Add layout properties for Bokeh backend.

[ahartikainen]

Description

Use bokeh.layouts.layout by default, when show is True. This enables user to define rcParams for sizing_mode and also setting the toolbar_location.

Checklist

• Follows official PR format
• Includes a sample plot to visually illustrate the changes (only for plot-related functions)
• New features are properly documented (with an example if appropriate)?
• Includes new or updated tests to cover the new feature
• Code style correct (follows pylint and black guidelines)
• Changes are listed in changelog

[ahartikainen]

We probably could add also rcParam for the plot output?

E.g. bokeh.layout.shape = default would use the default order given our plot. But user could set it for .. square which would unravel axes (remove None) and then reshape so the plots fill a square.

Maybe other options could be possible based on maximum columns and maximum rows?

layout also eats rugged list of lists, so we could have 3 / 2 / 3 structure etc?

Or maybe this could be a function that user can use manually (show=False)?

[ahartikainen]

cc @sethaxen @AlexAndorra

[ahartikainen]

import arviz as az
import bokeh.plotting as bkp
from bokeh.layouts import layout
import numpy as np

def make_square(arr):
    """Flat list to square"""
    n = np.ceil(len(arr)**0.5)
    new_arr = arr + [None for _ in range(n**2 - len(arr))]
    return np.array(new_arr).reshape(n,n).tolist()

az.backends.output_file("pair_plot_centered_eight.html")

data = az.load_arviz_data("centered_eight")
axes = az.plot_pair(data, backend="bokeh", divergences=True, show=False)
ravelled_axes = [ax for ax in axes.ravel() if ax is not None]
padded_square = make_square(ravelled_axes)
bkp.show(layout(padded_square, sizing_mode="stretch_both"))

[ahartikainen]

I think this has the main functionality. We need to add examples at some point to docs.

@OriolAbril @canyon289 and others, any better names / functionality for the row / column hack?

Should we use ncolumn / nrow structure --> 3column 2column 5row 6row etc for rcParam? How should we handle those cases?

row or column in param --> regex --> if not found --> normal test?

fyi

3column --> three columns --> shape == (-1, 3)

[codecov]

Codecov Report

Merging #1089 into master will decrease coverage by 14.22%.
The diff coverage is n/a.

@@             Coverage Diff             @@
##           master    #1089       +/-   ##
===========================================
- Coverage   93.45%   79.22%   -14.23%     
===========================================
  Files          93       93               
  Lines        8705     8775       +70     
===========================================
- Hits         8135     6952     -1183     
- Misses        570     1823     +1253     

Impacted Files	Coverage Δ
arviz/data/io_emcee.py	12.50% <0.00%> (-85.84%)	⬇️
arviz/data/io_pystan.py	16.47% <0.00%> (-82.60%)	⬇️
arviz/data/io_tfp.py	14.42% <0.00%> (-81.74%)	⬇️
arviz/data/io_cmdstanpy.py	21.69% <0.00%> (-78.31%)	⬇️
arviz/data/io_pymc3.py	22.94% <0.00%> (-71.68%)	⬇️
arviz/data/io_pyro.py	25.92% <0.00%> (-71.30%)	⬇️
arviz/data/io_numpyro.py	25.23% <0.00%> (-66.36%)	⬇️
arviz/data/io_cmdstan.py	43.77% <0.00%> (-51.08%)	⬇️
arviz/data/base.py	89.02% <0.00%> (-10.98%)	⬇️
arviz/data/converters.py	65.15% <0.00%> (-1.52%)	⬇️
... and 1 more

---

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 dc20591...971049a. Read the comment docs.

[codecov]

Codecov Report

Merging #1089 into master will decrease coverage by 14.22%.
The diff coverage is n/a.

@@             Coverage Diff             @@
##           master    #1089       +/-   ##
===========================================
- Coverage   93.45%   79.22%   -14.23%     
===========================================
  Files          93       93               
  Lines        8705     8775       +70     
===========================================
- Hits         8135     6952     -1183     
- Misses        570     1823     +1253     

Impacted Files	Coverage Δ
arviz/data/io_emcee.py	12.50% <0.00%> (-85.84%)	⬇️
arviz/data/io_pystan.py	16.47% <0.00%> (-82.60%)	⬇️
arviz/data/io_tfp.py	14.42% <0.00%> (-81.74%)	⬇️
arviz/data/io_cmdstanpy.py	21.69% <0.00%> (-78.31%)	⬇️
arviz/data/io_pymc3.py	22.94% <0.00%> (-71.68%)	⬇️
arviz/data/io_pyro.py	25.92% <0.00%> (-71.30%)	⬇️
arviz/data/io_numpyro.py	25.23% <0.00%> (-66.36%)	⬇️
arviz/data/io_cmdstan.py	43.77% <0.00%> (-51.08%)	⬇️
arviz/data/base.py	89.02% <0.00%> (-10.98%)	⬇️
arviz/data/converters.py	65.15% <0.00%> (-1.52%)	⬇️
... and 1 more

---

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 dc20591...971049a. Read the comment docs.

[codecov]

Codecov Report

Merging #1089 into master will decrease coverage by 12.15%.
The diff coverage is n/a.

@@             Coverage Diff             @@
##           master    #1089       +/-   ##
===========================================
- Coverage   91.58%   79.43%   -12.16%     
===========================================
  Files          93       93               
  Lines        8713     8850      +137     
===========================================
- Hits         7980     7030      -950     
- Misses        733     1820     +1087     

Impacted Files	Coverage Δ
arviz/data/io_emcee.py	12.50% <0.00%> (-83.34%)	⬇️
arviz/data/io_tfp.py	14.42% <0.00%> (-81.74%)	⬇️
arviz/data/io_cmdstanpy.py	21.69% <0.00%> (-78.31%)	⬇️
arviz/data/io_pyro.py	25.92% <0.00%> (-71.30%)	⬇️
arviz/data/io_pymc3.py	22.94% <0.00%> (-69.70%)	⬇️
arviz/data/io_numpyro.py	25.23% <0.00%> (-66.36%)	⬇️
arviz/data/io_cmdstan.py	43.77% <0.00%> (-51.08%)	⬇️
arviz/data/io_pystan.py	16.47% <0.00%> (-46.64%)	⬇️
arviz/data/base.py	89.02% <0.00%> (-10.98%)	⬇️
arviz/data/converters.py	65.15% <0.00%> (-1.52%)	⬇️
... and 1 more

---

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 0a0f18d...0382f65. Read the comment docs.

[OriolAbril]

LGTM