matplotlib-users Mailing List for matplotlib

>>>>> "Jeff" == Jeff Peery <jef...@se...> writes:
 Jeff> Hi again, my operating system is windows 2000 professional,
 Jeff> and I'm using python 2.4, wxpython 2.5.5, and
 Jeff> matplotlib-0.80.win32-py2.4.exe. I attached my code. The
 Jeff> problem is that the program crashes when I try to close a
 Jeff> plot or plot a second plot. Any ideas how to make this work?
 Jeff> Thanks.
Don't import pylab while using the OO API -- see
examples/embedding_in_wx*.py and 
 http://matplotlib.sourceforge.net/faq.html#OO
 
Should help...

#!/usr/bin/env python
#Boa:App:BoaApp
import wx
import sample
modules ={'sample': [1, 'Main frame of Application', 'none://sample.py']}
class BoaApp(wx.App):
 def OnInit(self):
 wx.InitAllImageHandlers()
 self.main = sample.create(None)
 # needed when running from Boa under Windows 9X
 self.SetTopWindow(self.main)
 self.main.Show();
 return True
def main():
 application = BoaApp(0)
 application.MainLoop()
if __name__ == '__main__':
 main()

>>>>> "John" == John Gill <jn...@eu...> writes:
 John> Attached is a patch to collections.py and legend.py to make
 John> the auto-legend stuff smarter.
 John> A couple of caveats.
 John> The new code doesn't makes as much use of C++ -- there might
 John> be performance issues for plots with large numbers of
 John> points, although I haven't encountered any.
 John> I don't think any of my tests include plots with
 John> LineCollections, so there is a reasonable chance that code
 John> in _auto_legend_data has not been exercised.
 John> Other than that, it seems to do the trick.
Thanks John, also added to CVS. I think this could be extremely slow
for very large lines (100k points, for example) but I don't see that
as a real problem because auto placement is not the default for
legends (for precisely this reason).
Cheers!
JDH

>>>>> "John" == John Gill <jn...@eu...> writes:
 John> Attached is a patch to the axes code that adds an extra
 John> paramter, linewidths, to the scatter method.
 John> My motivation for this wa I didn't want the black borders
 John> around all the points in my plot.
This is a common wish -- just as a warning, using a linewidth of 0
does not work across backends; for example, I think it fails on PS.
The standard way to do it is to set the edgecolors to be the same as
the facecolors. To support this, I also added a boolean kwarg,
faceted=True and when False, sets the edge colors to be the same as
the facecolor. I also added your linewidth changes.
axes.py revision 1.103 or later ...
Thanks!
JDH

Attached is a patch to collections.py and legend.py to make the 
auto-legend stuff smarter.
A couple of caveats.
The new code doesn't makes as much use of C++ -- there might be 
performance issues for plots with large numbers of points, although I 
haven't encountered any.
I don't think any of my tests include plots with LineCollections, so 
there is a reasonable chance that code in _auto_legend_data has not been 
exercised.
Other than that, it seems to do the trick.
John

Attached is a patch to the axes code that adds an extra paramter, 
linewidths, to the scatter method.
My motivation for this wa I didn't want the black borders around all the 
points in my plot.
This allows you to do:
scatter(y, x, c=z, linewidths=(0.0,))
and get some really pretty plots, similar to using pcolor/imshow, but 
works when your points are not neatly aligned in a grid.
John

>>>>> "Ryan" == Ryan Krauss <rya...@co...> writes:
 Ryan> I have a question about tick formatting. I have a semilogx
 Ryan> plot and if I resize the xticks using locs , labels = xticks
 Ryan> () set(labels , size=ticksize) I like the size and the font,
 Ryan> but the exponents (10^0) run into the axis.
The reasons for are complicated and have to do with an optimization to
make exponential ticking faster, and I won't go into them right now.
Suffice it to say that it is a bug, but there may be a workaround
Do any of the suggestions here help?
 http://matplotlib.sourceforge.net/faq.html#TEXTOVERLAP
Note you can also control the "pad" in points of the offset of the
ticks from the xaxis
 ticks = ax.xaxis.get_major_ticks()
 for tick in ticks:
 tick.set_pad(6)
or if you prefer
 set(ax.xaxis.get_major_ticks(), pad=6)
The default pad is controlled by an rc parameter
 tick.major.pad : 4 # distance to major tick label in points
 tick.minor.pad : 4 # distance to the minor tick label in points
See http://matplotlib.sf.net/.matplotlibrc
JDH

I have a question about tick formatting. I have a semilogx plot and if 
I resize the xticks using
locs , labels = xticks ()
set(labels , size=ticksize)
I like the size and the font, but the exponents (10^0) run into the axis.
If I size them using
xticks([1,10],[r'10ドル^0$',r'10ドル^1$'],size=ticksize,family='sans=serif')
I can get the position right (i.e. not running into the axis), but I 
have to set the ticks manually and I can't seem to get away from a math 
font I don't really like.
Is there a better way to do this?
Thanks,
Ryan

Ryan Krauss wrote:
> Looking at the code, it seems like SystemExit is raised any time the png doesn't 
> already exist.
> I am running these lines inside ipython, if that matters. Something about my 
> system doesn't like the SystemExit being raised.
yes, it's an annoyance of ipython whenever you run anything with a SystemExit 
in it. The .14 release of ipython has a new -e switch to ignore exit calls, 
so you don't get all that noise (use 'run -e fooscript').
If you feel adventurous, I put yesterday a release candidate here:
http://ipython.scipy.org/dist/testing/
Let me know of any problems you encounter though, I'd like this to really fix 
all known glitches for a while so I can concentrate only on new development.
Best,
f

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
 <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type">
 <title></title>
</head>
<body bgcolor="#ffffff" text="#000000">
Thanks again John.&nbsp; The newest texmanager.py worked great.<br>
<br>
John Hunter wrote:
<blockquote cite="mid...@pe..."
 type="cite">
 <blockquote type="cite">
 <blockquote type="cite">
 <blockquote type="cite">
 <blockquote type="cite">
 <blockquote type="cite">
 <pre wrap="">"Ryan" == Ryan Krauss <a class="moz-txt-link-rfc2396E" href="mailto:gt...@ma...">&lt;gt...@ma...&gt;</a> writes:
 </pre>
 </blockquote>
 </blockquote>
 </blockquote>
 </blockquote>
 </blockquote>
 <pre wrap=""><!---->
 Ryan&gt; I would like to use Matplotlib for automatically generating
 Ryan&gt; reports in HTML. I would like to do this without having to
 Ryan&gt; use latex first and the convert from there (it would be much
 Ryan&gt; faster to make my own HTML directly and I don't need lots of
 Ryan&gt; complicated features yet). Has anyone already done this who
 Ryan&gt; is willing to share code with me?
 Ryan&gt; One thing I need to do that would make this work really well
 Ryan&gt; is to generate little PNG's of symbols and formulas to use
 Ryan&gt; in line with text in the HTML (sort of how LaTeX2HTML
 Ryan&gt; handles using $\theta$ in line. Is there a way to use the
 Ryan&gt; TeX rendering system used on figures to make little PNG's
 Ryan&gt; with just TeX expressions on them (i.e. theta.png)?
TeX/LaTeX plus dvipng is really the right way to solve this problem.
Coincidentally, I have been working to incorporate tex into
backend_agg via dvipng (and into backend_ps via psfrag) and matplotlib
has a tex manager class
So if you don't mind installing tex and dvipng (on my Ubuntu system is
is simply 
 &gt; sudo apt-get install dvipng
then you can use the matplotlib texmanager class to handle the system
calls, cacheing results it's seen before and so on
 &gt;&gt;&gt; from matplotlib.texmanager import TexManager
 &gt;&gt;&gt; m = TexManager()
 &gt;&gt;&gt; pngfile = m.make_png("\TeX\ is Number $e^{-i\pi}$!", dpi=100)
 &gt;&gt;&gt; print pngfile
 /home/jdhunter/.tex.cache/5b723d2ea8d0f15af94ec585aece1582_100.png
You need to make sure you have texmanager revision 1.2 from CVS (or
later). 
Of course if you are on a platform where TeX is not easily installed,
this won't help much. In that case, you can use matplotlib to create
the math images for you, but I would use the mathtext parser directly
rather than the whole figure / axes api.
Let me know if you still want/need to use mpl for this and I'll give
you some pointers.
JDH
-------------------------------------------------------
This SF.Net email is sponsored by Oracle Space Sweepstakes
Want to be the first software developer in space?
Enter now for the Oracle Space Sweepstakes!
<a class="moz-txt-link-freetext" href="http://ads.osdn.com/?ad_id=7412&alloc_id=16344&op=click">http://ads.osdn.com/?ad_id=7412&amp;alloc_id=16344&amp;op=click</a>
_______________________________________________
Matplotlib-users mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Mat...@li...">Mat...@li...</a>
<a class="moz-txt-link-freetext" href="https://lists.sourceforge.net/lists/listinfo/matplotlib-users">https://lists.sourceforge.net/lists/listinfo/matplotlib-users</a>
 </pre>
</blockquote>
<br>
</body>
</html>