions useable with the g.map_dataframe() method attached to seaborn grids, but they offer more flexibility when used alone as well.

The main things:

• Removed all fig and dpi keyword arguments (users should be advised to generate their own figures first, eg with plt.subplot and then pass the ax to pingouin)
• Removed the plot_blandaltmann keyword argument scatter_kws and now any kwarg gets passed through. This was done because gmap_dataframe requires the passing of a color keyword (and label), and it can be useful at times.
• Note that now, without passing fig through, the default axis that pops up is not always square. So for the few plotting functions that probably should be square (qqplot, circmean), I added an ax.set_aspect("equal") in the docstring examples.

If you want to see some arbitrary examples of the use-case I'm talking about:

import pingouin as pg
import seaborn as sns
import matplotlib.pyplot as plt

iris = sns.load_dataset("iris")

g = sns.FacetGrid(iris, col="species", aspect=1)
g.map(pg.qqplot, "petal_width")
plt.show()

g = sns.FacetGrid(iris, col="species", hue="species", sharey=False)
g.map(pg.plot_blandaltman, "petal_width", "sepal_width")
plt.show()

[raphaelvallat]

Hi @remrama,

Thanks for the PR, great idea! A few comments:

• Can you please update the unit tests?
• Update the changelog
• Does this require to bump seaborn's version in the requirements/setup?
• I do feel thatax.set_aspect("equal") should be the default for qqplot and circmean. Is there a way that we could ensure that, e.g. by including a square=True parameter like in seaborn.heatmap?

[raphaelvallat]

@remrama I've rebased the PR to master, make sure you git pull before making new changes!

[codecov]

Codecov Report

Base: 98.67% // Head: 98.58% // Decreases project coverage by -0.08% ⚠️

Coverage data is based on head (56f7924) compared to base (6816d02).
Patch coverage: 83.33% of modified lines in pull request are covered.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #314      +/-   ##
==========================================
- Coverage   98.67%   98.58%   -0.09%     
==========================================
  Files          18       18              
  Lines        3312     3320       +8     
  Branches      537      540       +3     
==========================================
+ Hits         3268     3273       +5     
  Misses         25       25              
- Partials       19       22       +3     

Impacted Files	Coverage Δ
pingouin/plotting.py	95.72% <83.33%> (-0.78%)	⬇️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[remrama]

@raphaelvallat I think this is good to go?

It passes unittests, changelog is updated, and square keyword arguments were added to qqplot() and plot_circmean(). Nothing to worry about for seaborn version requirements. map() was present in version 0.10, and you already require 0.11.

[raphaelvallat]

@remrama just made some very minor edits to documentation. Merging when the CIs are complete