matplotlib-users Mailing List for matplotlib (Page 2)

Hackstein,
Unfortunately, I'm not sure of an 'elegant' way to do what your asking 
with a single call to scatter. Others may know a better way. However, 
you can use rectangle patches and patch collections. (Requires a bit 
more code than scatter but is ultimately more flexible.)
I think the example below does what you need, but with random numbers.
Hope it helps a little.
Ryan
#######################
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.patches import Rectangle
from matplotlib.collections import PatchCollection
n = 100
# Get your xy data points, which are the centers of the rectangles.
xy = np.random.rand(n,2)
# Set a fixed height
height = 0.02
# The variable widths of the rectangles
widths = np.random.rand(n)*0.1
# Get a color map and color values (normalized between 0 and 1)
cmap = plt.cm.jet
colors = np.random.rand(n)
rects = []
for p, w, c in zip(xy, widths, colors):
 xpos = p[0] - w/2 # The x position will be half the width from the 
center
 ypos = p[1] - height/2 # same for the y position, but with height
 rect = Rectangle( (xpos, ypos), w, height ) # Create a rectangle
 rects.append(rect) # Add the rectangle patch to our list
# Create a collection from the rectangles
col = PatchCollection(rects)
# set the alpha for all rectangles
col.set_alpha(0.3)
# Set the colors using the colormap
col.set_facecolor( cmap(colors) )
# Make a figure and add the collection to the axis.
ax = plt.subplot(111)
ax.add_collection(col)
plt.show()
###############################
On 4/24/2013 5:35 PM, Hackstein wrote:
>
> Hi all,
>
> I am trying to get a scatter plot using a colormap. Additionally, I 
> need to define every marker for every data point individually -- each 
> being a rectangle with fixed height but varying width as a function of 
> the y-value. X and y being the data coordinates, z being a number to 
> be color coded with the colormap.
>
> Ideally, I would like to create a list of width and height values for 
> each data point and tell the scatter plot to use those.
>
> So far I got colormapped data with custom markers (simplified):
>
> [code]
>
> import numpy as np
>
> import matplotlib.pyplot as plt
>
> from pylab import *
>
> x = y = [1,2,3,4,5]
>
> z = [2,4,6,8,10]
>
> colors = cm.gnuplot2
>
> verts_vec = list(zip([-10.,10.,10.,-10.],[-5.,-5.,5.,5.]))
>
> fig = plt.figure(1, figsize=(14.40, 9.00))
>
> ax = fig.add_subplot(1,1,1)
>
> sc = ax.scatter(x, y, c=np.asarray(z), marker=None, edgecolor='None', 
> verts=verts_vec, cmap=colors, alpha=1.)
>
> plt.colorbar(sc, orientation='horizontal')
>
> plt.savefig('test.png', dpi=200)
>
> plt.close(1)
>
> [/code]
>
> But I need to define a marker size for each point, and I also need to 
> do that in axis scale values, not in points.
>
> I imagine giving verts a list of N*2 tuples instead of 2 tuples, N 
> being len(x), to define N individual markers.
>
> But when doing that I get the error that vertices.ndim==2.
>
> A less elegant way would be to plot every data point in an individual 
> scatter plot function, using a for-loop iterating over all data 
> points. Then, however, I see no way to apply a colormap and colorbar.
>
> What is the best way to accomplish that then?
>
> Thanks,
>
> -Hackstein
>
>
>
> ------------------------------------------------------------------------------
> Try New Relic Now & We'll Send You this Cool Shirt
> New Relic is the only SaaS-based application performance monitoring service
> that delivers powerful full stack analytics. Optimize and monitor your
> browser, app, & servers with just a few lines of code. Try New Relic
> and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

I believe this PR fixes this bug:
https://github.com/matplotlib/matplotlib/pull/1884
I had been waiting for the original poster to confirm before merging, 
but I think I'll go ahead and do this anyway at this point.
Mike
On 04/23/2013 02:57 PM, Nils Wagner wrote:
> Hi all,
>
> I cannot install matplotlib. Please find enclosed the logfile of
> python setup.py install --prefix=$HOME/local >& log.txt
>
> Any idea how to resolve the problem is appreciated.
>
> Thanks in advance.
> Nils
>
>
>
> ------------------------------------------------------------------------------
> Try New Relic Now & We'll Send You this Cool Shirt
> New Relic is the only SaaS-based application performance monitoring service
> that delivers powerful full stack analytics. Optimize and monitor your
> browser, app, & servers with just a few lines of code. Try New Relic
> and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Hi all,
 
I am trying to get a scatter plot using a colormap. Additionally, I need to
define every marker for every data point individually - each being a
rectangle with fixed height but varying width as a function of the y-value.
X and y being the data coordinates, z being a number to be color coded with
the colormap.
Ideally, I would like to create a list of width and height values for each
data point and tell the scatter plot to use those.
So far I got colormapped data with custom markers (simplified):
 
[code]
 
import numpy as np
import matplotlib.pyplot as plt
from pylab import *
 
x = y = [1,2,3,4,5]
z = [2,4,6,8,10]
 
colors = cm.gnuplot2
 
verts_vec = list(zip([-10.,10.,10.,-10.],[-5.,-5.,5.,5.]))
 
fig = plt.figure(1, figsize=(14.40, 9.00))
ax = fig.add_subplot(1,1,1)
sc = ax.scatter(x, y, c=np.asarray(z), marker=None, edgecolor='None',
verts=verts_vec, cmap=colors, alpha=1.)
plt.colorbar(sc, orientation='horizontal')
plt.savefig('test.png', dpi=200)
plt.close(1)
 
[/code]
 
But I need to define a marker size for each point, and I also need to do
that in axis scale values, not in points.
I imagine giving verts a list of N*2 tuples instead of 2 tuples, N being
len(x), to define N individual markers.
But when doing that I get the error that vertices.ndim==2.
A less elegant way would be to plot every data point in an individual
scatter plot function, using a for-loop iterating over all data points.
Then, however, I see no way to apply a colormap and colorbar.
What is the best way to accomplish that then?
 
Thanks, 
 
-Hackstein
 
 
 
 

Not quite a module, but I think it does what you want:
http://matplotlib.1069221.n5.nabble.com/function-to-create-a-colormap-from-cpt-palette-td2165.html
Cheers, Jody
On Apr 24, 2013, at 9:16 AM, Andreas Hilboll <li...@hi...> wrote:
> Hi,
> I'd like to have the cpt-city colormaps available in matplotlib. Is
> there already a module for this? If not, I'll try to patch something
> together. What do you think?
> Cheers, Andreas
> 
> ------------------------------------------------------------------------------
> Try New Relic Now & We'll Send You this Cool Shirt
> New Relic is the only SaaS-based application performance monitoring service 
> that delivers powerful full stack analytics. Optimize and monitor your
> browser, app, & servers with just a few lines of code. Try New Relic
> and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_apr
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
--
Jody Klymak 
http://web.uvic.ca/~jklymak/

Hi,
I'd like to have the cpt-city colormaps available in matplotlib. Is
there already a module for this? If not, I'll try to patch something
together. What do you think?
Cheers, Andreas

Hi all,
I cannot install matplotlib. Please find enclosed the logfile of
python setup.py install --prefix=$HOME/local >& log.txt
Any idea how to resolve the problem is appreciated.
Thanks in advance.
 Nils

Hi all-
I've been working on a plot that puts the bottom and right spines at zero
(adapting some code from the example at
http://matplotlib.org/examples/pylab_examples/spine_placement_demo.html) ,
and I've discovered that setting the position of the right spine to 'zero'
seems to locate it at x=1.
Is this a bug? Or is there something that I'm missing?
Thanks,
--Chad
Here's an example that demonstrates the behavior:
import matplotlib.pyplot as plt
f=plt.figure(1)
ax=plt.subplot(111)
ax.spines['left'].set_color('none')
ax.spines['top'].set_color('none')
ax.spines['right'].set_position('zero')
ax.spines['bottom'].set_position('zero')
ax.xaxis.set_ticks_position('bottom')
ax.yaxis.set_ticks_position('right')
ax.set_xlim([-2,2])
ax.set_ylim([-2,2])
plt.show()

Hi all,
is there any possibility to show only first quadrant in hammer
projection? If it is not implemented in matplotlib, have you any trick
for doing this?
Thanks in advance for your help. 
Best, 
Marian

Hey everyone,
Over on IRC (#scipy channel on freenode), Baribal asked this:
 <Baribal> I'm computing a dendrite on a continuous surface ([0,
 1[, [0, 1[) and want to visualize it. What module would you
 recommend? In practice, I need to draw lines from x_1/y_1 to
 x_2/y_2 with a color gradient applied.
And further specified that the this is just for straight lines,
and each point also has an RGB value associated with it.
I coded up a solution (plot_gradient_rbg_pairs), and thought I'd
share it here. I've also posted it to a gist, in case I end up
updating it later. https://gist.github.com/5439438
#!/usr/bin/env python
"""A quick hack to draw gradient lines using a colormap.
This was written in response to <Baribal>'s question on IRC.
There are two functions provided here:
`plot_gradient_hack` takes two arguments, p0 and p1, which are both (x,y)
pairs, and plots a gradient between them that spans the full colormap.
`plot_gradient_rbg_pairs` does the same thing, but also takes rgb0 and rgb1
arguments, makes a new colormap that spans between those two values, and uses
that colormap for the plot.
There's an alternative solution over here [1], but that uses many more points.
1. http://matplotlib.1069221.n5.nabble.com/Gradient-color-on-a-line-plot-td17643.html
"""
import numpy as np
import matplotlib.pyplot as plt
import matplotlib
from matplotlib.colors import LinearSegmentedColormap
def plot_gradient_hack( p0, p1, npts=20, cmap=None, **kw):
 """
 Draw a gradient between p0 and p1 using a colormap
 The **kw dictionary gets passed to plt.plot, so things like linestyle,
 linewidth, labels, etc can be modified directly.
 """
 x_1, y_1 = p0
 x_2, y_2 = p1
 
 X = np.linspace(x_1, x_2, npts)
 Xs = X[:-1]
 Xf = X[1:]
 Xpairs = zip(Xs, Xf)
 
 Y = np.linspace(y_1, y_2, npts)
 Ys = Y[:-1]
 Yf = Y[1:]
 Ypairs = zip(Ys, Yf)
 C = np.linspace(0,1, npts)
 cmap = plt.get_cmap(cmap)
 # the simplest way of doing this is to just do the following:
 for x, y, c in zip(Xpairs, Ypairs, C):
 plt.plot(x, y, '-', c=cmap(c), **kw)
 # But for cases when that will be too slow, you can make this go faster,
 # follow along with this example:
 # http://matplotlib.org/examples/pylab_examples/line_collection2.html
def plot_gradient_rbg_pairs(p0, p1, rgb0, rgb1, **kw):
 """Form the gradient from RGB values at each point
 The **kw dictionary gets passed to plt.plot, so things like linestyle,
 linewidth, labels, etc can be modified directly.
 """
 cmap = LinearSegmentedColormap.from_list('tmp', (rgb0, rgb1))
 plot_gradient_hack(p0, p1, cmap=cmap, **kw)
# plot gradient that just spans the full colormap
plot_gradient_hack( (1,2), (5,6) )
# we can specify the colormap, and set some properties for the plot
plot_gradient_hack( (2,5), (5,3), cmap='bwr', linewidth=3.)
# We also have a simple wrapper to specify the two rgb points to interpolate
# the gradient between
plot_gradient_rbg_pairs( (1.1,2), (5.1,6), (0,0,0), (1,1,1) ) # black to white
plot_gradient_rbg_pairs( (1.2,2), (5.2,6), (0,0,0), (0,0,1), # black to blue
 linestyle='--', linewidth=9) 
plot_gradient_rbg_pairs( (1.3,2), (5.3,6), (1,0,0), (0,1,0), # red to green
 linewidth=4 )
plt.show()
# we can use this gradient plot to display all colormaps on one plot easily
plt.figure()
with matplotlib.rc_context({'lines.solid_capstyle':'butt'}):
 # the default projecting capstyle looks kind of ugly. rc_context was
 # introduced in matpltolib 1.2.0, if you are running a version older than
 # that, you can ignore this line and remove one level of indentation from
 # the for loop
 for i, map_name in enumerate(plt.cm.cmap_d):
 plot_gradient_hack((0, i), (1, i), cmap = map_name, linewidth=4)
 plt.text(1,i, map_name, va='center')
 # comment out this last line to plot all ~140 colormaps
 if i==25: break
plt.show()
best,
-- 
Paul Ivanov
314 address only used for lists, off-list direct email at:
http://pirsquared.org | GPG/PGP key id: 0x0F3E28F7 

Folks,
I am trying to make a plot with axes rotated by 45 degrees, so that the 
plot looks like a romb. I set the rotating transform for the subplot, 
but it is still plotted in a normal orientation. Could someone tell my 
why the set_transform function does not work?
Many thanks,
Nick Gnedin
import matplotlib.pyplot as plt
from matplotlib.transforms import Affine2D
tr = Affine2D().rotate_deg(45)
fig = plt.figure()
ax = fig.add_subplot(111)
ax.set_transform(tr)
fig.show()

The problem was mix axes and pyplot. Plt.savefig was locking the app.
I removed all pyplot reference and changed to self.axes.xxx() and
self.fig.savefig.(xxx). Worked.
--
View this message in context: http://matplotlib.1069221.n5.nabble.com/After-close-the-plot-window-the-process-keeps-running-tp40919p40940.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

many thanks Paul,
this did work:
I uninstalled all python in my computer:
sudo rm -rf python*
sudo rm -rf /Library/Frameworks/Python.framework/Versions/2.7
sudo rm -rf "/Applications/Python 2.7"
and installed python and matplotlib afresh. Numby had been installed
Python 2.7.4 (v2.7.4:026ee0057e2d, Apr 6 2013, 10:15:50) 
[GCC 4.0.1 (Apple Inc. build 5493)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy as np
>>> import pylab as pl
>>> x=[1,2,3,4]
>>> y=[2,3,4,5]
>>> pl.plot(x,y)
[<matplotlib.lines.Line2D object at 0x50b8990>]
>>> pl.show()
Thank you,
Charles
On 22 Apr 2013, at 06:50, Paul Ivanov <piv...@gm...> wrote:
> Hi Charles,
> 
> I'm sending my reply to the matplotlib-users list - please follow
> up there, so that others benefit from your experience, or folks
> who have run into the same issue can also pitch in with their
> help. 
> 
> Here's where to go to sign up: 
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
> Once you're signed up, you hit "reply all" to this email (it
> would be kind of you to remove my name from the addressee list
> when you do that, so that the email goes only to the
> matplotlib-users list, of which I'm a member)
> 
> See the rest of my reply below your original message.
> 
> cny...@gm..., on 2013年04月21日 09:43, wrote:
>> Hello Paul
>> 
>> I wished to use matplotlip, but I am facing diffuculties and I
>> was wondering if you can help.
>> 
>> Matplot lib is installed but pylab has the following error. I
>> will appreciate a suggestion if you have some please
>> 
>> Charles
>> 
>> 
>> 
>>>>> import numpy as np
>>>>> import pylab as pl
>> Traceback (most recent call last):
>> File "<stdin>", line 1, in <module>
>> File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/pylab.py", line 1, in <module>
>> from matplotlib.pylab import *
>> File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/pylab.py", line 222, in <module>
>> from matplotlib import mpl # pulls in most modules
>> File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/mpl.py", line 1, in <module>
>> from matplotlib import artist
>> File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/artist.py", line 7, in <module>
>> from transforms import Bbox, IdentityTransform, TransformedBbox, \
>> File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/transforms.py", line 35, in <module>
>> from matplotlib._path import (affine_transform, count_bboxes_overlapping_bbox,
>> ImportError: dlopen(/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/_path.so, 2): no suitable image found. Did find:
>> 	/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/_path.so: no matching architecture in universal wrapper
> 
> This looks like conflicting versions of python are being used (64
> bit and 32 bit). One trick for getting to the bottom of Python
> stack traces is to take the text of the *last* line of the
> traceback, remove all but the last few bits of the file path
> (since other folks won't necessarily be installing to the extact
> path you've installeed to), and feed it to google. 
> 
> The first google hit for "_path.so: no matching architecture in
> universal wrapper" is this stackoverflow post:
> http://stackoverflow.com/questions/5419439/matplotlib-pyplot-on-os-x-with-64-bit-python-from-python-org
> 
> See if the solution suggested there (installing 32-bit version
> from python.org)
> 
> best,
> -- 
> Paul Ivanov
> 314 address only used for lists, off-list direct email at:
> http://pirsquared.org | GPG/PGP key id: 0x0F3E28F7 

Hi Charles,
I'm sending my reply to the matplotlib-users list - please follow
up there, so that others benefit from your experience, or folks
who have run into the same issue can also pitch in with their
help. 
Here's where to go to sign up: 
https://lists.sourceforge.net/lists/listinfo/matplotlib-users
Once you're signed up, you hit "reply all" to this email (it
would be kind of you to remove my name from the addressee list
when you do that, so that the email goes only to the
matplotlib-users list, of which I'm a member)
See the rest of my reply below your original message.
cny...@gm..., on 2013年04月21日 09:43, wrote:
> Hello Paul
> 
> I wished to use matplotlip, but I am facing diffuculties and I
> was wondering if you can help.
> 
> Matplot lib is installed but pylab has the following error. I
> will appreciate a suggestion if you have some please
> 
> Charles
> 
> 
> 
> >>> import numpy as np
> >>> import pylab as pl
> Traceback (most recent call last):
> File "<stdin>", line 1, in <module>
> File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/pylab.py", line 1, in <module>
> from matplotlib.pylab import *
> File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/pylab.py", line 222, in <module>
> from matplotlib import mpl # pulls in most modules
> File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/mpl.py", line 1, in <module>
> from matplotlib import artist
> File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/artist.py", line 7, in <module>
> from transforms import Bbox, IdentityTransform, TransformedBbox, \
> File "/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/transforms.py", line 35, in <module>
> from matplotlib._path import (affine_transform, count_bboxes_overlapping_bbox,
> ImportError: dlopen(/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/_path.so, 2): no suitable image found. Did find:
> 	/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/site-packages/matplotlib/_path.so: no matching architecture in universal wrapper
This looks like conflicting versions of python are being used (64
bit and 32 bit). One trick for getting to the bottom of Python
stack traces is to take the text of the *last* line of the
traceback, remove all but the last few bits of the file path
(since other folks won't necessarily be installing to the extact
path you've installeed to), and feed it to google. 
The first google hit for "_path.so: no matching architecture in
universal wrapper" is this stackoverflow post:
http://stackoverflow.com/questions/5419439/matplotlib-pyplot-on-os-x-with-64-bit-python-from-python-org
See if the solution suggested there (installing 32-bit version
from python.org)
best,
-- 
Paul Ivanov
314 address only used for lists, off-list direct email at:
http://pirsquared.org | GPG/PGP key id: 0x0F3E28F7 

Just curious -- where is the formula for matplotlib in homebrew? I 
can't find it. I thought I would look into why that was failing -- it 
may just be simply that it's an old version of matplotlib and this bug 
is now fixed in the latest release.
Mike
On 04/20/2013 11:12 PM, Derek Thomas wrote:
> I was able to fix this by uninstalling the matplotlib from homebrew 
> and installing with pip.
>
>
> On Sat, Apr 20, 2013 at 9:33 AM, Derek Thomas <der...@gm... 
> <mailto:der...@gm...>> wrote:
>
> This may be known, but the following modified example from
> http://matplotlib.org/mpl_toolkits/mplot3d/tutorial.html fails
> with a TypeError at matplotlib/backends/backend_pdf.pyc in
> draw_path_collection. Is it possible to save pdf files with
> surface plots?
>
> from mpl_toolkits.mplot3d import Axes3D
> from matplotlib import cm
> from matplotlib.ticker import LinearLocator, FormatStrFormatter
> import matplotlib.pyplot as plt
> import numpy as np
>
> fig = plt.figure()
> ax = fig.gca(projection='3d')
> X = np.arange(-5, 5, 0.25)
> Y = np.arange(-5, 5, 0.25)
> X, Y = np.meshgrid(X, Y)
> R = np.sqrt(X**2 + Y**2)
> Z = np.sin(R)
> surf = ax.plot_surface(X, Y, Z, rstride=1, cstride=1,
> cmap=cm.coolwarm,
> linewidth=0, antialiased=False)
> ax.set_zlim(-1.01, 1.01)
>
> ax.zaxis.set_major_locator(LinearLocator(10))
> ax.zaxis.set_major_formatter(FormatStrFormatter('%.02f'))
>
> fig.colorbar(surf, shrink=0.5, aspect=5)
> fig.savefig('test.pdf')
> plt.show()
>
>
>
>
> ------------------------------------------------------------------------------
> Precog is a next-generation analytics platform capable of advanced
> analytics on semi-structured data. The platform includes APIs for building
> apps and a phenomenal toolset for data science. Developers can use
> our toolset for easy data analysis & visualization. Get a free account!
> http://www2.precog.com/precogplatform/slashdotnewsletter
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

I was able to fix this by uninstalling the matplotlib from homebrew and
installing with pip.
On Sat, Apr 20, 2013 at 9:33 AM, Derek Thomas <der...@gm...>wrote:
> This may be known, but the following modified example from
> http://matplotlib.org/mpl_toolkits/mplot3d/tutorial.html fails with a
> TypeError at matplotlib/backends/backend_pdf.pyc in draw_path_collection.
> Is it possible to save pdf files with surface plots?
>
> from mpl_toolkits.mplot3d import Axes3D
> from matplotlib import cm
> from matplotlib.ticker import LinearLocator, FormatStrFormatter
> import matplotlib.pyplot as plt
> import numpy as np
>
> fig = plt.figure()
> ax = fig.gca(projection='3d')
> X = np.arange(-5, 5, 0.25)
> Y = np.arange(-5, 5, 0.25)
> X, Y = np.meshgrid(X, Y)
> R = np.sqrt(X**2 + Y**2)
> Z = np.sin(R)
> surf = ax.plot_surface(X, Y, Z, rstride=1, cstride=1, cmap=cm.coolwarm,
> linewidth=0, antialiased=False)
> ax.set_zlim(-1.01, 1.01)
>
> ax.zaxis.set_major_locator(LinearLocator(10))
> ax.zaxis.set_major_formatter(FormatStrFormatter('%.02f'))
>
> fig.colorbar(surf, shrink=0.5, aspect=5)
> fig.savefig('test.pdf')
> plt.show()
>
>

This may be known, but the following modified example from
http://matplotlib.org/mpl_toolkits/mplot3d/tutorial.html fails with a
TypeError at matplotlib/backends/backend_pdf.pyc in draw_path_collection.
 Is it possible to save pdf files with surface plots?
from mpl_toolkits.mplot3d import Axes3D
from matplotlib import cm
from matplotlib.ticker import LinearLocator, FormatStrFormatter
import matplotlib.pyplot as plt
import numpy as np
fig = plt.figure()
ax = fig.gca(projection='3d')
X = np.arange(-5, 5, 0.25)
Y = np.arange(-5, 5, 0.25)
X, Y = np.meshgrid(X, Y)
R = np.sqrt(X**2 + Y**2)
Z = np.sin(R)
surf = ax.plot_surface(X, Y, Z, rstride=1, cstride=1, cmap=cm.coolwarm,
 linewidth=0, antialiased=False)
ax.set_zlim(-1.01, 1.01)
ax.zaxis.set_major_locator(LinearLocator(10))
ax.zaxis.set_major_formatter(FormatStrFormatter('%.02f'))
fig.colorbar(surf, shrink=0.5, aspect=5)
fig.savefig('test.pdf')
plt.show()

On 04/19/2013 03:26 AM, Mark Lawrence wrote:
> On 19/04/2013 04:03, John Ladasky wrote:
>
>> Reading more, I realize that the way I was getting GUI output previously
>> (with Python 2.7 and Matplotlib 1.1) was through wxPython.
>> Unfortunately, it appears that wxPython's star is fading, and a Python
>> 3-compatible version will not be written. In fact, wxPython hasn't
>> released a new version in nine months.
>>
> I'm surprised that you say this as months of work have gone into
> updating wxPython to make in Python 3 compatible. Please see
> http://wxpython.org/Phoenix/snapshot-builds/ for the latest and greatest.
>
Well, that is good news. Just before my last message to this list, I 
visited wxpython.org, as I have done many times over the years. I saw no 
announcement that there would be a Python 3.x-compatible version. You 
would think that the developers would want everyone to know!

Hi,
I am trying to get matplotlib 1.2.0 to work with wxPython Phoenix - will 
provide a patch when it is working.
Made the changes to backend_wx* for things like EmptyImage/EmptyBitmap 
and Toolbar but I am stuck on the following.
 if bbox is None:
 # agg => rgba buffer -> bitmap
 if 'phoenix' in wx.PlatformInfo:
 return wx.Bitmap.FromBufferRGBA(int(agg.width), 
int(agg.height),
memoryview(agg.buffer_rgba()))
 else:
 return wx.BitmapFromBufferRGBA(int(agg.width), int(agg.height),
 agg.buffer_rgba())
 else:
 # agg => rgba buffer -> bitmap => clipped bitmap
 return _WX28_clipped_agg_as_bitmap(agg, bbox)
TypeError: cannot make memory view because object does not have the 
buffer interface
File "h:\devProjectsT\aaTests\matplotlib\wxembedding-5.py", line 63, in 
<module>
 demo()
File "h:\devProjectsT\aaTests\matplotlib\wxembedding-5.py", line 60, in demo
 app.MainLoop()
File "c:\Python27\Lib\site-packages\wx-2.9.6-msw-phoenix\wx\core.py", 
line 1841, in MainLoop
 rv = wx.PyApp.MainLoop(self)
File "c:\Python27\Lib\site-packages\matplotlib\backends\backend_wx.py", 
line 1209, in _onPaint
 self.draw(drawDC=drawDC)
File 
"c:\Python27\Lib\site-packages\matplotlib\backends\backend_wxagg.py", 
line 61, in draw
 self.bitmap = _convert_agg_to_wx_bitmap(self.get_renderer(), None)
File 
"c:\Python27\Lib\site-packages\matplotlib\backends\backend_wxagg.py", 
line 173, in _convert_agg_to_wx_bitmap
 memoryview(agg.buffer_rgba()))
I tried using "memoryview" based on a suggestion by Robin Dunn, and 
based on the following info I see in the debugger that should work no?
agg.buffer_rgba()
<read-write buffer ptr 0x05400638, size 229200 at 0x055FC680>
type(agg.buffer_rgba())
<type 'buffer'>
agg
<matplotlib.backends.backend_agg.RendererAgg instance at 0x04BA0670>
If I don't use "memoryview" (which would probably be preferred) I get 
the following exception.
Can someone help us figure this one out.
Thanks
Werner
TypeError: Bitmap.FromBufferRGBA(): argument 3 has unexpected type 'buffer'
File "h:\devProjectsT\aaTests\matplotlib\wxembedding-5.py", line 63, in 
<module>
 demo()
File "h:\devProjectsT\aaTests\matplotlib\wxembedding-5.py", line 60, in demo
 app.MainLoop()
File "c:\Python27\Lib\site-packages\wx-2.9.6-msw-phoenix\wx\core.py", 
line 1841, in MainLoop
 rv = wx.PyApp.MainLoop(self)
File "c:\Python27\Lib\site-packages\matplotlib\backends\backend_wx.py", 
line 1209, in _onPaint
 self.draw(drawDC=drawDC)
File 
"c:\Python27\Lib\site-packages\matplotlib\backends\backend_wxagg.py", 
line 61, in draw
 self.bitmap = _convert_agg_to_wx_bitmap(self.get_renderer(), None)
File 
"c:\Python27\Lib\site-packages\matplotlib\backends\backend_wxagg.py", 
line 173, in _convert_agg_to_wx_bitmap
 agg.buffer_rgba())

Hello,
Is there a way to find out the optimal resolution that an array (of a given
aspect ratio) should have, so that imshow will not re-scale it on a
pixel-based backend?
Some background: I'm preparing an array that has a native resolution, so for
PDF output I use imshow with interpolation='none'. For PNG-output, however,
the array has to be rescaled. I would prefer to prepare the array in the
correct resolution right away.
Thanks in advance,
Christoph

On 04/19/2013 01:59 AM, C M wrote:
>
>
> On Thu, Apr 18, 2013 at 11:03 PM, John Ladasky 
> <joh...@sb... <mailto:joh...@sb...>> wrote:
>
> .
> Reading more, I realize that the way I was getting GUI output
> previously
> (with Python 2.7 and Matplotlib 1.1) was through wxPython.
> Unfortunately, it appears that wxPython's star is fading, and a Python
> 3-compatible version will not be written. In fact, wxPython hasn't
> released a new version in nine months.
>
>
> wxPython is alive and well and the newest developmental version of it 
> ("Phoenix") runs on Python 3. It should be released fairly soon. One 
> of the wxPython list regulars mentioned getting his software to run 
> with it, with a few minor issues, just six days ago. So you might 
> want to give Phoenix a try.
That's great news. Unfortunately, I'm not aware of any testing with it 
and matplotlib on Python 3, so for someone who just wants to get things 
to work, I would still recommend Tk, gtk or Qt4 on Python3.
Mike

On 19/04/2013 12:26, Mark Lawrence wrote:
> On 19/04/2013 04:03, John Ladasky wrote:
>
>> Reading more, I realize that the way I was getting GUI output previously
>> (with Python 2.7 and Matplotlib 1.1) was through wxPython.
>> Unfortunately, it appears that wxPython's star is fading, and a Python
>> 3-compatible version will not be written. In fact, wxPython hasn't
>> released a new version in nine months.
>>
> I'm surprised that you say this as months of work have gone into
> updating wxPython to make in Python 3 compatible. Please see
> http://wxpython.org/Phoenix/snapshot-builds/ for the latest and greatest.
This one to the list - sorry Mark used the wrong reply button.
I don't work on Python 3 yet but I am trying to get Phoenix to work with 
matplotlib.
Made a few changes to backend_wx:
Following don't exist in Phoenix (build r73823) - will research later 
if they should exist.
 #wx.WXK_PRIOR : 'pageup',
 #wx.WXK_NEXT : 'pagedown',
 #wx.WXK_NUMPAD_PRIOR : 'pageup',
 #wx.WXK_NUMPAD_NEXT : 'pagedown',
There are a few EmptyBitMap creates which need to be changed to:
 if 'phoenix' in wx.PlatformInfo:
 self.bitmap = wx.Bitmap(w, h)
 else:
 self.bitmap = wx.EmptyBitmap(w, h)
But now I am stuck on the following exception and I haven't found a 
solution to this yet - any pointers would be very welcome.
Werner
AttributeError: 'BaseApp' object has no attribute 'ProcessIdle'
File "c:\dev\twcbv4\twcbsrc\controllers\app_stats.py", line 864, in <module>
 apprb = Appstats(None, standalone=True)
File "c:\dev\twcbv4\twcbsrc\controllers\app_stats.py", line 203, in __init__
 self.setupControls()
File "c:\dev\twcbv4\twcbsrc\controllers\app_stats.py", line 236, in 
setupControls
 self.createStatsPanes()
File "c:\dev\twcbv4\twcbsrc\controllers\app_stats.py", line 401, in 
createStatsPanes
 self.paneStatsDtypeCons.axes = 
self.paneStatsDtypeCons.figure.add_subplot(noRow, noCol, 1)
File "c:\Python27\Lib\site-packages\matplotlib\figure.py", line 882, in 
add_subplot
 a = subplot_class_factory(projection_class)(self, *args, **kwargs)
File "c:\Python27\Lib\site-packages\matplotlib\axes.py", line 8938, in 
__init__
 self._axes_class.__init__(self, fig, self.figbox, **kwargs)
File "c:\Python27\Lib\site-packages\matplotlib\axes.py", line 461, in 
__init__
 self._init_axis()
File "c:\Python27\Lib\site-packages\matplotlib\axes.py", line 523, in 
_init_axis
 self.spines['right'].register_axis(self.yaxis)
File "c:\Python27\Lib\site-packages\matplotlib\spines.py", line 151, in 
register_axis
 self.axis.cla()
File "c:\Python27\Lib\site-packages\matplotlib\axis.py", line 732, in cla
 self.reset_ticks()
File "c:\Python27\Lib\site-packages\matplotlib\axis.py", line 746, in 
reset_ticks
 self.minorTicks.extend([self._get_tick(major=False)])
File "c:\Python27\Lib\site-packages\matplotlib\axis.py", line 1839, in 
_get_tick
 return YTick(self.axes, 0, '', major=major, **tick_kw)
File "c:\Python27\Lib\site-packages\matplotlib\axis.py", line 140, in 
__init__
 self.tick2line = self._get_tick2line()
File "c:\Python27\Lib\site-packages\matplotlib\axis.py", line 541, in 
_get_tick2line
 l.set_transform(self.axes.get_yaxis_transform(which='tick2'))
File "c:\Python27\Lib\site-packages\matplotlib\lines.py", line 476, in 
set_transform
 Artist.set_transform(self, t)
File "c:\Python27\Lib\site-packages\matplotlib\artist.py", line 235, in 
set_transform
 self.pchanged()
File "c:\Python27\Lib\site-packages\matplotlib\artist.py", line 216, in 
pchanged
 for oid, func in self._propobservers.iteritems():

On 19/04/2013 04:03, John Ladasky wrote:
> Reading more, I realize that the way I was getting GUI output previously
> (with Python 2.7 and Matplotlib 1.1) was through wxPython.
> Unfortunately, it appears that wxPython's star is fading, and a Python
> 3-compatible version will not be written. In fact, wxPython hasn't
> released a new version in nine months.
>
I'm surprised that you say this as months of work have gone into 
updating wxPython to make in Python 3 compatible. Please see 
http://wxpython.org/Phoenix/snapshot-builds/ for the latest and greatest.
-- 
If you're using GoogleCrapTM please read this 
http://wiki.python.org/moin/GoogleGroupsPython.
Mark Lawrence

Hi John,
on Kubuntu Precise the standard repo has at least:
python3-pyqt4
python3-pyside
python3-tk
The first two should enable Qt4Agg backend, the last TkAgg
Fra
2013年4月19日 Sterling Smith <sm...@fu...>
> I have used the TkAgg backend in python2, installing the dependencies by
> hand. Is this backend not available for python3?
>
> -Sterling
>
> On Apr 18, 2013, at 8:03PM, John Ladasky wrote:
>
> > Thanks to both Francesco Montesano and Benjamin Root. I have done some
> > reading. And I have made some progress, though I am not quite where I
> > want to be yet.
> >
> > So the problem appears to be that the only backend for which I had
> > suitable Python 3 libraries was agg. It only requires libpng, which I
> > have. I can render a Matplotlib canvas, but it appears that the only
> > output that agg offers is in the form of PNG files to disk. I cannot
> > create a live window on the screen.
> >
> > Reading more, I realize that the way I was getting GUI output previously
> > (with Python 2.7 and Matplotlib 1.1) was through wxPython.
> > Unfortunately, it appears that wxPython's star is fading, and a Python
> > 3-compatible version will not be written. In fact, wxPython hasn't
> > released a new version in nine months.
> >
> > The other choices for Matplotlib GUI output on Linux appear to be
> > through GTK, PySide, and PyQt. I am not familiar with GTK, but I know
> > that it is widely-used. Also, GTK appears to be Python 3-compatible,
> > and so that is where I need to go.
> >
> > I'm going through a trial and error process. Unfortunately, the names
> > of the repositories in Ubuntu are not very helpful. I installed a few
> > GTK and python-gtk related packages that I thought were relevant. On my
> > first build attempt I got no errors, but also, I didn't get a GTKAgg
> > backend. Upon re-reading, I saw that I should modify matplotlib's
> > setup.cfg file to force a GTK build attempt, and to report errors if it
> > fails. That's what it does. In the "optional backend dependencies"
> > section I am not seeing any GTK libraries listed, even though I have
> > installed python-gtk2-dev (2.24.0), python-gobject-2-dev (2.28.6),
> > libgtk2.0-dev (2.24.10), libglib2.0-dev (2.32.3), python-gi-dev (3.2.2),
> > python-gobject-dev (3.2.2), python3-gi (3.2.2), and a few DOZEN packages
> > on which these depend.
> >
> > If anyone knows the way forward from here, I would appreciate your
> > advice. Thanks again.
> >
> >
> ------------------------------------------------------------------------------
> > Precog is a next-generation analytics platform capable of advanced
> > analytics on semi-structured data. The platform includes APIs for
> building
> > apps and a phenomenal toolset for data science. Developers can use
> > our toolset for easy data analysis & visualization. Get a free account!
> > http://www2.precog.com/precogplatform/slashdotnewsletter
> > _______________________________________________
> > Matplotlib-users mailing list
> > Mat...@li...
> > https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>
>
> ------------------------------------------------------------------------------
> Precog is a next-generation analytics platform capable of advanced
> analytics on semi-structured data. The platform includes APIs for building
> apps and a phenomenal toolset for data science. Developers can use
> our toolset for easy data analysis & visualization. Get a free account!
> http://www2.precog.com/precogplatform/slashdotnewsletter
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

I have used the TkAgg backend in python2, installing the dependencies by hand. Is this backend not available for python3?
-Sterling
On Apr 18, 2013, at 8:03PM, John Ladasky wrote:
> Thanks to both Francesco Montesano and Benjamin Root. I have done some 
> reading. And I have made some progress, though I am not quite where I 
> want to be yet.
> 
> So the problem appears to be that the only backend for which I had 
> suitable Python 3 libraries was agg. It only requires libpng, which I 
> have. I can render a Matplotlib canvas, but it appears that the only 
> output that agg offers is in the form of PNG files to disk. I cannot 
> create a live window on the screen.
> 
> Reading more, I realize that the way I was getting GUI output previously 
> (with Python 2.7 and Matplotlib 1.1) was through wxPython. 
> Unfortunately, it appears that wxPython's star is fading, and a Python 
> 3-compatible version will not be written. In fact, wxPython hasn't 
> released a new version in nine months.
> 
> The other choices for Matplotlib GUI output on Linux appear to be 
> through GTK, PySide, and PyQt. I am not familiar with GTK, but I know 
> that it is widely-used. Also, GTK appears to be Python 3-compatible, 
> and so that is where I need to go.
> 
> I'm going through a trial and error process. Unfortunately, the names 
> of the repositories in Ubuntu are not very helpful. I installed a few 
> GTK and python-gtk related packages that I thought were relevant. On my 
> first build attempt I got no errors, but also, I didn't get a GTKAgg 
> backend. Upon re-reading, I saw that I should modify matplotlib's 
> setup.cfg file to force a GTK build attempt, and to report errors if it 
> fails. That's what it does. In the "optional backend dependencies" 
> section I am not seeing any GTK libraries listed, even though I have 
> installed python-gtk2-dev (2.24.0), python-gobject-2-dev (2.28.6), 
> libgtk2.0-dev (2.24.10), libglib2.0-dev (2.32.3), python-gi-dev (3.2.2), 
> python-gobject-dev (3.2.2), python3-gi (3.2.2), and a few DOZEN packages 
> on which these depend.
> 
> If anyone knows the way forward from here, I would appreciate your 
> advice. Thanks again.
> 
> ------------------------------------------------------------------------------
> Precog is a next-generation analytics platform capable of advanced
> analytics on semi-structured data. The platform includes APIs for building
> apps and a phenomenal toolset for data science. Developers can use
> our toolset for easy data analysis & visualization. Get a free account!
> http://www2.precog.com/precogplatform/slashdotnewsletter
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

On Thu, Apr 18, 2013 at 11:03 PM, John Ladasky
<joh...@sb...>wrote:
> .
> Reading more, I realize that the way I was getting GUI output previously
> (with Python 2.7 and Matplotlib 1.1) was through wxPython.
> Unfortunately, it appears that wxPython's star is fading, and a Python
> 3-compatible version will not be written. In fact, wxPython hasn't
> released a new version in nine months.
>
wxPython is alive and well and the newest developmental version of it
("Phoenix") runs on Python 3. It should be released fairly soon. One of
the wxPython list regulars mentioned getting his software to run with it,
with a few minor issues, just six days ago. So you might want to give
Phoenix a try.
Che