[Matplotlib-users] set_clip_mask() single polygon for multiple subplots

[Scott H. <st...@co...>]

I'd like to use the same patch to clip two images that share the same 
axes, and extract values from the un-clipped region of both arrays. 
Unfortunately this seems harder than expected. Code & questions below, 
Thanks!

from matplotlib.patches import Polygon
import matplotlib.pyplot as plt
from numpy import random

poly = patches.Circle((5,5), radius=3, fill=False, ec='none')

fig = plt.figure()
ax = fig.add_subplot(121)
ax.autoscale_view(0,0,0)
test = random.rand(10,10)
im = ax.imshow(test)

test1 = random.rand(10,10)
# How to prevent automatic axis scaling?
ax1 = fig.add_subplot(122, sharex=ax, sharey=ax)
im1 = ax1.imshow(test1)

ax.add_patch(poly)
im.set_clip_path(poly)

# Doesn't work b/c poly vertices auto-transformed to display coords by 
add_patch?
ax1.add_patch(poly)
im1.set_clip_path(poly)

# How to extract non-clipped values instead of full array?
clipped_data = im.get_array()

plt.show()