tight_layout does not work with set_visible(False)

Bug summary

tight_layout() cooperates with axis('off') but not with set_visible(False).
cf. pandas#9351

Code for reproduction

This code fails:

import matplotlib.pyplot as plt
f, axarr = plt.subplots(2, 2)
axarr[1][1].set_visible(False)
plt.tight_layout()

This code succeeds:

import matplotlib.pyplot as plt
f, axarr = plt.subplots(2, 2)
axarr[1][1].axis('off')
plt.tight_layout()

Actual outcome

AttributeError: 'NoneType' object has no attribute 'is_bbox'

Expected outcome

Looks like the tight_layout code is iterating over all of the subplots rather than just the visible ones

My suspicion is that adding a condition to the list comprehension:

tight_bbox_raw = union([ax.get_tightbbox(renderer) for ax in subplots])

at line 126 of matplotlib.tight_layout.py should do the trick.

Those two commands do very different things, axis('off') makes the spines invisible, but everything else is still visible ex

fig, ax = plt.subplots()
ax.plot(range(5))
ax.axis('off')

gives

where are ax.set_visible(False) makes the Axes object (and all of it's children) are not included in the draw tree.

Which is why they look the same output-wise when the plot is empty in the first place (except for dimensional differences), gotcha.

Navigation Menu