matplotlib-users Mailing List for matplotlib (Page 10)

I have sets of data to plot on semilogy. I want the minimum y axis set to
some value, say 10e-10.
I do:
 axis([0,1,1e-10,1])
 hold(True)
 for (whatever):
 semilogy (x, y)
 grid()
 legend()
 show()
But the data is not clipped in y from [1e-10..1] as I wanted. What's wrong
here?

On Thu, May 8, 2008 at 10:10 AM, dertom <tw...@ar...> wrote:
> Please, can anybody tell me, if there is a solution, to call the show()
> method more than one time,while my main script is running?
To do this right, you will probably need to write a tk application and
embed matplotlib in it, rather than try to use pylab. See the
embedding_in_tk*.py examples at http://matplotlib.sf.net/examples/.
There is also a tutorial on working with the matplotlib artist API at
http://matplotlib.sf.net/pycon
 http://matplotlib.sourceforge.net/pycon/artist_api_tut.pdf
There are a couple of other FAQs that are relevant
 http://matplotlib.sf.net/faq.html#OO
 http://matplotlib.sf.net/faq.html#SHOW
JDH

On Thu, May 8, 2008 at 11:39 PM, Tim Mitchell <tim...@gm...> wrote:
> I can use the axes position to work I guess but that seems rather
> unsatisfactory as I'll have to figure out when it's being displayed or
> not, font sizing and spacing blah blah blah.
> Is there an easier way around this (or a fix?)? I'm using v0.90.0
There is no current way to automate this easily -- Michael has done
some experimental work for autosizing the axes to make room for the
text on the svn trunk, but it is not complete. There is a parameter
that you can set in your matplotlibrc figure.subplot.bottom to make
the bottom of the axes higher by default, or you can manually change
this with
fig.subplots_adjust(bottom=0.2)
JDH

Søren Nielsen wrote:
>
> Søren Nielsen wrote:
>
> Also, if I want to use matplotlib.patches on a canvas where I
> can actually see individual pixels, is it possible to only
> draw on the available pixels showing, and not just make a nice
> high resolution line that is plotted on top of my image?
>
> I'm not sure I understand the question. You mean to draw a line
> using large pixels, rather than as a vector? I don't think
> there's an easy way on that one -- but maybe someone else here can
> think of a creative way to do it.
>
>
> Yes you understod it correctly, I guess I could figure out what pixels 
> are involved in a line or a circle for instance, and then do a scatter 
> plot on those pixels to form the shape.. but it sounds really 
> computationally heavy.. and I need something that works fast, (the 
> user should be able to mask a circle in an image using the pixels 
> available in the image) Also the circle should be shown as it is 
> created (like choosing a center and moving the mouse enlarges the 
> circle).
>
> Maybe someone tried this? masking stuff in an image seems like a 
> fairly common thing :)
Of the top of my head and hand waving over all of the gory details -- 
you could try drawing the line to one plot, saving the image (to a 
CStringIO in-memory "file"), and then loading that into another plot 
using imshow(). (Easier said than done, but should be theoretically 
possible, and not extremely slow).
Mike
>
> Thanks for the help!
> Soren
>
>
> 
>
>
> Cheers,
> Mike
>
>
> Thanks,
> Soren
>
>
> ------------------------------------------------------------------------
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by the 2008 JavaOne(SM)
> Conference Don't miss this year's exciting event. There's
> still time to save 100ドル. Use priority code J8TL2D2.
> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> ------------------------------------------------------------------------
>
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> <mailto: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
>
>
-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

Jouni K. Seppänen wrote:
> Christopher Brown <c-...@as...> writes:
>
> 
>> I have read a little bit online about the difference between types 1
>> and 3, but I can't find a description of which font is which type. So
>> what font should I use?
>> 
>
> Try setting the pdf.fonttype parameter in your matplotlibrc file to 42
> and see if the resulting file passes validation. Matplotlib currently
> only uses TrueType fonts, which can be embedded in either Type 3 or Type
> 42 format. I seem to recall there was some problem with the Type 42
> format so Type 3 was made the default; Michael Droettboom probably
> remembers it better.
Type 42 fonts are currently working, as far as I know. There was a 
transient problem with them but I think that's fixed. Type 3 is the 
default just because it makes the files smaller.
Cheers,
Mike
-- 
Michael Droettboom
Science Software Branch
Operations and Engineering Division
Space Telescope Science Institute
Operated by AURA for NASA

Hi,
First I want to thanks for this amazing Matplotlib. It's fantastic!
My newbie question is how I can show the (zoom, pan, etc) toolbar in
embedding_in_qt.py example?
I have noticed that in animation_blit_qt.py the toolbar is available, but it
seems to be done in a different way.
Appreciate your help!
Thank you,
/Behnam

Hi,
When matplotlib uses an axis multiplier, as in x-axis in the script 
below, the scale factor is not positioned properly on the figure as is 
shown in the attached picture. The 'x1e+30' in the bottom right corner 
is clipped unless I enlarge the window.
--------------------
import Numeric, gtk
import matplotlib
matplotlib.use('GTK')
import matplotlib.figure, matplotlib.axes, matplotlib.backends.backend_gtk
# data
values = Numeric.array([-1e30, -1e30, 0,1,2,3,4,5,6,7,8,9], 'd')
# figure
figsize=(6,3)
figure = matplotlib.figure.Figure(figsize=figsize)
canvas = matplotlib.backends.backend_gtk.FigureCanvasGTK(figure)
axes = matplotlib.axes.Subplot(figure, 111)
figure.add_axes(axes)
axes.hist(values, bins=10)
axes.set_title('X-scale multiplier is hidden', fontsize=10)
# window to display
window = gtk.Window()
window.connect('destroy', gtk.main_quit)
window.add(canvas)
window.show_all()
gtk.main()
--------------------
I can use the axes position to work I guess but that seems rather 
unsatisfactory as I'll have to figure out when it's being displayed or 
not, font sizing and spacing blah blah blah.
Is there an easier way around this (or a fix?)? I'm using v0.90.0
Thanks
Tim
-- 
Tim Mitchell
Software Engineer
Applied Research Associates (NZ) Ltd. (www.aranz.com)
Ph: +64 (3) 374-6120 ext: 203
Fax: +64 (3) 374-6130
Skype: tim-mitchell

Hi,
When matplotlib uses an axis multiplier, as in x-axis in the script
below, the scale factor is not positioned properly on the figure as is
shown in the attached picture. The 'x1e+30' in the bottom right corner
is clipped unless I enlarge the window.
--------------------
import Numeric, gtk
import matplotlib
matplotlib.use('GTK')
import matplotlib.figure, matplotlib.axes, matplotlib.backends.backend_gtk
# data
values = Numeric.array([-1e30, -1e30, 0,1,2,3,4,5,6,7,8,9], 'd')
# figure
figsize=(6,3)
figure = matplotlib.figure.Figure(figsize=figsize)
canvas = matplotlib.backends.backend_gtk.FigureCanvasGTK(figure)
axes = matplotlib.axes.Subplot(figure, 111)
figure.add_axes(axes)
axes.hist(values, bins=10)
axes.set_title('X-scale multiplier is hidden', fontsize=10)
# window to display
window = gtk.Window()
window.connect('destroy', gtk.main_quit)
window.add(canvas)
window.show_all()
gtk.main()
--------------------
I can use the axes position to work I guess but that seems rather
unsatisfactory as I'll have to figure out when it's being displayed or
not, font sizing and spacing blah blah blah.
Is there an easier way around this (or a fix?)? I'm using v0.90.0
Thanks
Tim

Hi Jouni,
JKS> Try setting the pdf.fonttype parameter in your matplotlibrc file to
JKS> 42 and see if the resulting file passes validation. Matplotlib
This produces a good-looking pdf. Unfortunately, It'll be a few weeks 
before I can test it with the online figure check program, since the 
that manuscript is already submitted and I can no longer update the 
figures (for now). Once I can try again, I'll let you know.
-- 
Chris

The toolbar is just a widget so you can do that the same way you show any
other widget in Qt. Something along the lines of:
 
- build a widget
- add a layout
- add the figure widget to the layout
- add the toolbar to the layout
- connect them together (I forget exactly how but backend_qt.py
has all the code that does this)
 
From: mat...@li...
[mailto:mat...@li...] On Behalf Of Behnam
Kamrani
Sent: Thursday, May 08, 2008 11:19 AM
To: mat...@li...
Subject: [Matplotlib-users] Toolbar in QT
 
Hi, 
First I want to thanks for this amazing Matplotlib. It's fantastic!
My newbie question is how I can show the (zoom, pan, etc) toolbar in
embedding_in_qt.py example?
I have noticed that in animation_blit_qt.py the toolbar is available, but it
seems to be done in a different way. 
Appreciate your help! 
Thank you,
/Ben

On Thu, May 8, 2008 at 3:52 PM, Michael Droettboom <md...@st...> wrote:
> You can pass an "interpolation" parameter to imshow, or set a default in
> "image.interpolation" in your matplotlibrc. From the imshow docstring:
>
> * interpolation is one of:
>
> 'nearest', 'bilinear', 'bicubic', 'spline16', 'spline36',
> 'hanning', 'hamming', 'hermite', 'kaiser', 'quadric',
> 'catrom', 'gaussian', 'bessel', 'mitchell', 'sinc',
> 'lanczos', 'blackman'
>
> if interpolation is None, default to rc
> image.interpolation. See also th the filternorm and
> filterrad parameters
>
> 'nearest' is the one you want if you want to see sharp square pixels.
>
Thanks mike, thats exactly what I needed :)
>
> Søren Nielsen wrote:
>
>> Also, if I want to use matplotlib.patches on a canvas where I can actually
>> see individual pixels, is it possible to only draw on the available pixels
>> showing, and not just make a nice high resolution line that is plotted on
>> top of my image?
>>
> I'm not sure I understand the question. You mean to draw a line using
> large pixels, rather than as a vector? I don't think there's an easy way on
> that one -- but maybe someone else here can think of a creative way to do
> it.
Yes you understod it correctly, I guess I could figure out what pixels are
involved in a line or a circle for instance, and then do a scatter plot on
those pixels to form the shape.. but it sounds really computationally
heavy.. and I need something that works fast, (the user should be able to
mask a circle in an image using the pixels available in the image) Also the
circle should be shown as it is created (like choosing a center and moving
the mouse enlarges the circle).
Maybe someone tried this? masking stuff in an image seems like a fairly
common thing :)
Thanks for the help!
Soren
>
> Cheers,
> Mike
>
>>
>> Thanks,
>> Soren
>>
>>
>> ------------------------------------------------------------------------
>>
>> -------------------------------------------------------------------------
>> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't
>> miss this year's exciting event. There's still time to save 100ドル. Use
>> priority code J8TL2D2.
>> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> 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
>
>

Christopher Brown <c-...@as...> writes:
> I have read a little bit online about the difference between types 1
> and 3, but I can't find a description of which font is which type. So
> what font should I use?
Try setting the pdf.fonttype parameter in your matplotlibrc file to 42
and see if the resulting file passes validation. Matplotlib currently
only uses TrueType fonts, which can be embedded in either Type 3 or Type
42 format. I seem to recall there was some problem with the Type 42
format so Type 3 was made the default; Michael Droettboom probably
remembers it better.
If you really need to use Type 1 fonts as opposed to TrueType, there
actually is some limited support for them in the pdf backend, currently
used only by the dviread approach to using LaTeX. You could try setting
text.usetex to True - if it works out for you, you should see Type 1 TeX
fonts in the output. If not, it will probably take a little hacking to
smuggle a Type 1 font (a pfa or pfb file) past the front end, but the
pdf backend should then embed it.
-- 
Jouni K. Seppänen
http://www.iki.fi/jks

Christopher Brown <c-...@as...> writes:
> I have read a little bit online about the difference between types 1
> and 3, but I can't find a description of which font is which type. So
> what font should I use?
Try setting the pdf.fonttype parameter in your matplotlibrc file to 42
and see if the resulting file passes validation. Matplotlib currently
only uses TrueType fonts, which can be embedded in either Type 3 or Type
42 format. I seem to recall there was some problem with the Type 42
format so Type 3 was made the default; Michael Droettboom probably
remembers it better.
If you really need to use Type 1 fonts as opposed to TrueType, there
actually is some limited support for them in the pdf backend, currently
used only by the dviread approach to using LaTeX. You could try setting
text.usetex to True - if it works out for you, you should see Type 1 TeX
fonts in the output. If not, it will probably take a little hacking to
smuggle a Type 1 font (a pfa or pfb file) past the front end, but the
pdf backend should then embed it.
-- 
Jouni K. Seppänen
http://www.iki.fi/jks

Hi,
First I want to thanks for this amazing Matplotlib. It's fantastic!
My newbie question is how I can show the (zoom, pan, etc) toolbar in
embedding_in_qt.py example?
I have noticed that in animation_blit_qt.py the toolbar is available, but it
seems to be done in a different way.
Appreciate your help!
Thank you,
/Ben

Hi List,
I recently subjected some figures to an online 'quality analysis' done 
by a journal I am submitting a paper to. The analysis stated that my 
figures (created with mpl's pdf backend, and using whatever default 
fonts are used) used type 3 fonts, and that I should use type 1 whenever 
possible. I figured Arial would be a safer bet, but that also drew 
complaints. I have read a little bit online about the difference between 
types 1 and 3, but I can't find a description of which font is which 
type. So what font should I use?
-- 
Chris

You can use \AA for angstrom (to conform to the (La)TeX standard).
Cheers,
Mike
Maxim Fedorovsky wrote:
> Thanks a lot for the help) I have noticed that \angstrom is not 
> anymore supported in 0.91.
>
>
> With best regards,
> Maxim Fedorovsky.
>> Forgot to mention -- you could also apply this patch to your local 
>> copy of mathtext.py
>>
>> http://matplotlib.svn.sourceforge.net/viewvc/matplotlib/branches/v0_91_maint/lib/matplotlib/mathtext.py?r1=4816&r2=4854 
>>
>>
>> Michael Droettboom wrote:
>>> It's a bug in 0.91.2. This has been fixed on the maintenance branch 
>>> (if you're comfortable using SVN), or we do plan to release a bugfix 
>>> 0.91.3 release soon.
>>>
>>> Cheers,
>>> Mike
>>>
>>> Maxim Fedorovsky wrote:
>>> 
>>>> Dear all,
>>>>
>>>> The following TeX expression did not produce any error message 
>>>> with matplotlib versions 0.87.3 - 0.90.2 :
>>>>
>>>> r'$Wavenumber,\ [\ cm^{-1}\ ]$'
>>>>
>>>> The following error is produced with version 0.91.2 :
>>>>
>>>> <class 'matplotlib.pyparsing.ParseFatalException'>: Expected 
>>>> end of math '$'
>>>> $Wavenumber,\ [\ cm^{-1}\ ]$ (at char 0), (line:1, col:1)
>>>>
>>>> The problem comes about because of the square brackets. Could 
>>>> everyone please explain me how to handle this situation (without 
>>>> placing the brackets outside of '$$') ?
>>>>
>>>>
>>>> Thanks in advance,
>>>> Dr. Maxim Fedorovsky.
>>>>
>>>> ------------------------------------------------------------------------- 
>>>>
>>>> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
>>>> Don't miss this year's exciting event. There's still time to save 
>>>> 100ドル. Use priority code J8TL2D2. 
>>>> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone 
>>>>
>>>> _______________________________________________
>>>> 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

Hallo,
I'm quite new to matplotlib and have a problem, I was searching a solution
for the whole day :(
The thing is, I want to show different Figures at different states of my
main application.
So it is not possible for me, to perform the standard procedure, that may
looks like:
<code>
figure(1)
plot(bla)
figure(2)
plot(bla2)
show()
</code>
I have a method, in a class, that should be responsible for drawing a Figure
and its content; looks like that:
<code>
def drawFigure(self, el):
 figure(self.__figure)
 title(self.__headline)
 
 plot( el.getXValues(), el.getYValues(),str(el.getColor())+
str(el.getLineStyle())+str(el.getMarkerStyle()) )
 show()
</code>
This method exists in several instances of that class, that are created at
different times
Let I run it this way, the 1st Figure window appears. When this method is
called a 2nd time, not 'Figure 2' appears, a window named 'Tk' appears.
It is safe that the figure(f) method gets a uniquie number at every time the
method is called.
I also tried to let all this run in separate threads, using python Threads, 
but there was no success.
Please, can anybody tell me, if there is a solution, to call the show()
method more than one time,while my main script is running? 
The reason why I want to have it this way is, that between different states,
some time is elapsing. In this time the user shall have the possibility to
watch the figures, and zoom and pan and so on. When the next state is
reached, the next Figure window shall appear, while the others are kept on
the screen.
Thanks, very much.
Regards,
Thomas
-- 
View this message in context: http://www.nabble.com/How-to-handle-different-Figures%2C-created-from-different-instances-tp17127297p17127297.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Great. You may still want to let the scisoft guys know about this so 
others won't hit up against it.
Cheers,
Mike
Antonino Cucchiara wrote:
> Thanks,
> I just had to replace the /usr/lib/libiconv.2.dylib with the one I 
> have in /sw/lib/ directory .
> Now the GTK, GTKAgg backend work.
>
> Nino
>
>
> Michael Droettboom wrote:
>> You could try rebuilding matplotlib yourself. There are some OS-X 
>> directions here:
>>
>> http://www.astro.washington.edu/owen/BuildingMatplotlibForMac.html
>>
>> This may mean, however, that it won't play nice with other parts of 
>> scisoft. But this really is a matter of how and where scisoft was 
>> built (which I don't know) and where you're running it.
>>
>> Cheers,
>> Mike
>>
>> Antonino Cucchiara wrote:
>>> Is there anything I can do to update it by myself in the meanwhile?
>>> Thanks,
>>> Nino
>>>
>>> Michael Droettboom wrote:
>>>> It looks like matplotlib was built for a different version of iconv 
>>>> than what you have on your machine. This may be a scisoft 
>>>> packaging problem, you may want to bring this question to their 
>>>> attention. Perhaps the version of OSX that scisoft was built 
>>>> on/for is different than what you're running it on.
>>>>
>>>> ImportError: 
>>>> dlopen(/scisoft/i386/library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/matplotlib/backends/_backend_gdk.so, 
>>>> 2): Library not loaded: /usr/lib/libiconv.2.dylib
>>>> Referenced from: 
>>>> /scisoft/i386/library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/matplotlib/backends/_backend_gdk.so 
>>>>
>>>> Reason: Incompatible library version: _backend_gdk.so requires 
>>>> version 7.0.0 or later, but libiconv.2.dylib provides version 5.0.0
>>>>
>>>> Cheers,
>>>> Mike
>>>>
>>>> Antonino Cucchiara wrote:
>>>> 
>>>>> Hi Mike,
>>>>> I re-install the astronomy scisoft package and now the 
>>>>> segmentation fault is not anymore present when I import 
>>>>> matplotlib.mathtext.
>>>>> But when I import pylab I have the following error.
>>>>> I think it's something related with my backend choise. I tried to 
>>>>> change all the GUI possibilities in my matplotlibrc file but the 
>>>>> error is still present.
>>>>> I attached the gdb python output (I don't know if it helps) and 
>>>>> the output of my "import pylab" command
>>>>>
>>>>>
>>>>>
>>>>> Michael Droettboom wrote:
>>>>> 
>>>>>> We need some more information.
>>>>>>
>>>>>> Can you set "verbose.level" to "debug-annoying" in your 
>>>>>> ~/.matplotlib/matplotlibrc file?
>>>>>>
>>>>>> Also, if you have the development tools (XCode) installed on your 
>>>>>> mac, can you run python in gdb?
>>>>>>
>>>>>> 
>>>>>>> gdb python
>>>>>>> 
>>>>>> GNU gdb Red Hat Linux (6.3.0.0-1.153.el4_6.2rh)
>>>>>> Copyright 2004 Free Software Foundation, Inc.
>>>>>> GDB is free software, covered by the GNU General Public License, 
>>>>>> and you are
>>>>>> welcome to change it and/or distribute copies of it under certain 
>>>>>> conditions.
>>>>>> Type "show copying" to see the conditions.
>>>>>> There is absolutely no warranty for GDB. Type "show warranty" 
>>>>>> for details.
>>>>>> This GDB was configured as "i386-redhat-linux-gnu"...Using host 
>>>>>> libthread_db library "/lib/tls/libthread_db.so.1".
>>>>>>
>>>>>> (gdb) run -c "import pylab"
>>>>>>
>>>>>> ... and send us the output...
>>>>>>
>>>>>> Cheers,
>>>>>> Mike
>>>>>>
>>>>>>
>>>>>> Antonino Cucchiara wrote:
>>>>>> 
>>>>>>> Hello,
>>>>>>> I am experiencing a problem recently on my Mac OsX.
>>>>>>> When I try to import mathplotlib.mathtext or pylab I received a 
>>>>>>> segmentation fault error.
>>>>>>>
>>>>>>> It wasn't like that before, and I recently update my matplotlib 
>>>>>>> and scipy via fink.
>>>>>>> It worked fine few weeks ago.
>>>>>>>
>>>>>>> I hope someone can help me.
>>>>>>> Thanks,
>>>>>>> Nino
>>>>>>>
>>>>>>>
>>>>>>> ------------------------------------------------------------------------- 
>>>>>>>
>>>>>>> This SF.net email is sponsored by the 2008 JavaOne(SM) 
>>>>>>> Conference Don't miss this year's exciting event. There's still 
>>>>>>> time to save 100ドル. Use priority code J8TL2D2. 
>>>>>>> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone 
>>>>>>>
>>>>>>> _______________________________________________
>>>>>>> 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

Thanks a lot for the help) I have noticed that \angstrom is not anymore 
supported in 0.91.
With best regards,
Maxim Fedorovsky.
> Forgot to mention -- you could also apply this patch to your local 
> copy of mathtext.py
>
> http://matplotlib.svn.sourceforge.net/viewvc/matplotlib/branches/v0_91_maint/lib/matplotlib/mathtext.py?r1=4816&r2=4854 
>
>
> Michael Droettboom wrote:
>> It's a bug in 0.91.2. This has been fixed on the maintenance branch 
>> (if you're comfortable using SVN), or we do plan to release a bugfix 
>> 0.91.3 release soon.
>>
>> Cheers,
>> Mike
>>
>> Maxim Fedorovsky wrote:
>> 
>>> Dear all,
>>>
>>> The following TeX expression did not produce any error message 
>>> with matplotlib versions 0.87.3 - 0.90.2 :
>>>
>>> r'$Wavenumber,\ [\ cm^{-1}\ ]$'
>>>
>>> The following error is produced with version 0.91.2 :
>>>
>>> <class 'matplotlib.pyparsing.ParseFatalException'>: Expected end 
>>> of math '$'
>>> $Wavenumber,\ [\ cm^{-1}\ ]$ (at char 0), (line:1, col:1)
>>>
>>> The problem comes about because of the square brackets. Could 
>>> everyone please explain me how to handle this situation (without 
>>> placing the brackets outside of '$$') ?
>>>
>>>
>>> Thanks in advance,
>>> Dr. Maxim Fedorovsky.
>>>
>>> ------------------------------------------------------------------------- 
>>>
>>> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
>>> Don't miss this year's exciting event. There's still time to save 
>>> 100ドル. Use priority code J8TL2D2. 
>>> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone 
>>>
>>> _______________________________________________
>>> Matplotlib-users mailing list
>>> Mat...@li...
>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>> 
>>
>> 
>

Thanks,
I just had to replace the /usr/lib/libiconv.2.dylib with the one I have 
in /sw/lib/ directory .
Now the GTK, GTKAgg backend work.
Nino
Michael Droettboom wrote:
> You could try rebuilding matplotlib yourself. There are some OS-X 
> directions here:
>
> http://www.astro.washington.edu/owen/BuildingMatplotlibForMac.html
>
> This may mean, however, that it won't play nice with other parts of 
> scisoft. But this really is a matter of how and where scisoft was 
> built (which I don't know) and where you're running it.
>
> Cheers,
> Mike
>
> Antonino Cucchiara wrote:
>> Is there anything I can do to update it by myself in the meanwhile?
>> Thanks,
>> Nino
>>
>> Michael Droettboom wrote:
>>> It looks like matplotlib was built for a different version of iconv 
>>> than what you have on your machine. This may be a scisoft packaging 
>>> problem, you may want to bring this question to their attention. 
>>> Perhaps the version of OSX that scisoft was built on/for is 
>>> different than what you're running it on.
>>>
>>> ImportError: 
>>> dlopen(/scisoft/i386/library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/matplotlib/backends/_backend_gdk.so, 
>>> 2): Library not loaded: /usr/lib/libiconv.2.dylib
>>> Referenced from: 
>>> /scisoft/i386/library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/matplotlib/backends/_backend_gdk.so 
>>>
>>> Reason: Incompatible library version: _backend_gdk.so requires 
>>> version 7.0.0 or later, but libiconv.2.dylib provides version 5.0.0
>>>
>>> Cheers,
>>> Mike
>>>
>>> Antonino Cucchiara wrote:
>>> 
>>>> Hi Mike,
>>>> I re-install the astronomy scisoft package and now the segmentation 
>>>> fault is not anymore present when I import matplotlib.mathtext.
>>>> But when I import pylab I have the following error.
>>>> I think it's something related with my backend choise. I tried to 
>>>> change all the GUI possibilities in my matplotlibrc file but the 
>>>> error is still present.
>>>> I attached the gdb python output (I don't know if it helps) and the 
>>>> output of my "import pylab" command
>>>>
>>>>
>>>>
>>>> Michael Droettboom wrote:
>>>> 
>>>>> We need some more information.
>>>>>
>>>>> Can you set "verbose.level" to "debug-annoying" in your 
>>>>> ~/.matplotlib/matplotlibrc file?
>>>>>
>>>>> Also, if you have the development tools (XCode) installed on your 
>>>>> mac, can you run python in gdb?
>>>>>
>>>>> 
>>>>>> gdb python
>>>>>> 
>>>>> GNU gdb Red Hat Linux (6.3.0.0-1.153.el4_6.2rh)
>>>>> Copyright 2004 Free Software Foundation, Inc.
>>>>> GDB is free software, covered by the GNU General Public License, 
>>>>> and you are
>>>>> welcome to change it and/or distribute copies of it under certain 
>>>>> conditions.
>>>>> Type "show copying" to see the conditions.
>>>>> There is absolutely no warranty for GDB. Type "show warranty" for 
>>>>> details.
>>>>> This GDB was configured as "i386-redhat-linux-gnu"...Using host 
>>>>> libthread_db library "/lib/tls/libthread_db.so.1".
>>>>>
>>>>> (gdb) run -c "import pylab"
>>>>>
>>>>> ... and send us the output...
>>>>>
>>>>> Cheers,
>>>>> Mike
>>>>>
>>>>>
>>>>> Antonino Cucchiara wrote:
>>>>> 
>>>>>> Hello,
>>>>>> I am experiencing a problem recently on my Mac OsX.
>>>>>> When I try to import mathplotlib.mathtext or pylab I received a 
>>>>>> segmentation fault error.
>>>>>>
>>>>>> It wasn't like that before, and I recently update my matplotlib 
>>>>>> and scipy via fink.
>>>>>> It worked fine few weeks ago.
>>>>>>
>>>>>> I hope someone can help me.
>>>>>> Thanks,
>>>>>> Nino
>>>>>>
>>>>>>
>>>>>> ------------------------------------------------------------------------- 
>>>>>>
>>>>>> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
>>>>>> Don't miss this year's exciting event. There's still time to save 
>>>>>> 100ドル. Use priority code J8TL2D2. 
>>>>>> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone 
>>>>>>
>>>>>> _______________________________________________
>>>>>> Matplotlib-users mailing list
>>>>>> Mat...@li...
>>>>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>>>>> 
>>>> 
>>>
>>> 
>>
>>
>
-- 
Antonino Cucchiara
PhD candidate
Department of Astronomy&Astrophysics
Penn State University
website: www.astro.psu.edu/~cucchiara/

You could try rebuilding matplotlib yourself. There are some OS-X 
directions here:
http://www.astro.washington.edu/owen/BuildingMatplotlibForMac.html
This may mean, however, that it won't play nice with other parts of 
scisoft. But this really is a matter of how and where scisoft was built 
(which I don't know) and where you're running it.
Cheers,
Mike
Antonino Cucchiara wrote:
> Is there anything I can do to update it by myself in the meanwhile?
> Thanks,
> Nino
>
> Michael Droettboom wrote:
>> It looks like matplotlib was built for a different version of iconv 
>> than what you have on your machine. This may be a scisoft packaging 
>> problem, you may want to bring this question to their attention. 
>> Perhaps the version of OSX that scisoft was built on/for is different 
>> than what you're running it on.
>>
>> ImportError: 
>> dlopen(/scisoft/i386/library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/matplotlib/backends/_backend_gdk.so, 
>> 2): Library not loaded: /usr/lib/libiconv.2.dylib
>> Referenced from: 
>> /scisoft/i386/library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/matplotlib/backends/_backend_gdk.so 
>>
>> Reason: Incompatible library version: _backend_gdk.so requires 
>> version 7.0.0 or later, but libiconv.2.dylib provides version 5.0.0
>>
>> Cheers,
>> Mike
>>
>> Antonino Cucchiara wrote:
>> 
>>> Hi Mike,
>>> I re-install the astronomy scisoft package and now the segmentation 
>>> fault is not anymore present when I import matplotlib.mathtext.
>>> But when I import pylab I have the following error.
>>> I think it's something related with my backend choise. I tried to 
>>> change all the GUI possibilities in my matplotlibrc file but the 
>>> error is still present.
>>> I attached the gdb python output (I don't know if it helps) and the 
>>> output of my "import pylab" command
>>>
>>>
>>>
>>> Michael Droettboom wrote:
>>> 
>>>> We need some more information.
>>>>
>>>> Can you set "verbose.level" to "debug-annoying" in your 
>>>> ~/.matplotlib/matplotlibrc file?
>>>>
>>>> Also, if you have the development tools (XCode) installed on your 
>>>> mac, can you run python in gdb?
>>>>
>>>> 
>>>>> gdb python
>>>>> 
>>>> GNU gdb Red Hat Linux (6.3.0.0-1.153.el4_6.2rh)
>>>> Copyright 2004 Free Software Foundation, Inc.
>>>> GDB is free software, covered by the GNU General Public License, 
>>>> and you are
>>>> welcome to change it and/or distribute copies of it under certain 
>>>> conditions.
>>>> Type "show copying" to see the conditions.
>>>> There is absolutely no warranty for GDB. Type "show warranty" for 
>>>> details.
>>>> This GDB was configured as "i386-redhat-linux-gnu"...Using host 
>>>> libthread_db library "/lib/tls/libthread_db.so.1".
>>>>
>>>> (gdb) run -c "import pylab"
>>>>
>>>> ... and send us the output...
>>>>
>>>> Cheers,
>>>> Mike
>>>>
>>>>
>>>> Antonino Cucchiara wrote:
>>>> 
>>>>> Hello,
>>>>> I am experiencing a problem recently on my Mac OsX.
>>>>> When I try to import mathplotlib.mathtext or pylab I received a 
>>>>> segmentation fault error.
>>>>>
>>>>> It wasn't like that before, and I recently update my matplotlib 
>>>>> and scipy via fink.
>>>>> It worked fine few weeks ago.
>>>>>
>>>>> I hope someone can help me.
>>>>> Thanks,
>>>>> Nino
>>>>>
>>>>>
>>>>> ------------------------------------------------------------------------- 
>>>>>
>>>>> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
>>>>> Don't miss this year's exciting event. There's still time to save 
>>>>> 100ドル. Use priority code J8TL2D2. 
>>>>> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone 
>>>>>
>>>>> _______________________________________________
>>>>> 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

You can pass an "interpolation" parameter to imshow, or set a default in 
"image.interpolation" in your matplotlibrc. From the imshow docstring:
 * interpolation is one of:
 'nearest', 'bilinear', 'bicubic', 'spline16', 'spline36',
 'hanning', 'hamming', 'hermite', 'kaiser', 'quadric',
 'catrom', 'gaussian', 'bessel', 'mitchell', 'sinc',
 'lanczos', 'blackman'
 if interpolation is None, default to rc
 image.interpolation. See also th the filternorm and
 filterrad parameters
'nearest' is the one you want if you want to see sharp square pixels.
Søren Nielsen wrote:
> Also, if I want to use matplotlib.patches on a canvas where I can 
> actually see individual pixels, is it possible to only draw on the 
> available pixels showing, and not just make a nice high resolution 
> line that is plotted on top of my image?
I'm not sure I understand the question. You mean to draw a line using 
large pixels, rather than as a vector? I don't think there's an easy 
way on that one -- but maybe someone else here can think of a creative 
way to do it.
Cheers,
Mike
>
> Thanks,
> Soren
>
>
> ------------------------------------------------------------------------
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
> Don't miss this year's exciting event. There's still time to save 100ドル. 
> Use priority code J8TL2D2. 
> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> ------------------------------------------------------------------------
>
> _______________________________________________
> 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 anything I can do to update it by myself in the meanwhile?
Thanks,
Nino
Michael Droettboom wrote:
> It looks like matplotlib was built for a different version of iconv than 
> what you have on your machine. This may be a scisoft packaging problem, 
> you may want to bring this question to their attention. Perhaps the 
> version of OSX that scisoft was built on/for is different than what 
> you're running it on.
>
> ImportError: 
> dlopen(/scisoft/i386/library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/matplotlib/backends/_backend_gdk.so, 
> 2): Library not loaded: /usr/lib/libiconv.2.dylib
> Referenced from: 
> /scisoft/i386/library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/matplotlib/backends/_backend_gdk.so
> Reason: Incompatible library version: _backend_gdk.so requires version 
> 7.0.0 or later, but libiconv.2.dylib provides version 5.0.0
>
> Cheers,
> Mike
>
> Antonino Cucchiara wrote:
> 
>> Hi Mike,
>> I re-install the astronomy scisoft package and now the segmentation 
>> fault is not anymore present when I import matplotlib.mathtext.
>> But when I import pylab I have the following error.
>> I think it's something related with my backend choise. I tried to 
>> change all the GUI possibilities in my matplotlibrc file but the error 
>> is still present.
>> I attached the gdb python output (I don't know if it helps) and the 
>> output of my "import pylab" command
>>
>>
>>
>> Michael Droettboom wrote:
>> 
>>> We need some more information.
>>>
>>> Can you set "verbose.level" to "debug-annoying" in your 
>>> ~/.matplotlib/matplotlibrc file?
>>>
>>> Also, if you have the development tools (XCode) installed on your 
>>> mac, can you run python in gdb?
>>>
>>> 
>>>> gdb python
>>>> 
>>> GNU gdb Red Hat Linux (6.3.0.0-1.153.el4_6.2rh)
>>> Copyright 2004 Free Software Foundation, Inc.
>>> GDB is free software, covered by the GNU General Public License, and 
>>> you are
>>> welcome to change it and/or distribute copies of it under certain 
>>> conditions.
>>> Type "show copying" to see the conditions.
>>> There is absolutely no warranty for GDB. Type "show warranty" for 
>>> details.
>>> This GDB was configured as "i386-redhat-linux-gnu"...Using host 
>>> libthread_db library "/lib/tls/libthread_db.so.1".
>>>
>>> (gdb) run -c "import pylab"
>>>
>>> ... and send us the output...
>>>
>>> Cheers,
>>> Mike
>>>
>>>
>>> Antonino Cucchiara wrote:
>>> 
>>>> Hello,
>>>> I am experiencing a problem recently on my Mac OsX.
>>>> When I try to import mathplotlib.mathtext or pylab I received a 
>>>> segmentation fault error.
>>>>
>>>> It wasn't like that before, and I recently update my matplotlib and 
>>>> scipy via fink.
>>>> It worked fine few weeks ago.
>>>>
>>>> I hope someone can help me.
>>>> Thanks,
>>>> Nino
>>>>
>>>>
>>>> ------------------------------------------------------------------------- 
>>>>
>>>> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
>>>> Don't miss this year's exciting event. There's still time to save 
>>>> 100ドル. Use priority code J8TL2D2. 
>>>> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone 
>>>>
>>>> _______________________________________________
>>>> Matplotlib-users mailing list
>>>> Mat...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>>> 
>>>> 
>> 
>
> 
-- 
Antonino Cucchiara
PhD candidate
Department of Astronomy&Astrophysics
Penn State University
website: www.astro.psu.edu/~cucchiara/

It looks like matplotlib was built for a different version of iconv than 
what you have on your machine. This may be a scisoft packaging problem, 
you may want to bring this question to their attention. Perhaps the 
version of OSX that scisoft was built on/for is different than what 
you're running it on.
ImportError: 
dlopen(/scisoft/i386/library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/matplotlib/backends/_backend_gdk.so, 
2): Library not loaded: /usr/lib/libiconv.2.dylib
 Referenced from: 
/scisoft/i386/library/Frameworks/Python.framework/Versions/2.5/lib/python2.5/site-packages/matplotlib/backends/_backend_gdk.so
 Reason: Incompatible library version: _backend_gdk.so requires version 
7.0.0 or later, but libiconv.2.dylib provides version 5.0.0
Cheers,
Mike
Antonino Cucchiara wrote:
> Hi Mike,
> I re-install the astronomy scisoft package and now the segmentation 
> fault is not anymore present when I import matplotlib.mathtext.
> But when I import pylab I have the following error.
> I think it's something related with my backend choise. I tried to 
> change all the GUI possibilities in my matplotlibrc file but the error 
> is still present.
> I attached the gdb python output (I don't know if it helps) and the 
> output of my "import pylab" command
>
>
>
> Michael Droettboom wrote:
>> We need some more information.
>>
>> Can you set "verbose.level" to "debug-annoying" in your 
>> ~/.matplotlib/matplotlibrc file?
>>
>> Also, if you have the development tools (XCode) installed on your 
>> mac, can you run python in gdb?
>>
>> > gdb python
>> GNU gdb Red Hat Linux (6.3.0.0-1.153.el4_6.2rh)
>> Copyright 2004 Free Software Foundation, Inc.
>> GDB is free software, covered by the GNU General Public License, and 
>> you are
>> welcome to change it and/or distribute copies of it under certain 
>> conditions.
>> Type "show copying" to see the conditions.
>> There is absolutely no warranty for GDB. Type "show warranty" for 
>> details.
>> This GDB was configured as "i386-redhat-linux-gnu"...Using host 
>> libthread_db library "/lib/tls/libthread_db.so.1".
>>
>> (gdb) run -c "import pylab"
>>
>> ... and send us the output...
>>
>> Cheers,
>> Mike
>>
>>
>> Antonino Cucchiara wrote:
>>> Hello,
>>> I am experiencing a problem recently on my Mac OsX.
>>> When I try to import mathplotlib.mathtext or pylab I received a 
>>> segmentation fault error.
>>>
>>> It wasn't like that before, and I recently update my matplotlib and 
>>> scipy via fink.
>>> It worked fine few weeks ago.
>>>
>>> I hope someone can help me.
>>> Thanks,
>>> Nino
>>>
>>>
>>> ------------------------------------------------------------------------- 
>>>
>>> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
>>> Don't miss this year's exciting event. There's still time to save 
>>> 100ドル. Use priority code J8TL2D2. 
>>> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone 
>>>
>>> _______________________________________________
>>> 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

Hi,
I have a displayed an image with imshow() under wxPython, and it looks great
with the WxAgg figure canvas.. but sometimes I want to actually see pixels
when I zoom.. and not just a nice smoothed image. Is it possible to disable
the smoothing? The way it is now, I will never see square pixels no matter
how close i zoom in..
Also, if I want to use matplotlib.patches on a canvas where I can actually
see individual pixels, is it possible to only draw on the available pixels
showing, and not just make a nice high resolution line that is plotted on
top of my image?
Thanks,
Soren