[Bug]: Saving a fig with a colorbar using a TwoSlopeNorm sometimes results in 'posx and posy should be finite values' #22300
Description
Bug summary
When saving a figure with a colorbar and a mappable using a TwoSlopeNorm, it sometimes results in posx and posy should be finite values. The saved figure then lacks of a label for the colorbar.
Note that:
- you need to specify a
vmaxsmaller than or equat to0.9989999999999999 - this looks independent of the max and min values of
Z. Tested withZ = 10000*(X**3 - Y**4)andZ = (X**3 - Y**4)/10000
it dos not happen:
- if the figure is not saved but only shown
- with previous versions of matplotlib
Code for reproduction
import matplotlib.pyplot as plt
import numpy as np
import matplotlib.colors as colors
x = np.linspace(0, 1, 100)
y = x.copy()
X, Y = np.meshgrid(x, y)
Z = X**3 - Y**4
divnorm = colors.TwoSlopeNorm(vcenter=0, vmax=0.1)
fig = plt.figure()
c = plt.pcolormesh(x, y, Z, cmap='seismic', norm=divnorm)
plt.colorbar(c, label='Velocity [cm/s]')
plt.savefig('Spatio_temporal_velocity.png')
plt.show()Actual outcome
Expected outcome
It should display the colorbar label, and not return in the terminal posx and posy should be finite values.
Additional information
- you need to specify a
vmaxsmaller than or equat to0.9989999999999999 - this looks independent of the max and min values of
Z. Tested withZ = 10000*(X**3 - Y**4)andZ = (X**3 - Y**4)/10000
it dos not happen:
- if the figure is not saved but only shown
- with previous versions of matplotlib
Operating system
Ubuntu 20
Matplotlib Version
3.5.0
Matplotlib Backend
QtAgg
Python version
3.8.10
Jupyter version
No response
Installation
pip