az.plot_dist(kind=hist) and bins arg raises exception

Following code raises an exception

import arviz as az
newcomb = pd.read_csv("https://raw.githubusercontent.com/avehtari/ROS-Examples/master/Newcomb/data/newcomb.txt")
az.plot_dist(newcomb.values, hist_kwargs=dict(bins=30))

Matplotlib version does not

import matplotlib.pyplot as plt
newcomb = pd.read_csv("https://raw.githubusercontent.com/avehtari/ROS-Examples/master/Newcomb/data/newcomb.txt")
plt.hist(newcomb.values,bins=30)

[ahartikainen]

Yes, there is an error. We probably should grab the bins returned from the plot call.

https://github.com/arviz-devs/arviz/blob/d046cea6d2847c63f9207826d8e478e429ae2374/arviz/plots/backends/matplotlib/distplot.py#L126

to

        n, bins, _ = ax.hist(np.asarray(values).flatten(), bins=bins, **hist_kwargs)

[OriolAbril]

The idea would be something like this test where instead of using continuous_model["y"] as data a list was passed as data. Similar test with matplotlib here too in case it helps.

We want to make sure the preprocessing works and the figure is generated, what is not checked is the generated figure so things like color, axis limits… are generally not chekced. In this particular case we actually want to check the preprocessing basically so everything should work fine.