Plotting with alpha=0 with rasterized=True causes ValueError on saving to pdf

[henrysky]

Bug report

Bug summary

A recent upgrade to matplotlib 2.2.3 causes a crash when the following conditions are true:

• plot with alpha=0
• rasterized=True
• savefig to pdf.

similar bug report: #12003

Code for reproduction

import numpy as np
from matplotlib.pyplot import *

x = np.random.normal(0,1,5000)
y = np.random.normal(0,1,5000)

ax = subplot(111)
ax.plot(x,y,'g.',rasterized=True, alpha=0.)
savefig("bug.pdf",format='pdf',dpi=300)
show()

Actual outcome

ValueError                                Traceback (most recent call last)
<ipython-input-1-6378414f85f8> in <module>()
      7 ax = subplot(111)
      8 ax.plot(x,y,'g.',rasterized=True, alpha=0.)
----> 9 savefig("bug.pdf",format='pdf',dpi=300)
     10 show()

~\Anaconda3\lib\site-packages\matplotlib\pyplot.py in savefig(*args, **kwargs)
    693 def savefig(*args, **kwargs):
    694     fig = gcf()
--> 695     res = fig.savefig(*args, **kwargs)
    696     fig.canvas.draw_idle()   # need this if 'transparent=True' to reset colors
    697     return res

~\Anaconda3\lib\site-packages\matplotlib\figure.py in savefig(self, fname, **kwargs)
   2060             self.set_frameon(frameon)
   2061
-> 2062         self.canvas.print_figure(fname, **kwargs)
   2063
   2064         if frameon:

~\Anaconda3\lib\site-packages\matplotlib\backends\backend_qt5agg.py in print_figure(self, *args, **kwargs)
     92
     93     def print_figure(self, *args, **kwargs):
---> 94         super(FigureCanvasQTAgg, self).print_figure(*args, **kwargs)
     95         self.draw()
     96

~\Anaconda3\lib\site-packages\matplotlib\backend_bases.py in print_figure(self, filename, dpi, facecolor, edgecolor, orientation, format, **kwargs)
   2261                 orientation=orientation,
   2262                 bbox_inches_restore=_bbox_inches_restore,
-> 2263                 **kwargs)
   2264         finally:
   2265             if bbox_inches and restore_bbox:

~\Anaconda3\lib\site-packages\matplotlib\backends\backend_pdf.py in print_pdf(self, filename, **kwargs)
   2584                 RendererPdf(file, image_dpi, height, width),
   2585                 bbox_inches_restore=_bbox_inches_restore)
-> 2586             self.figure.draw(renderer)
   2587             renderer.finalize()
   2588             if not isinstance(filename, PdfPages):

~\Anaconda3\lib\site-packages\matplotlib\artist.py in draw_wrapper(artist, renderer, *args, **kwargs)
     53                 renderer.start_filter()
     54
---> 55             return draw(artist, renderer, *args, **kwargs)
     56         finally:
     57             if artist.get_agg_filter() is not None:

~\Anaconda3\lib\site-packages\matplotlib\figure.py in draw(self, renderer)
   1491
   1492             mimage._draw_list_compositing_images(
-> 1493                 renderer, self, artists, self.suppressComposite)
   1494
   1495             renderer.close_group('figure')

~\Anaconda3\lib\site-packages\matplotlib\image.py in _draw_list_compositing_images(renderer, parent, artists, suppress_composite)
    139     if not_composite or not has_images:
    140         for a in artists:
--> 141             a.draw(renderer)
    142     else:
    143         # Composite any adjacent images together

~\Anaconda3\lib\site-packages\matplotlib\artist.py in draw_wrapper(artist, renderer, *args, **kwargs)
     53                 renderer.start_filter()
     54
---> 55             return draw(artist, renderer, *args, **kwargs)
     56         finally:
     57             if artist.get_agg_filter() is not None:

~\Anaconda3\lib\site-packages\matplotlib\axes\_base.py in draw(self, renderer, inframe)
   2633             renderer.stop_rasterizing()
   2634
-> 2635         mimage._draw_list_compositing_images(renderer, self, artists)
   2636
   2637         renderer.close_group('axes')

~\Anaconda3\lib\site-packages\matplotlib\image.py in _draw_list_compositing_images(renderer, parent, artists, suppress_composite)
    139     if not_composite or not has_images:
    140         for a in artists:
--> 141             a.draw(renderer)
    142     else:
    143         # Composite any adjacent images together

~\Anaconda3\lib\site-packages\matplotlib\artist.py in draw_wrapper(artist, renderer, *args, **kwargs)
     58                 renderer.stop_filter(artist.get_agg_filter())
     59             if artist.get_rasterized():
---> 60                 renderer.stop_rasterizing()
     61
     62     draw_wrapper._supports_rasterization = True

~\Anaconda3\lib\site-packages\matplotlib\backends\backend_mixed.py in stop_rasterizing(self)
    128
    129             height = self._height * self.dpi
--> 130             buffer, bounds = self._raster_renderer.tostring_rgba_minimized()
    131             l, b, w, h = bounds
    132             if w > 0 and h > 0:

~\Anaconda3\lib\site-packages\matplotlib\backends\backend_agg.py in tostring_rgba_minimized(self)
    142                 [extents[0] + extents[2], self.height - extents[1]]]
    143         region = self.copy_from_bbox(bbox)
--> 144         return np.array(region), extents
    145
    146     def draw_path(self, gc, path, transform, rgbFace=None):

ValueError: negative dimensions are not allowed

Expected outcome

A pdf is saved successfully

Matplotlib version

• Operating system: Win10 1809 x64
• Matplotlib version: 2.2.3
• Matplotlib backend (print(matplotlib.get_backend())): Qt5Agg
• Python version: 3.7.0
• Jupyter version (if applicable):
• Other libraries: numpy 1.5.1

Anaconda 5.3.0 x64

[QuLogic]

Have you tried with 3.0.0? This sounds familiar though; it may be fixed for 3.0.0 or 3.0.1, when it's released.

[QuLogic]

Other libraries: numpy 1.5.1

I think you mean 1.15.1, no? 1.5.1 is rather old and unsupported.

Just tried with 2.2.3 and it fails, with 3.0.0 and it works, and v2.2.x now works, so the fix from #12003 does seems sufficient, so this will be fixed in 2.2.4.

[henrysky]

@QuLogic, yea sorry it is numpy 1.15.1

Can you confirm this case is also fixed

import numpy as np
from matplotlib.pyplot import *

x = np.random.normal(0,1,1)
y = np.random.normal(0,1,1)

ax = subplot(111)
ax.errorbar(x,y, xerr=x, yerr=y, color='blue',rasterized=True)  # Only work if specify marker='...'
savefig("bug.pdf",format='pdf',dpi=300)
show()

[QuLogic]

I do not see any error with that example using 2.2.3.