'AxesSubplot' object has no attribute 'sharey' on Colab

[jecampagne]

Does someone else experience trace on simple arviz exemple see nb on Colab
Here is the snippet

import arviz as az  #0.12.1
import numpy as np

# This Plot is ok
data = {'a':np.random.normal(size=(100,)), 
        "b":np.random.normal(size=(100,)), 
         }
az.plot_pair(
    data,
    kind="kde",
    kde_kwargs={
        "hdi_probs": [0.393, 0.865, 0.989],  # Plot 1sigma, 2sigma, 3sigma HDI contours
        "contourf_kwargs": {"cmap": "Reds"},
    },
    marginal_kwargs={"color": "Black"},
    marginals=True,
    textsize=30,
);

#But here it crashes
data = {'a':np.random.normal(size=(100,)), 
        "b":np.random.normal(size=(100,)), 
        "c":np.random.normal(size=(100,)), 
         }
az.plot_pair(
    data,
    kind="kde",
    kde_kwargs={
        "hdi_probs": [0.393, 0.865, 0.989],  # Plot 1sigma, 2sigma, 3sigma HDI contours
        "contourf_kwargs": {"cmap": "Reds"},
    },
    marginal_kwargs={"color": "Black"},
    marginals=True,
    textsize=30,
);

The crash is

/usr/local/lib/python3.8/dist-packages/arviz/plots/backends/matplotlib/pairplot.py in plot_pair(ax, plotters, numvars, figsize, textsize, kind, scatter_kwargs, kde_kwargs, hexbin_kwargs, gridsize, colorbar, divergences, diverging_mask, divergences_kwargs, flat_var_names, backend_kwargs, marginal_kwargs, show, marginals, point_estimate, point_estimate_kwargs, point_estimate_marker_kwargs, reference_values, reference_values_kwargs)
    267                 for j in range(0, vars_to_plot):
    268                     for i in range(0, j):
--> 269                         ax[j, i].axes.sharey(ax[j, 0])
    270 
    271         hexbin_values = []

AttributeError: 'AxesSubplot' object has no attribute 'sharey'

If someone has a workaround that would be nice. (I have googled and not found a solution yet). Thanks

[OriolAbril]

I can't seem to reproduce the issue with the notebook you shared. The default versions might have been updated.

Can you try again with default libraries? And making sure both arviz and matplotlib are on their latest versions?

[jecampagne]

Yes you're right it woks now on Colab. I have not changed anything...
But Colab as also change jax lib few days ago. SO it may have change / cure some incompatible versons...
Thanks