You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
X, Y = meshgrid(linspace(0, 1, 2048), linspace(0, 1, 1024))
Y *= 2 + sin(linspace(0, 2*np.pi, 2048))
C = np.sin(X + Y)
pcolormesh(X, Y, C, rasterized=True)
gcf().savefig('test.eps')
gcf().savefig('test.png')
Saved as eps (converted the eps -> png for upload with inkscape)
saved as png (and matches screen):
My guess is that the agg layer renders the parts of the bounding box around the artist where the quad-mesh isn't as transparent and iirc postscript can't deal with the alpha channel. I hope this is an easy fix (change the background to (1, 1, 1, 0)?), but I don't know the guts of the rendering layer well enough to tell what to change.
The text was updated successfully, but these errors were encountered:
tacaswell
added a commit
to tacaswell/matplotlib
that referenced
this issue
Sep 30, 2013
The change to the agg propagate to how `imsave` works so one test had
to be changed to account for the fact that pixels with alpha=0 are now
set to (1, 1, 1, 0) instead of (0, 0, 0, 0).
* upstream/v1.3.x:
Update version to 1.3.1
tiny correction of matplotlibrc.template
minor docstring fix in dates.py
Updated the docs of pyplot.gca.
Fix#2472: Draw the spines on top of twin axes
changes to c++ because gcc 4.8 seems to allow things the compiler on travis will not.
added class variable `_fill_color` to RendererAgg to hold the color to use in `clear` calls to the render_base.
patch to fix issue #2473.
Conflicts:
lib/matplotlib/__init__.py
lib/matplotlib/axes/_axes.py
This reproduces the problem:
Saved as eps (converted the eps -> png for upload with inkscape)
saved as png (and matches screen):
My guess is that the agg layer renders the parts of the bounding box around the artist where the quad-mesh isn't as transparent and iirc postscript can't deal with the alpha channel. I hope this is an easy fix (change the background to (1, 1, 1, 0)?), but I don't know the guts of the rendering layer well enough to tell what to change.
The text was updated successfully, but these errors were encountered: