Different plot when running as script vs Jupyter/Console

[LucaMarconato]

It may be related to this other bug: #68

from spatialdata.datasets import blobs_annotating_element
import spatialdata_plot
import matplotlib.pyplot as plt

axes = plt.subplots(1, 2)[1]

sdata = blobs_annotating_element("blobs_circles")
sdata.pl.render_images("blobs_image").pl.render_shapes("blobs_circles", color="radius").pl.show(ax=axes[0])
sdata.pl.render_images("blobs_image").pl.render_shapes("blobs_circles", color="radius").pl.show(ax=axes[1])

plt.show()

In a Jupyter Notebook or Python console gives the following (correct) plot:

While running the code as a script gives this plot:

An observation, I guess that tests behaves like running code as scripts.