matplotlib-users Mailing List for matplotlib

I verified that matplotlib-0.70 works w/ numarray-1.1.1 on
Windows XP using a quick "plot([1,2,3,4]); show();" test. I
noted that TkAgg is now raising an exception when the plot
window is closed, but the plot looked fine to me.
Cheers,
Todd
Exception in Tkinter callback
Traceback (most recent call last):
 File "h:\python24\lib\lib-tk\Tkinter.py", line 1345, in __call__
 return self.func(*args)
 File
"H:\python24\lib\site-packages\matplotlib\backends\backend_tkagg.py",
lin
e 215, in key_release
 FigureCanvasBase.key_release_event(self, key)
 File
"H:\python24\lib\site-packages\matplotlib\backend_bases.py",
line 677, in
 key_release_event
 event = KeyEvent('key_release_event', self, key,
self._lastx, self._lasty)
 File
"H:\python24\lib\site-packages\matplotlib\backend_bases.py",
line 640, in
 __init__
 LocationEvent.__init__(self, name, canvas, x, y)
 File
"H:\python24\lib\site-packages\matplotlib\backend_bases.py",
line 566, in
 __init__
 if a.in_axes(self.x, self.y):
 File "H:\python24\lib\site-packages\matplotlib\axes.py",
line 1544, in in_axes
 return self.bbox.contains(xwin, ywin)
TypeError: float() argument must be a string or a number

>>>>> "Darren" == Darren Dale <dd...@co...> writes:
 Darren> I think the "Installing on Windows" note is out of
 Darren> date. Recently, the MPL dev's changed the install code to
 Darren> automatically detect numarray and Numeric, and to install
 Darren> accordingly. This is why you dont see an installer
 Darren> specifically for numarray or numeric.
Yep, that's right. Perhaps someone else can verify that matplotlib
with numerux : numarray works on python2.4 / win32 just to verify that
I didn't screw up the 2.4 installer.
 Darren> As for solving your problem, I don't have an answer but
 Darren> let me offer a couple suggestions. Can you import numarray
 Darren> from an interactive python session? Could you try to
 Darren> temporarily install Numeric, and see if MPL can install?
Also verify that your http://matplotlib.sf.net/.matplotlibrc setting
for numerix is numarray. While testing, you may want to run with
--verbose-helpful which may provide some additional diagnoistic
information. Note you can also select numeric vs numarray from the
DOS shell, which can be helpful while debugging this problem
 c:> python myscript.py --verbose-helpful --Numeric
 c:> python myscript.py --verbose-helpful --numarray
Hope this helps,
JDH

I'm new to both Python and matplotlib.
I'm using this on MS Win with Python 2.3 and matplotlib 0.7. and ipython 
as IDE.
A substantial amount of memory is allocated when a array of constantly 
varying values is plotted, as example:
# Sawtooth signal 32768 elements long to show memory usage.
y = ones(2**15)
y[1::2] = 0
plot(y)
if the figure created is now killed, the memory is not released, and 
more memory is allocated each time the plot() command is executed. Thr 
memory is only returned to the system once python is closed.
Plotting a signal with less amplitude variation or slow amplitude change 
( for instance plot(ones2**15) or plot(sin(arange(0,10,10./2**15)) ) 
uses a lot less memory, so the leak is not as obvious, but still there.
Any suggestions from the group on resolving this?
Thanks,
Hennie

Hi Greg,
> Hi folks. Just setting up a new machine with Python 2.4, and wanted to
> check something. Under "Installing on Windows", the Matplotlib pages at
> sf.net say:
>
> For standard python installations, you will also need to install
> either Numeric or numarray in addition to the matplotlib installer.
> matplotlib provides installers for Numeric and numarray users. It is
> important that you pick the matplotlib installer that corresponds to
> your array package. Ie, if you mostly work with numarray arrays, use
> the matplotlib numarray installer. matplotlib has a numerix setting in
> the matplotlib rc file (which by default resides in
> c:\python23\share\matplotlitb\.matplotlibrc) and you should make sure
> this setting corresponds to your preferred array package.
>
> However, the download page:
>
> 
> http://sourceforge.net/project/showfiles.php?group_id=80706&package_id=8247
>4
>
> only differentiates installers by Python version; there's no hint which
> one is for Numpy, numarray, etc. The one I grabbed
> (matplotlib-0.70.1-win32.py2.4.exe) installed, but complains about not
> being able to find numarray (which is installed, for Python 2.4). I'd be
> happy to use Numpy, but the only releases available are for Python 2.3.
>
> Any suggestions?
>
I think the "Installing on Windows" note is out of date. Recently, the MPL 
dev's changed the install code to automatically detect numarray and Numeric, 
and to install accordingly. This is why you dont see an installer 
specifically for numarray or numeric.
As for solving your problem, I don't have an answer but let me offer a couple 
suggestions. Can you import numarray from an interactive python session? 
Could you try to temporarily install Numeric, and see if MPL can install?
Darren

Hi folks. Just setting up a new machine with Python 2.4, and wanted to
check something. Under "Installing on Windows", the Matplotlib pages at
sf.net say:
 For standard python installations, you will also need to install
 either Numeric or numarray in addition to the matplotlib installer.
 matplotlib provides installers for Numeric and numarray users. It is
 important that you pick the matplotlib installer that corresponds to
 your array package. Ie, if you mostly work with numarray arrays, use
 the matplotlib numarray installer. matplotlib has a numerix setting in
 the matplotlib rc file (which by default resides in
 c:\python23\share\matplotlitb\.matplotlibrc) and you should make sure
 this setting corresponds to your preferred array package.
However, the download page:
 http://sourceforge.net/project/showfiles.php?group_id=80706&package_id=82474
only differentiates installers by Python version; there's no hint which
one is for Numpy, numarray, etc. The one I grabbed
(matplotlib-0.70.1-win32.py2.4.exe) installed, but complains about not
being able to find numarray (which is installed, for Python 2.4). I'd be
happy to use Numpy, but the only releases available are for Python 2.3.
Any suggestions?
Thanks,
Greg Wilson