Repository navigation
Rasterizing patch changes filling of hatches in pdf backend #6228
Description
I want to plot a patch with the '*' hatch that is filled. The hatch is always filled when using
the agg backend but in the pdf backend only when it is rasterized.
As the legend artists do not use the rasterized prop (correct?), it is inconsistent.
mpl: 1.5.1, python 3.5.1, arch linux
Code to reproduce:
import matplotlib
import matplotlib.pyplot as plt
fig = plt.figure()
ax = fig.add_subplot(111)
# patch without hatching
patch1 = matplotlib.patches.Rectangle((0.5, 0.5), 1., 1., color='red', fill=False, hatch='*', label='patch1')
# patch with hatching
patch2 = matplotlib.patches.Rectangle((2.0, 0.5), 1., 1., color='blue', fill=False, hatch='*', rasterized=True, label='patch2')
ax.add_patch(patch1)
ax.add_patch(patch2)
ax.set_xlim(0., 3.5)
ax.set_ylim(0., 3)
ax.legend()
fig.savefig('asdf.png')
fig.savefig('asdf.pdf')
PNG output:
PDF output:
Activity
tacaswell commented on Mar 27, 2016
This may be addressed by #6198
tacaswell commented on Mar 27, 2016
No, not fixed.
In pdf when an artist is rasterized that just means that artist is rendered via the Agg backend (the same as png) and then that is inserted into the pdf (hence why they match). We probably should propagate the rasterized flag to the legend. That isn't fool-proof as the artists maybe rasterized based on zorder as well, but better than nothing.
tacaswell commented on Mar 27, 2016
Of course doing that exposes #6227 ;)
QuLogic commented on Nov 8, 2016
So the legend is kind of unrelated here; it's that Agg renders with a filled hatch pattern when fill=False while PDF does not. Probably either one can be fixed directly rather easily, but which one is the expected behaviour here?
tacaswell commented on Nov 8, 2016
I would go with the Agg behavior as 'correct'.
QuLogic commented on Nov 8, 2016
Well that's good, because PDF's the easier one to fix.
Fill hatches on PDF and PS backends.
NelleV commented on Nov 14, 2016
Closed in #7421