matplotlib-users Mailing List for matplotlib

Eric Firing wrote:
> 
> Some time in the last year I made changes in mpl and in numpy to speed 
> this up (the colormapping). What versions of mpl and numpy are you using?
> 
> Eric
Sorry, I was still using matplotlib 0.91.4 (numpy 1.1.0). I upgraded to 
0.98.3 and now the Colormap.__call__ is 5 times faster and no longer a 
problem. motion_notify_event of backend_gtk.py still takes much more 
time than it should, but I haven't looked into that, and anyway, my 
application runs fast enough now.
Thanks,
João Luís Silva

João Luís Silva wrote:
> Hello all,
> 
> I'm creating a python application what does some calculation and 
> displays the result as two imshow() with respective colorbars as the 
> calculation progresses. Each array is ~8192x50. Updating the image is 
> slow, so I only update every 2 seconds or so, but even doing this a 
> significant part of the program's time is spent on Colormap.__call__ of 
> colors.py. Another significant time sink is motion_notify_event of 
> backend_gtk.py. A small example follows to illustrate the problem. Any 
> ideia on how to circunvent or correct this?
Some time in the last year I made changes in mpl and in numpy to speed 
this up (the colormapping). What versions of mpl and numpy are you using?
Eric
> 
> Thank you,
> João Luís Silva
> 
> ----------------------------------------
> 
> import numpy as np
> from numpy.random import rand
> import pygtk
> pygtk.require('2.0')
> import gtk
> import matplotlib
> matplotlib.use('GTKAgg')
> from matplotlib.backends.backend_gtkagg import FigureCanvasGTKAgg as 
> FigureCanvas
> from matplotlib.figure import Figure
> 
> def main():
> N = 2000
> 
> win = gtk.Window()
> vbox = gtk.VBox()
> fig = Figure()
> ax = fig.add_subplot(111)
> canvas = FigureCanvas(fig)
> im = 
> ax.imshow(rand(N,N),origin='lower',aspect='auto',extent=(0.0,1.0,0.0,1.0),interpolation='bilinear')
> cb = fig.colorbar(im)
> vbox.pack_start(canvas)
> win.add(vbox)
> win.connect("destroy", lambda x: gtk.main_quit())
> win.show_all()
> gtk.main()
> 
> if __name__ == '__main__':
> #Use hotshot2calltree and kcachegrind to see the stats
> import hotshot
> prof = hotshot.Profile("hotshot_stats")
> prof.runcall(main)
> prof.close()
> 
> 
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

On Oct 17, 2008, at 11:06 AM, Michael Droettboom wrote:
> No, it's not possible now. Just so I'm clear -- you're suggesting 
> plotting on a wedge from theta_1 to theta_2? Seems doable with a 
> little work.
Yes, exactly. For instance if I wanted to make a polar bar chart 
showing the distribution of angles separating a collection of lines at 
their intersections, I would only need to display one quarter of the 
polar plot, the first quadrant, from 0-90 degrees. If I wanted to 
display the orientation of, say, a fault on the surface of the Earth, 
I'd need 180 degrees, but I might want to use the northern half of the 
circle (orientations ranging from West to East (-90, 90)) or perhaps 
the eastern half (ranging from (0, 180)).
Ideally, the user would be able to specify both thetamin and thetamax, 
and also an offset describing the location of theta=0, because in some 
applications it doesn't make sense for zero to be in its traditional 
polar-coordinates location at the right hand side. E.g. in a 
geographic context, where the angles being displayed are compass 
headings, it is traditional to put zero (north) at the top. Putting 
zero at the top would also make sense if you were plotting data that 
corresponded to angular deviations to the left and right from some 
specified path.
Some portion of this functionality has been implemented in the 
"windrose" module, see:
http://youarealegend.blogspot.com/2008/09/windrose.html
http://www.nabble.com/windrose-OO-td16298025.html
but with a mind toward a particular kind of application (windrose 
diagrams obviously :). But I think allowing users to specify 
(thetamin, thetamax, offset) would be a very generally applicable 
improvement.
Thanks again,
Zane
--
Zane Selvans
Amateur Earthling
za...@id...
303/815-6866
http://zaneselvans.org
PGP Key: 55E0815F

I can fill between curves using either:
xs, ys = poly_between(x, y1, y2)
ax = gca()
ax.fill(xs,ys, facecolor='b', alpha=0.5)
or 
xs = concatenate((x,x[::-1]))
ys = concatenate((y1,y2[::-1]))
p = fill(xs,ys,facecolor='b', alpha=0.5)
giving
http://www.nabble.com/file/p20038095/good.png 
But, if I try to do either on a semilogy plot, I get a strange looking
combination of reflected-looking polygons.
http://www.nabble.com/file/p20038095/bad.png 
Any ideas?
-- 
View this message in context: http://www.nabble.com/fill-between-curves-on-log-scale--tp20038095p20038095.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

No, it's not possible now. Just so I'm clear -- you're suggesting 
plotting on a wedge from theta_1 to theta_2? Seems doable with a little 
work.
Mike
Zane Selvans wrote:
> Is there any way to limit the theta extent in a polar axis plot?
>
> There are a variety of angular contexts in which 360 degrees are not 
> necessary. To convey the orientation of an undirected line, you need 
> only 0-180 degrees. The minimum separation between two intersecting 
> lines only ranges from 0-90 degrees.
>
> It seems like the ideal way for this to work would be to have a 
> function analogous to xlim(), maybe called thetalim(thetamin, 
> thetamax), and then treat the two bounding radial lines as the primary 
> and secondary y axes.
>
> Didn't see anything like this in any of the examples or documentation 
> though. If it's not possible now, I think it would be a wonderful 
> addition.
>
> Thanks,
> Zane
>
> --
> Zane Selvans
> Amateur Earthling
> za...@id... <mailto:za...@id...>
> 303/815-6866
> http://zaneselvans.org
> PGP Key: 55E0815F
>
>
>
>
>
>
> ------------------------------------------------------------------------
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> ------------------------------------------------------------------------
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

Is there any way to limit the theta extent in a polar axis plot?
There are a variety of angular contexts in which 360 degrees are not 
necessary. To convey the orientation of an undirected line, you need 
only 0-180 degrees. The minimum separation between two intersecting 
lines only ranges from 0-90 degrees.
It seems like the ideal way for this to work would be to have a 
function analogous to xlim(), maybe called thetalim(thetamin, 
thetamax), and then treat the two bounding radial lines as the primary 
and secondary y axes.
Didn't see anything like this in any of the examples or documentation 
though. If it's not possible now, I think it would be a wonderful 
addition.
Thanks,
Zane
--
Zane Selvans
Amateur Earthling
za...@id...
303/815-6866
http://zaneselvans.org
PGP Key: 55E0815F

>> First, are there plans to eventually embed autogenerated plots in 
>> the the "examples" section (http://matplotlib.sourceforge.net/examples/ 
>> ) <http://matplotlib.sourceforge.net/examples/%29> in the same way 
>> that they're embedded within the user guide?
> There are plans to do that, and probably to have a "thumbnail 
> gallery" of examples as well. The sticking point is that many of 
> the examples generate multiple figures, and we currently don't have 
> a way to save those out automatically from the documentation. Just 
> need to make that work.
A thumbnail gallery of all the examples in the User's Guide and 
Examples is a great idea!
Plots are inherently visual, and humans have a huge amount of built-in 
visual processing power. Searching a even an enormous page of images 
to find something that looks like what you're trying to create will be 
much more efficient that trying to come up with the right keyword if 
you don't know what it is already, or scanning a table of contents.
Actually, it reminds me of the "graphical abstracts" that many 
chemistry journals use - one figure, often showing the molecule(s) 
that the paper is about, which readers can browse online in lieu of 
searching or reading long ISO approved chemical names... which often 
allows almost instant recognition of papers you'll be interested in, 
or at least a serious pruning of the search space.
>> Second, in the sidebar on the main page, in the "Need Help?" 
>> section, I'm seeing a bunch of weird gunk that's probably not 
>> supposed to be getting through to the web...:
>
> That seems to have been fixed in the meantime. Still seeing it?
Yep, looks good.
Thanks for all your work!
--
Zane Selvans
Amateur Earthling
za...@id...
303/815-6866
http://zaneselvans.org
PGP Key: 55E0815F

> "John Hunter" <jd...@gm...> 10/16/08 6:39 PM
> We've been working behind the scenes on a new documentation system for
> matplotlib, which integrates the web site, API documentation and PDF
> As an example we can include plots in our API documentation, see
> http://matplotlib.sourceforge.net/api/axes_api.html#matplotlib.axes.Axes.acorr
Awesome!!
Of course you chose to advertise the docstring with a typo ;-)
In the acorr docstring the sentence "If normed = True, normalize the data *but* the autocorrelation at 0-th lag.", should read "If normed = True, normalize the data *by* the autocorrelation at 0-th lag.".
Yours in sheer impressedness,
Scott
Please find our Email Disclaimer here: http://www.ukzn.ac.za/disclaimer/

Hello all,
I'm creating a python application what does some calculation and 
displays the result as two imshow() with respective colorbars as the 
calculation progresses. Each array is ~8192x50. Updating the image is 
slow, so I only update every 2 seconds or so, but even doing this a 
significant part of the program's time is spent on Colormap.__call__ of 
colors.py. Another significant time sink is motion_notify_event of 
backend_gtk.py. A small example follows to illustrate the problem. Any 
ideia on how to circunvent or correct this?
Thank you,
João Luís Silva
----------------------------------------
import numpy as np
from numpy.random import rand
import pygtk
pygtk.require('2.0')
import gtk
import matplotlib
matplotlib.use('GTKAgg')
from matplotlib.backends.backend_gtkagg import FigureCanvasGTKAgg as 
FigureCanvas
from matplotlib.figure import Figure
def main():
 N = 2000
 win = gtk.Window()
 vbox = gtk.VBox()
 fig = Figure()
 ax = fig.add_subplot(111)
 canvas = FigureCanvas(fig)
 im = 
ax.imshow(rand(N,N),origin='lower',aspect='auto',extent=(0.0,1.0,0.0,1.0),interpolation='bilinear')
 cb = fig.colorbar(im)
 vbox.pack_start(canvas)
 win.add(vbox)
 win.connect("destroy", lambda x: gtk.main_quit())
 win.show_all()
 gtk.main()
if __name__ == '__main__':
 #Use hotshot2calltree and kcachegrind to see the stats
 import hotshot
 prof = hotshot.Profile("hotshot_stats")
 prof.runcall(main)
 prof.close()

I'm not a regular Windows user, so I'm probably not of much help, but...
It's trying to write a default .matplotlibrc file to C:\ Is that 
directory writable, by the user that is running matplotlib?
Mike
David Springate wrote:
>
> Hi,
>
> I am a fairly new Python user and am trying to set up matplot lib to 
> work with scipy as an alternative to matlab. I installed matplotlib 
> on my home pc with no problems but I followed the same protocol on my 
> work pc and got the following error when I tried to call in
>
> "From pylab import *" :
>
> Traceback (most recent call last):
> File "<pyshell#0>", line 1, in <module>
> from pylab import *
> File "C:\Python25\Lib\site-packages\pylab.py", line 1, in <module>
> from matplotlib.pylab import *
> File "C:\Python25\lib\site-packages\matplotlib\__init__.py", line 677, in
> <module>
> rcParams = rc_params()
> File "C:\Python25\lib\site-packages\matplotlib\__init__.py", line 598, in
> rc_params
> fname = matplotlib_fname()
> File "C:\Python25\lib\site-packages\matplotlib\__init__.py", line 548, in
> matplotlib_fname
> fname = os.path.join(get_configdir(), 'matplotlibrc')
> File "C:\Python25\lib\site-packages\matplotlib\__init__.py", line 242, in
> wrapper
> ret = func(*args, **kwargs)
> File "C:\Python25\lib\site-packages\matplotlib\__init__.py", line 438, in
> _get_configdir
> raise RuntimeError("Failed to create %s/.matplotlib; consider setting
> MPLCONFIGDIR to a writable directory for matplotlib configuration data"%h)
> RuntimeError: Failed to create C:\/.matplotlib; consider setting
> MPLCONFIGDIR to a writable directory for matplotlib configuration data
>
> ...I checked and as far as I can tell, none of the python directories 
> are read-only. I also
> followed the instructions on the thread 'runtime error' which dealt with I
> think the same problems, but after patching the cutils.py (line 79) 
> and the
> __init__ (line 435) files, I still have the same problem.
>
>
> Any advice would be apreciated. My home pc (where it works fine) is 
> Vista, while work is XP pro. Python
> 2.5.2, scipy, numpy and matplotlib all installed from binaries. I 
> also tried the enthought binary installation with the same result.
>
> Thanks,
>
> Dave
>
> ------------------------------------------------------------------------
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> ------------------------------------------------------------------------
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

The title says it all. I need the simplest blit animation I can find and 
when it works animation_blit.py works quite well, but sometimes it comes 
up and runs to completion at almost 400 FPS, other times it comes up and 
nothing is drawn at all except an empty plot.
Ubuntu Hardy
matplotlib 0.98.3
python 2.5
python-wxGtk 2.8
Any help would be appreciated.
...Thanks,
...Ken

This has now been fixed in SVN.
index.html is the only page that includes hand-written HTML. If you see 
any errors of this nature on other pages, please file bugs with Sphinx 
and/or docutils.
Cheers,
Mike
Xavier Gnata wrote:
> Looks great but there are too many errors:
> http://validator.w3.org/check?uri=http%3A%2F%2Fmatplotlib.sourceforge.net%2F&charset=(detect+automatically)&doctype=Inline&group=0
>
> I'm not a geek and I do not care about w3c small warnings but it would 
> be so nice to have a xhtml compliant website (as close as possible)
>
> From an "artistic" point of view, I would put more emphasis on the 
> screenshot (pylab purpose is to produce *very* nice images...)
>
> xavier
>
>
> 
>> We've been working behind the scenes on a new documentation system for
>> matplotlib, which integrates the web site, API documentation and PDF
>> guide into a single source of sphinx/rest documents which are easier
>> to maintain and extend, hopefully leading to better and more
>> up-to-date docs.
>>
>> We went live with the new site yesterday:
>>
>> http://matplotlib.sf.net
>>
>> so check it out and let us know if something is broken or missing. We
>> don't have everything that was on the old site (some stuff from the
>> FAQ, "what's new" and "user's guide" has not been ported over) but we
>> do have should be current, searchable, indexed and cross-linked.
>>
>> Thanks to Darren Dale who spear-headed the effort to use the sphinx
>> documentation, and to the developers who have contributed, especially
>> Michael Droettboom, who has developed several nice sphinx extensions
>> to do inheritance diagrams, syntax highlighting of ipython sessions,
>> and inline plotting. As an example we can include plots in our API
>> documentation, see
>>
>> http://matplotlib.sourceforge.net/api/axes_api.html#matplotlib.axes.Axes.acorr
>>
>> We embed these plots with a "plot" directive that generates the
>> figures from external code at documentation build time, which
>> guarantees that the example code you see in the docs generate the
>> figures you see in the docs. For example, in the acorr docstring, all
>> we have to do is::
>>
>> **Example:**
>>
>> .. plot:: ../mpl_examples/pylab_examples/xcorr_demo.py
>>
>> and the figure and source code links automagically appear in the docs.
>>
>> Because some of these extensions are generally useful, Michael,
>> Fernando and I have been working on a "sphinx_template" which contains
>> the template of a sphinx documentation project with these extensions
>> in place, so people who want to get started using sphinx (the official
>> documentation system for python, numpy, ipython and matplotlib) can do
>> so more easily. Right now it is available in svn
>>
>> > svn co https://matplotlib.svn.sourceforge.net/svnroot/matplotlib/trunk/py4science/examples/sphinx_template2
>>
>> and see the README in the checkout directory. Michael also did a talk
>> on matplotlib's use of sphinx and the sphinx template at the last
>> scipy conference. We're still waiting for the videos of the talks to
>> be posted (can someone poke someone?) but you can see the talk PDF
>> from the proceedings here:
>>
>> http://conference.scipy.org/proceedings/SciPy2008/paper_6/
>>
>> JDH
>>
>> -------------------------------------------------------------------------
>> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
>> Build the coolest Linux based applications with Moblin SDK & win great prizes
>> Grand prize is a trip for two to an Open Source event anywhere in the world
>> http://moblin-contest.org/redirect.php?banner_id=100&url=/
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>> 
>> 
>
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

Zane Selvans wrote:
> Wow! The new site looks beautiful. Thank you!
>
> The apparent lack of searchable, online, well organized, 
> cross-referenced documentation has been my main frustration with 
> Matplotlib. I'm very glad it's being worked on.
>
> A couple of quick things:
>
> First, are there plans to eventually embed autogenerated plots in the 
> the "examples" section (http://matplotlib.sourceforge.net/examples/) 
> <http://matplotlib.sourceforge.net/examples/%29> in the same way that 
> they're embedded within the user guide?
There are plans to do that, and probably to have a "thumbnail gallery" 
of examples as well. The sticking point is that many of the examples 
generate multiple figures, and we currently don't have a way to save 
those out automatically from the documentation. Just need to make that 
work.
>
> Second, in the sidebar on the main page, in the "Need Help?" section, 
> I'm seeing a bunch of weird gunk that's probably not supposed to be 
> getting through to the web...:
>
> <<<<<<< .mine <<<<<<< .mine
>
> Check the docs, the faq, and join the matplotlib mailing lists. There 
> is a also community wiki cookbook
> =======
>
> Check the user guide, the
>
> Check the user guide, the >>>>>> .r6222 href="faq/index.html">faq, the 
> api docs, archives, and join the matplotlib mailing lists >>>>>>> .r6221
>
> You can file bugs, patches and feature requests on the sourceforge 
> tracker, but it is a good idea to ping us on the mailing list too.
That seems to have been fixed in the meantime. Still seeing it?
Cheers,
Mike
-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

Hi,
I am a fairly new Python user and am trying to set up matplot lib to work
with scipy as an alternative to matlab. I installed matplotlib on my home
pc with no problems but I followed the same protocol on my work pc and got
the following error when I tried to call in
"From pylab import *" :
 Traceback (most recent call last):
 File "<pyshell#0>", line 1, in <module>
 from pylab import *
 File "C:\Python25\Lib\site-packages\pylab.py", line 1, in <module>
 from matplotlib.pylab import *
 File "C:\Python25\lib\site-packages\matplotlib\__init__.py", line 677, in
<module>
 rcParams = rc_params()
 File "C:\Python25\lib\site-packages\matplotlib\__init__.py", line 598, in
rc_params
 fname = matplotlib_fname()
 File "C:\Python25\lib\site-packages\matplotlib\__init__.py", line 548, in
matplotlib_fname
 fname = os.path.join(get_configdir(), 'matplotlibrc')
 File "C:\Python25\lib\site-packages\matplotlib\__init__.py", line 242, in
wrapper
 ret = func(*args, **kwargs)
 File "C:\Python25\lib\site-packages\matplotlib\__init__.py", line 438, in
_get_configdir
 raise RuntimeError("Failed to create %s/.matplotlib; consider setting
MPLCONFIGDIR to a writable directory for matplotlib configuration data"%h)
RuntimeError: Failed to create C:\/.matplotlib; consider setting
MPLCONFIGDIR to a writable directory for matplotlib configuration data
...I checked and as far as I can tell, none of the python directories are
read-only. I also
followed the instructions on the thread 'runtime error' which dealt with I
think the same problems, but after patching the cutils.py (line 79) and the
__init__ (line 435) files, I still have the same problem.
Any advice would be apreciated. My home pc (where it works fine) is Vista,
while work is XP pro. Python
2.5.2, scipy, numpy and matplotlib all installed from binaries. I also
tried the enthought binary installation with the same result.
Thanks,
Dave