Leftmost column of PNG output is transparent

Hi All,

I'm trying to plot evenly grided data to tiles for google maps overlays.
Each tile needs to be a 256x256 pixel PNG image, without any borders.

The problem that I'm having is that the leftmost column of the resultant PNG
image (saved using FigureCanvasAgg.print_figure) is transparent.

My real code uses basemap because the output has to be in the mercator
projection, but the sample code below is simplified to exclude it. You can
view the sample image that is produced at this address:
http://earth.usc.edu/ftp/kmilner/trans_test.png

Any ideas?

Thanks!
Kevin

#!/usr/bin/env python

from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
from matplotlib.figure import Figure
import numpy

dpi = 64
pixels = 256
fig = Figure(linewidth=0, frameon=False)
canvas = FigureCanvas(fig)

fig.figurePatch.set_alpha(0.0)

ax_bounds = [0, 0, 1, 1]

ax = fig.add_axes(ax_bounds, axisbg=(1.0,1.0,0.0,0.0),
            frameon=False )

inches = pixels / dpi
fig.set_size_inches(256 / dpi, 256 / dpi)

num = 31

z = numpy.empty( (num, num) )

for x in xrange(num):
  for y in xrange(num):
    z[x, y] = x * y

x, y = numpy.meshgrid( range(num), range(num) )

image = ax.pcolor(x, y, z)
image = ax.contourf(x, y, z)

canvas.print_figure("/tmp/trans_test.png", dpi=dpi)

···

--
View this message in context: http://www.nabble.com/Leftmost-column-of-PNG-output-is-transparent-tp22669232p22669232.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

I couldn't reproduce it (my output has no transparent column). I'm
running the current svn. I wonder if others can reproduce it.

Kevin,
what happen if you only do one of the pcolor or the contourf (it is
not clear why you're calling pcolor as it will be overridden by
contourf)?

-JJ

···

On Mon, Mar 23, 2009 at 5:10 PM, kmilner <kmilner@...1887...> wrote:

Hi All,

I'm trying to plot evenly grided data to tiles for google maps overlays.
Each tile needs to be a 256x256 pixel PNG image, without any borders.

The problem that I'm having is that the leftmost column of the resultant PNG
image (saved using FigureCanvasAgg.print_figure) is transparent.

My real code uses basemap because the output has to be in the mercator
projection, but the sample code below is simplified to exclude it. You can
view the sample image that is produced at this address:
http://earth.usc.edu/ftp/kmilner/trans_test.png

Any ideas?

Thanks!
Kevin

#!/usr/bin/env python

from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
from matplotlib.figure import Figure
import numpy

dpi = 64
pixels = 256
fig = Figure(linewidth=0, frameon=False)
canvas = FigureCanvas(fig)

fig.figurePatch.set_alpha(0.0)

ax_bounds = [0, 0, 1, 1]

ax = fig.add_axes(ax_bounds, axisbg=(1.0,1.0,0.0,0.0),
frameon=False )

inches = pixels / dpi
fig.set_size_inches(256 / dpi, 256 / dpi)

num = 31

z = numpy.empty( (num, num) )

for x in xrange(num):
for y in xrange(num):
z[x, y] = x * y

x, y = numpy.meshgrid( range(num), range(num) )

image = ax.pcolor(x, y, z)
image = ax.contourf(x, y, z)

canvas.print_figure("/tmp/trans_test.png", dpi=dpi)
--
View this message in context: http://www.nabble.com/Leftmost-column-of-PNG-output-is-transparent-tp22669232p22669232.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

------------------------------------------------------------------------------
Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
easily build your RIAs with Flex Builder, the Eclipse(TM)based development
software that enables intelligent coding and step-through debugging.
Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
matplotlib-users List Signup and Options

Maybe I'll try the latest SVN version. I'm running the version in the ubuntu 8.10 repositories, which appears to be 0.98.3

As for both contourf and pcolor, I was just testing to see that it was messed up either way, leaving one commented and one uncommented. Having them both uncommented was just a typo, but the problem persists either way.

Thanks,
Kevin

Jae-Joon Lee wrote:

···

I couldn't reproduce it (my output has no transparent column). I'm
running the current svn. I wonder if others can reproduce it.

Kevin,
what happen if you only do one of the pcolor or the contourf (it is
not clear why you're calling pcolor as it will be overridden by
contourf)?

-JJ

On Mon, Mar 23, 2009 at 5:10 PM, kmilner <kmilner@...1887...> wrote:
  

Hi All,

I'm trying to plot evenly grided data to tiles for google maps overlays.
Each tile needs to be a 256x256 pixel PNG image, without any borders.

The problem that I'm having is that the leftmost column of the resultant PNG
image (saved using FigureCanvasAgg.print_figure) is transparent.

My real code uses basemap because the output has to be in the mercator
projection, but the sample code below is simplified to exclude it. You can
view the sample image that is produced at this address:
http://earth.usc.edu/ftp/kmilner/trans_test.png

Any ideas?

Thanks!
Kevin

#!/usr/bin/env python

from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
from matplotlib.figure import Figure
import numpy

dpi = 64
pixels = 256
fig = Figure(linewidth=0, frameon=False)
canvas = FigureCanvas(fig)

fig.figurePatch.set_alpha(0.0)

ax_bounds = [0, 0, 1, 1]

ax = fig.add_axes(ax_bounds, axisbg=(1.0,1.0,0.0,0.0),
                                               frameon=False )

inches = pixels / dpi
fig.set_size_inches(256 / dpi, 256 / dpi)

num = 31

z = numpy.empty( (num, num) )

for x in xrange(num):
       for y in xrange(num):
               z[x, y] = x * y

x, y = numpy.meshgrid( range(num), range(num) )

image = ax.pcolor(x, y, z)
image = ax.contourf(x, y, z)

canvas.print_figure("/tmp/trans_test.png", dpi=dpi)
--
View this message in context: http://www.nabble.com/Leftmost-column-of-PNG-output-is-transparent-tp22669232p22669232.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

------------------------------------------------------------------------------
Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
easily build your RIAs with Flex Builder, the Eclipse(TM)based development
software that enables intelligent coding and step-through debugging.
Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
matplotlib-users List Signup and Options