matplotlib-users Mailing List for matplotlib

Hi Everyone,

I'm using plot_date to scatter a ton of data retrieved from a database
and have encountered an error with the date ticks.

I more or less copied the format of the example in
http://matplotlib.sourceforge.net/examples/date_demo2.py and my ticks
ended up being months in order but never making it past the year 2001
(they just recycle) though my data begin in 2000 and end in 2004.

I took another look at the plot generated by date_demo2.py and it turns
out that the same is true there.

I'm thinking this is not intentional, but if so does anyone have any
advice or example scripts that might help me fix my plot?

Thank you,

Travis Brady
-- 
  Travis Brady
  td...@fa...

hi, I am plotting two figures in a row and I am getting a duplicated colorb=
ar=20
in the second figure:

imshow(stats[0])
title('Distance Matrix')
colorbar()
savefig('dm.png')
imshow(stats[12])
colorbar()
title('Conectivity Matrix')
savefig('cm.png')

I tried omitting the second colorbar() command. Then I get a single colorba=
r=20
in the second figure but with same scale of figure 1.

What am I doing wrong?

=46l=E1vio

>>>>> "Graeme" == Graeme O'Keefe <gj...@ne...> writes:

    Graeme> Hi, I've installed matplotlib 0.71 on Mac OS X 10.3 which
    Graeme> completed without errors.

    Graeme> when I run python "import matplotlib" is fine, but "import
    Graeme> pylab" bombs with "ImportError: cannot import name POLAR"


rm -rf the build directory *and* site-packages/matplotlib and
rebuild/reinstall.  This cures the problem 95% of the time.  It is
usually caused from some detritus from a previous build or install
that distutils doesn't handle so well.  Let mw know...

JDH

>>>>> "Eric" == Eric Emsellem <ems...@ob...> writes:

    Eric> Hi John, sorry to bother you about this again, but is there
    Eric> a fix to the bug in the imshow routine (using the
    Eric> aspect='preserve' option).  If not, let me know if you have
    Eric> an idea when this could be done (no pushing there, just to
    Eric> plan things on my side: I am at the moment using an
    Eric> artificial way to replace that option by calculating
    Eric> coordinates on the plot, just to know if I need to go on
    Eric> with that or not).

Hi Eric, there is not a fix yet, but there is a workaround in CVS.  

I played around with some things that got the initial coords right but
then broke under window resizes and I rolled these changes back.  It
is a somewhat tough problem, to preserve aspect ration under possible
figure and axes resizes, and/or view limit navigation.  It is a very
high priority bug for me, but because it is also a hard one its taking
a little while.  SO I can't give you a firm date -- I'll try and get
something serviceable in the next couple of weeks.

This is the same bug that causes zoom to rect on aspect=preserve plots
to get the wrong limits.

Fernando Perez submitted a patch in CVS that works around the problem
for now using the "matshow" command.  This command takes the same
arguments as imshow and creates a figure and an axes which are scaled
to have the same dimensions as the array.  Thus if you don't resize
your window, you can plot an array with matshow with either free or
preserve and the aspect ratio will be correct.  But, if you zoom to
rect or resize your window, even with matshow you'll lose the aspect
ratio.  I'll include the docstring for matshow below so you can see if
you're interested

I've very aware of the scientists need to "just see the data" w/o any
fancy stuff.  Hopefully a fix will be in soon.

JDH


matshow:

    """Display an array as a matrix.

    The origin is set at the upper left hand corner and rows (first dimension
    of the array) are displayed horizontally.  The aspect ratio of the figure
    window is that of the array, as long as it is possible to fit it within
    the constraints of your figure.figsize_min/max parameters with no
    stretching.  If the window dimensions can't accomodate this (extremely
    tall/wide arrays), some stretching will inevitably occur.

    matshow() calls imshow() with args and **kwargs, but by default
    it sets interpolation='nearest' (unless you override it).  All
    other arguments and keywords are passed to imshow(), so see its
    docstring for further details.

    Tick labels for the xaxis are placed on top by default.
    
    return value is a (fig, ax, im) tuple

    Example usage:
    
    def samplemat(dims):
        aa = zeros(dims)
        for i in range(min(dims)):
            aa[i,i] = i
        return aa

    dimlist = [(12,12),(128,64),(64,512),(2048,256)]

    
    for d in dimlist:
        fig, ax, im = matshow(samplemat(d))
    show()
    
    """

Hi John,
sorry to bother you about this again, but is there a fix to the bug
in the imshow routine (using the aspect='preserve' option).
If not, let me know if you have an idea when this could be done
(no pushing there, just to plan things on my side: I am at the moment
using an artificial way to replace that option by calculating coordinates
on the plot, just to know if I need to go on with that or not).

Let me know; Thanks!

Eric

-- 
===============================================================
Observatoire de Lyon                 ems...@ob...
9 av. Charles-Andre                      tel: +33 4 78 86 83 84
69561 Saint-Genis Laval Cedex            fax: +33 4 78 86 83 86
France               http://www-obs.univ-lyon1.fr/eric.emsellem
===============================================================

Hi,

I've installed matplotlib 0.71 on Mac OS X 10.3 which completed without  
errors.

when I run python
	"import matplotlib"
is fine, but
	"import pylab"
bombs with
	"ImportError: cannot import name POLAR"

see below  for a screendump of the reported errors when I try to import  
pylab.
I did not run this in the build directory.

/System/Library/Frameworks/Python.framework/Versions/2.3/lib/python2.3/ 
site-packages
(which is linked to /Library/Python/2.3)

/Library/Python/2.3/matplotlib/transforms.py, line 189: from  
_transforms import IDENTITY, LOG10, POLAR, Func, FuncXY
/Library/Python/2.3/matplotlib/_transforms.so exists and is 3503536  
bytes in size.

here is a screendump of the reported errors when I try to import pylab.
I did not run this in the build directory.

can anyone help?

regards,

Graeme

[graeme:~/Desktop/downloads/matplotlib-0.71] gjok% python
Python 2.3 (#1, Sep 13 2003, 00:49:11)
[GCC 3.3 20030304 (Apple Computer, Inc. build 1495)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
 >>> import pylab
Traceback (most recent call last):
   File "<stdin>", line 1, in ?
   File  
"/System/Library/Frameworks/Python.framework/Versions/2.3/lib/ 
python2.3/site-packages/pylab.py", line 1, in ?
     from matplotlib.pylab import *
   File  
"/System/Library/Frameworks/Python.framework/Versions/2.3/lib/ 
python2.3/site-packages/matplotlib/pylab.py", line 190, in ?
     from axes import Axes, PolarAxes
   File  
"/System/Library/Frameworks/Python.framework/Versions/2.3/lib/ 
python2.3/site-packages/matplotlib/axes.py", line 12, in ?
     from artist import Artist
   File  
"/System/Library/Frameworks/Python.framework/Versions/2.3/lib/ 
python2.3/site-packages/matplotlib/artist.py", line 3, in ?
     from transforms import identity_transform
   File  
"/System/Library/Frameworks/Python.framework/Versions/2.3/lib/ 
python2.3/site-packages/matplotlib/transforms.py", line 189, in ?
     from _transforms import IDENTITY, LOG10, POLAR, Func, FuncXY
ImportError: cannot import name POLAR

>>>>> "Chris" == Chris Barker <Chr...@no...> writes:

    Chris> Hi all, I think I've successfully build a binary matplotlib
    Chris> package for use with Apple's Python on OS-X 10.3.* In
    Chris> theory the only other thing it requires is Numeric, and I
    Chris> have a package for that too (courtesy of Bob Ippolito).

    Chris> It's almost 5 MB, so I haven't included it here. I also
    Chris> don't have a web site to put it on at the moment, so:

    Chris> A) if you send me a note, I'll send it to you B) if you
    Chris> have a place to host it, please volunteer to do so.

    Chris> I'd like people to test it out, all I know now is that it
    Chris> works on my system.

    Chris> Once it's been better tested, perhaps we can get it put on
    Chris> the sourceforge download page.

High Chris,

Thanks for doing this.  Are either the GTK or Tk GUI backends compiled
in.  If not, wx* will still work on boxes that have wxpython
installed, but it would be a good idea to summarize what is available
in the package.

Thanks!
JDH

Hi all,

I think I've successfully build a binary matplotlib package for use with 
Apple's Python on OS-X 10.3.* In theory the only other thing it requires 
is Numeric, and I have a package for that too (courtesy of Bob Ippolito).

It's almost 5 MB, so I haven't included it here. I also don't have a web 
site to put it on at the moment, so:

A) if you send me a note, I'll send it to you
B) if you have a place to host it, please volunteer to do so.

I'd like people to test it out, all I know now is that it works on my 
system.

Once it's been better tested, perhaps we can get it put on the 
sourceforge download page.

thanks,
-Chris




-- 
Christopher Barker, Ph.D.
Oceanographer
                                     		
NOAA/OR&R/HAZMAT         (206) 526-6959   voice
7600 Sand Point Way NE   (206) 526-6329   fax
Seattle, WA  98115       (206) 526-6317   main reception

Chr...@no...