matplotlib-users Mailing List for matplotlib

In my interactive plots that connect with mouse scroll events
(event.button), I just noticed that TkAgg reports the opposite - 'up'
when the scroll direction is down and vice versa. GTKAgg and WxAgg
report the correct value.
I noticed this after installing ipython in a virtualenv and having
problems with interactive plotting using GTKAgg and WxAgg. Figure
windows would not appear, or would appear and hang the interpreter - a
threading issue I guess. So I switched to TkAgg, which solved the
problem, but then scrolling was mixed up.
This is using stock packages for Ubuntu 10.04, with ipython 0.10.1
installed in the virtualenv with pip. I compiled and installed
matplotlib 1.0 in the same virtualenv and saw the same behavior.
Anyone else had these issues?
-Rick

You might be able to do this (before importing pyplot or pylab):
matplotlib.rcParams['text.usetex'] = True
matplotlib.rcParams['text.latex.preamble'] = '\usepackage{libertine}'
and optionally (if you have xpdf or poppler installed):
matplotlib.rcParams['ps.usedistiller'] = 'xpdf'
Good luck,
Jeff
On Oct 20, 2010, at 3:32 PM, Thøger Emil Juul Thorsen wrote:
> Sorry, perhaps I should have specified; 
> 
> I'm using usetex, and Libertine is not among the supported fonts. In the
> example RC file, it said that if any font was wanted, that could be
> loaded with a \usepackage command, one should request it on the mailing
> list.
> 
> 
> 
> 
> On Wed, 2010年10月20日 at 14:29 -0400, Stan West wrote:
>>> From: Thøger Emil Juul Thorsen [mailto:th...@fy...] 
>>> Sent: Monday, October 18, 2010 14:04
>>> 
>>> I'm using the excellent font Linux Libertine for writing my thesis, in
>>> which I do my plotting in matplotlib. I would really love to 
>>> be ablu to
>>> have consistent fonts in both text and graphs - is this 
>>> possible? I know
>>> it can be loaded by \usepackage{libertine}.
>> 
>> For ordinary text objects, one way is to set the font.family rcParam to the
>> font name. You could do this either in a matplotlibrc file in an appropriate
>> location or in your code dynamically [1]. If you'll be using mathtext, also
>> see the "Fonts" section and "Custom fonts" subsection at [2]; for LaTeX, [3].
>> 
>> [1] http://matplotlib.sourceforge.net/users/customizing.html
>> 
>> [2] http://matplotlib.sourceforge.net/users/mathtext.html#fonts
>> 
>> [3] http://matplotlib.sourceforge.net/users/usetex.html
> 
> 
> 
> ------------------------------------------------------------------------------
> Nokia and AT&T present the 2010 Calling All Innovators-North America contest
> Create new apps & games for the Nokia N8 for consumers in U.S. and Canada
> 10ドル million total in prizes - 4ドルM cash, 500 devices, nearly 6ドルM in marketing
> Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store 
> http://p.sf.net/sfu/nokia-dev2dev
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Sorry, perhaps I should have specified; 
I'm using usetex, and Libertine is not among the supported fonts. In the
example RC file, it said that if any font was wanted, that could be
loaded with a \usepackage command, one should request it on the mailing
list.
On Wed, 2010年10月20日 at 14:29 -0400, Stan West wrote:
> > From: Thøger Emil Juul Thorsen [mailto:th...@fy...] 
> > Sent: Monday, October 18, 2010 14:04
> > 
> > I'm using the excellent font Linux Libertine for writing my thesis, in
> > which I do my plotting in matplotlib. I would really love to 
> > be ablu to
> > have consistent fonts in both text and graphs - is this 
> > possible? I know
> > it can be loaded by \usepackage{libertine}.
> 
> For ordinary text objects, one way is to set the font.family rcParam to the
> font name. You could do this either in a matplotlibrc file in an appropriate
> location or in your code dynamically [1]. If you'll be using mathtext, also
> see the "Fonts" section and "Custom fonts" subsection at [2]; for LaTeX, [3].
> 
> [1] http://matplotlib.sourceforge.net/users/customizing.html
> 
> [2] http://matplotlib.sourceforge.net/users/mathtext.html#fonts
> 
> [3] http://matplotlib.sourceforge.net/users/usetex.html

Hello,
Consider these two simple lines in IPython -pylab:
plt.hist(np.random.randn(1000), normed=1, histtype='step', label='test', lw=2)
plt.legend()
How could I change the appearance of the legend symbol in this case?
It auto-uses a patch object (rectangle in this case).
I would like to get a straight line instead.
Thanks.
-- 
Gökhan

> From: Thøger Emil Juul Thorsen [mailto:th...@fy...] 
> Sent: Monday, October 18, 2010 14:04
> 
> I'm using the excellent font Linux Libertine for writing my thesis, in
> which I do my plotting in matplotlib. I would really love to 
> be ablu to
> have consistent fonts in both text and graphs - is this 
> possible? I know
> it can be loaded by \usepackage{libertine}.
For ordinary text objects, one way is to set the font.family rcParam to the
font name. You could do this either in a matplotlibrc file in an appropriate
location or in your code dynamically [1]. If you'll be using mathtext, also
see the "Fonts" section and "Custom fonts" subsection at [2]; for LaTeX, [3].
[1] http://matplotlib.sourceforge.net/users/customizing.html
[2] http://matplotlib.sourceforge.net/users/mathtext.html#fonts
[3] http://matplotlib.sourceforge.net/users/usetex.html

Friends,
I tried to do a 3d scatter plot. The following is my code. However i dnt
know how to set different colors and marker for x, y and z. Someone kindly
help in doing the same. I tried to give a sequence of colors as
c=['black','red','blue'] but i dnt get what i want.
----------------------
import numpy as np
from mpl_toolkits.mplot3d import Axes3D
import matplotlib.pyplot as plt
fig = plt.figure(figsize=(4,4))
ax = Axes3D(fig)
x=np.reshape(np.loadtxt('PR1'),15000)
y=np.reshape(np.loadtxt('PR2'),15000)
z=np.reshape(np.loadtxt('PR3'),15000)
ax.scatter(x[:],y[:],z[:], marker='o')
plt.savefig('img.png',dpi=200)
plt.show()
Thanks,
Bala

On 10/19/2010 05:44 AM, vaishu wrote:
>
> Hey guys,
>
> I am a new user to the python& matplotlib, this might be a simple question
> but I searched the internet for hours and couldn't find a solution for this.
>
> I am plotting precipitation data from which is in the NetCDF format. What I
> find weird is that, the data doesn't have any negative values in it.(I
> checked that many times,just to make sure). But the value in the colorbar
> starts with a negative value (like -0.0000312 etc). It doesnt make sense
> because I dont do any manipulations to the data, other that just selecting a
> part of the data from the big file and plotting it.
>
> So my code doesn't much to it. Here is the code:
>
> from mpl_toolkits.basemap import Basemap
> import numpy as np
> import matplotlib.pyplot as plt
> from netCDF4 import Dataset
> cd progs
> f=Dataset('V21_GPCP.1979-2009.nc')
> lats=f.variables['lat'][:]
> lons=f.variables['lon'][:]
> prec=f.variables['PREC'][:]
> la=lats[31:52]
> lo=lons[18:83]
> pre=prec[0,31:52,18:83]
> m = Basemap(width=06.e6,height=05.e6,projection='gnom',lat_0=15.,lon_0=80.)
> x, y = m(*np.meshgrid(lo,la))
> m.drawcoastlines()
> m.drawmapboundary(fill_color='lightblue')
> m.drawparallels(np.arange(-90.,120.,5.),labels=[1,0,0,0])
> m.drawmeridians(np.arange(0.,420.,5.),labels=[0,0,0,1])
> cs=m.contourf(x,y,pre,50,cmap=plt.cm.jet)
> plt.colorbar()
>
> The output that I got for that code was a beautiful plot, with the colorbar
> starting from the value -0.00001893, and the rest are positive values, and I
> believe are correct. Its just the minimum value thats bugging me.
> http://old.nabble.com/file/p30001558/jan1979plot.png
>
> I would like to know:
>
> Is there anything wrong in my code? cos I know that the data is right.
I can't reproduce the problem without the data file; can you supply it?
Also, what version of matplotlib are you using?
> Is there a way to manually change the value to 0?
> Is it right for the values in the colorbar to change everytime we run the
> code, cos for the same data, the next time I run the code, the values go
> like this " -0.00001893, 2.00000000, 4.00000000, 6.00000000 etc"
> I want to customize them to "0.0, 2.0, 4.0, 6.0 etc"
In your call to contourf, instead of giving a number of levels, give a 
sequence of the desired levels, e.g.:
cs=m.contourf(x,y,pre,np.arange(0, 20, 0.5),cmap=plt.cm.jet)
Eric
>
>
> Thanks,
> Vaishu
>

On Tue, Oct 19, 2010 at 1:27 PM, Friedrich Romstedt <
fri...@gm...> wrote:
>
> My idea would be to add another layer. Meaning to provide a class
> which *creates* Axes, and which has other instances of *this class* in
> an attribute. The attrib would be scalar. Repetition is done via
> recursive call so some privat e method, initiated by a call to a
> public commit method. So you would do in pseudo-code:
>
> level0 = LevelAxesGenerator() # NOT actually creating the Axes, see below.
> level1 = LevelAxesGenerator(leaf=level0, repeat=10, direction='x') #
> NOT creating too.
>
> level1.create(width=foo, height=bar) # Subdividing automatically,
> generating Axes in the recursive calls on level0-level
>
> This shouldn't be super-hard.
>
> Sorry, I cannot help you with implementation currently.
>
> When you code it, think about general usability, also for other users,
> if you're inclined. I would maybe have been such a user in the past.
>
> ``.create()`` could maybe called ``.commit()``, this is more specific.
> Create can mean anything.
>
> You could even add a .populate() method, so that you can derive from
> ``LevelAxesGenerator`` and the instances automatically populate their
> axes based on the "coordinate" tuple (level0, level1, level2, ...).
> This would be really neat, and it's not too specific, because it'll be
> the main use of this level-functionality.
Hi Friedrich. Thanks for your comments and suggestions.
What I need to clarify is the following. First consider some simple plot A:
it has axes, data points, tick marks, labels, etc., and for all these
elements there are associated absolute x-y coordinates on the canvas. If
now we make this plot A one of the subplots in a collection of, say, 12
subplots, arranged as 3 rows of 4 subplots each, all the x-y coordinates
associated with the original plot A will have to be translated and scaled,
so that the subplot lands in the right place on the canvas, and has the
appropriate size. This process of translation and scaling is what I want to
pinpoint: What exactly is the connection between running the add_subplot
method and the translation+scaling that it entails?
The reason I need to do this is that for the hierarchical containment scheme
I described in my original post, this translation+scaling operation will
have to be done at every level. It is not clear to me how the scheme you
describe would take care of all these repeated translation/scaling
transformations.
I could use some guidance to the source code. What would be a good entry
point for me to answer the questions above by reading the source code?
(Maybe these questions would be more appropriate for the developers' list.)
~kj

2010年10月20日 James Battat <jb...@mi...>:
> Hi,
> I'm having trouble with my built-from-source installation of matplotlib on
> Snow Leopard (I also tried installing from the .dmg file, but that install
> failed -- see below for details).
*SIGH* *SIGH*
YEAH, this keeps bugging us.
Please do a search in Gmane (http:/gmane.org, search for
matplotlib-users ...) for Mac OS X 10.6 or Snow Leopard and from
source or compilation or whatever you feel like.
You'll find a bunch of messages, many of them quite long, but be sure:
It's solvable!
I'll write it up somewhen and put it to the webpage if the developers
allow me for. So many ASAP thingys currently ...
Anyway, don't be frustrated by the many details, and start with the
*recent* messages, and believe them more than the old ones. In the
end, when you got a complete and non-conflicting image, compile again.
 You'll also need to compile the libpng and libfreetype2 I believe.
Try to avoid or use as last rescue make.osx, it'll install old libs in
your /usr/local namespace!
Please apologise me, I'm getting tired of explaining it again and
again, sorry again, good success with your studies,
Friedrich
Don't hesitate to ask questions but I cannot explain it from the very
beginning, it'll also be wasting of storage :-)
Second: The dmg installer requires python.org Python, not Apple one.
The Apple one is said to be "whacky", but I don't know any details.
Just use python.org MacPython, where mpl dmg installer is compiled
against and should work from the very beginning. Compile if you need
svn bleeding-edge.
Third: The installer *might* still have the bug of wrong rights for
some images used in the toolbar, so if you find those, don't be
surprised. It's some missing read right in POSIX. But it might be
fixed now, though then I missed it.
> I downloaded:
>  matplotlib-1.0.0.tar.gz
> and then built and installed:
>> tar xvzf matplotlib-1.0.0.tar.gz
>> cd matplotlib-1.0.0
>> python setup.py build
>> python setup.py install
> Then tried using pylab:
>> cd
>> python
> Python 2.6.1 (r261:67515, Feb 11 2010, 00:51:29)
> [GCC 4.2.1 (Apple Inc. build 5646)] on darwin
> Type "help", "copyright", "credits" or "license" for more information.
>>>> import pylab
> Traceback (most recent call last):
>  File "<stdin>", line 1, in <module>
>  File "/Library/Python/2.6/site-packages/pylab.py", line 1, in <module>
>   from matplotlib.pylab import *
>  File "/Library/Python/2.6/site-packages/matplotlib/pylab.py", line 216, in
> <module>
>   from matplotlib import mpl # pulls in most modules
>  File "/Library/Python/2.6/site-packages/matplotlib/mpl.py", line 2, in
> <module>
>   from matplotlib import axis
>  File "/Library/Python/2.6/site-packages/matplotlib/axis.py", line 10, in
> <module>
>   import matplotlib.font_manager as font_manager
>  File "/Library/Python/2.6/site-packages/matplotlib/font_manager.py", line
> 52, in <module>
>   from matplotlib import ft2font
> ImportError: dlopen(/Library/Python/2.6/site-packages/matplotlib/ft2font.so,
> 2): Symbol not found: _FT_Attach_File
>  Referenced from: /Library/Python/2.6/site-packages/matplotlib/ft2font.so
>  Expected in: flat namespace
> in /Library/Python/2.6/site-packages/matplotlib/ft2font.so
> I've seen both of these problems listed elsewhere, but no solution posted:
>  http://www.mailinglistarchive.com/html/mat...@li.../2010-09/msg00091.html
> Any help is greatly appreciated.
> Here's information on my environment:
>
> Mac OS X 10.6.4
>> uname -a
> Darwin My-MacBook-Pro.local 10.4.0 Darwin Kernel Version 10.4.0: Fri Apr 23
> 18:28:53 PDT 2010; root:xnu-1504年7月4日~1/RELEASE_I386 i386
> I was building from source because I was not able to install from the DMG
> installer (matplotlib-1.0.0-python.org-py2.6-macosx10.4.dmg). When I tried
> the .dmg installer, I got the following message:
>  matplotlib 1.0.0-r0 can't be installed on this disk.
>  matplotlib requires System Python 2.6 to install
> Thanks for your help,
> James

2010年10月20日 Stefan Mauerberger <ste...@mn...>:
> I want to make a polar plot with grid not of the full circle but a
> section (e.g. r=[5:6], phi=[-20:30]). The result should look like this:
> http://homepages.physik.uni-muenchen.de/~Stefan.Mauerberger/example.png
> I have tried a lot and had also looked to the examples but my results
> are not satisfying. This example
> http://matplotlib.sourceforge.net/examples/axes_grid/demo_curvelinear_grid.html seems to be highly relevant for my needs but i do not understand it at all.
>
> I tried two different ways:
> Using the option polar=True works fine but I was not able to shrink the
> plot to the section.
> Otherwise transforming the data to cartesian coordinates (e.g.
> pcolor(r*np.sin(phi),r*np.cos(phi),data) ) the data are plotted into a
> box. But in this case I do not know how to draw a grid.
I believe only option (2) will be feasible, I remember a discussion in
past about this to having had the same outcome.
Try here: http://matplotlib.sourceforge.net/api/axes_api.html#matplotlib.axes.Axes.pcolor
(*faceted* option).
Notice the next functions on that page, they might help you saving
time, but some of them *might* not work in polar Axes, I'm not sure
which of them if any at all, just be warned.
Make sure you're using 1:1 aspect of the axes containing the pcolor()
Collection, or it will look distorted.
Friedrich
P.S.: If this is your first mpl-users thread, make sure to hit the
reply-to-all button ...

Hi,
I'm having trouble with my built-from-source installation of matplotlib on Snow Leopard (I also tried installing from the .dmg file, but that install failed -- see below for details).
I downloaded:
 matplotlib-1.0.0.tar.gz
and then built and installed:
> tar xvzf matplotlib-1.0.0.tar.gz
> cd matplotlib-1.0.0
> python setup.py build
> python setup.py install
Then tried using pylab:
> cd
> python
Python 2.6.1 (r261:67515, Feb 11 2010, 00:51:29) 
[GCC 4.2.1 (Apple Inc. build 5646)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import pylab
Traceback (most recent call last):
 File "<stdin>", line 1, in <module>
 File "/Library/Python/2.6/site-packages/pylab.py", line 1, in <module>
 from matplotlib.pylab import *
 File "/Library/Python/2.6/site-packages/matplotlib/pylab.py", line 216, in <module>
 from matplotlib import mpl # pulls in most modules
 File "/Library/Python/2.6/site-packages/matplotlib/mpl.py", line 2, in <module>
 from matplotlib import axis
 File "/Library/Python/2.6/site-packages/matplotlib/axis.py", line 10, in <module>
 import matplotlib.font_manager as font_manager
 File "/Library/Python/2.6/site-packages/matplotlib/font_manager.py", line 52, in <module>
 from matplotlib import ft2font
ImportError: dlopen(/Library/Python/2.6/site-packages/matplotlib/ft2font.so, 2): Symbol not found: _FT_Attach_File
 Referenced from: /Library/Python/2.6/site-packages/matplotlib/ft2font.so
 Expected in: flat namespace
 in /Library/Python/2.6/site-packages/matplotlib/ft2font.so
I've seen both of these problems listed elsewhere, but no solution posted:
 http://www.mailinglistarchive.com/html/mat...@li.../2010-09/msg00091.html
Any help is greatly appreciated.
Here's information on my environment: 
 
Mac OS X 10.6.4
> uname -a
Darwin My-MacBook-Pro.local 10.4.0 Darwin Kernel Version 10.4.0: Fri Apr 23 18:28:53 PDT 2010; root:xnu-1504年7月4日~1/RELEASE_I386 i386
I was building from source because I was not able to install from the DMG installer (matplotlib-1.0.0-python.org-py2.6-macosx10.4.dmg). When I tried the .dmg installer, I got the following message:
 matplotlib 1.0.0-r0 can't be installed on this disk.
 matplotlib requires System Python 2.6 to install
Thanks for your help,
James
 

Hello everyone, 
I want to make a polar plot with grid not of the full circle but a
section (e.g. r=[5:6], phi=[-20:30]). The result should look like this:
http://homepages.physik.uni-muenchen.de/~Stefan.Mauerberger/example.png
I have tried a lot and had also looked to the examples but my results
are not satisfying. This example
http://matplotlib.sourceforge.net/examples/axes_grid/demo_curvelinear_grid.html seems to be highly relevant for my needs but i do not understand it at all.
I tried two different ways: 
Using the option polar=True works fine but I was not able to shrink the
plot to the section. 
Otherwise transforming the data to cartesian coordinates (e.g.
pcolor(r*np.sin(phi),r*np.cos(phi),data) ) the data are plotted into a
box. But in this case I do not know how to draw a grid. 
Could anyone give me some advice how to do this?
Regards
Stefan

Ben, thanks a lot! Your way does exactly what I want. 
Scott, I do not understand your solution, unfortunately. I have already
known the example from the gallery. But up to now, I have not dealt with
masked arrays. Anyway ... 
Thanks a lot again. 
Stefan