matplotlib-users — Discussion related to using matplotlib

Re: [Matplotlib-users] slowdown when using pcolor and TkAgg; help?

[Perry G. <pe...@st...>]

My guess is that the hold mode is true (that's the default). I think 
this means that every time you call .pcolor that you are just appending 
another one to the axes. You should either change the default hold mode 
or call the .clear() method of the axes object before calling the 
.pcolor method. If you don't do that it is rendering n previous 
versions and the time it takes grows linearly.

Perry

On Mar 16, 2005, at 2:47 PM, Russell E. Owen wrote:

> I have appended some code which draws a varying pcolor plot in a 
> Tkinter
> window. The plot shows two 5x5 white rings which shrink to 3x3 and then
> 1x3, then cycles through this repeatedly.
>
> Unfortunately, it slows down with each cycle!!!
>
> Are we doing something wrong? Or can we work around the problem 
> somehow?
>
> (By the way, I did not write the code, though I cleaned up some 
> portions
> of it. I realize the data generation can be cleaned up, but it's not
> slowing things down).
>
> -- Russell
>
> P.S. At risk of distracting from the main problem: I am not entirely
> sure I'm commanding a redraw in the correct fashion. I'm fuzzy on the
> difference between show and plot. Using canvas.show() and canvas.plot()
> both seem to work. Just calling plot() doesn't. Any hints on this or
> pointers to documentation would be appreciated.
>
> Also, is pcolor documented anywhere?
>
> ----- the code -----
>
> import Tkinter as Tk
> from matplotlib.pylab import *
> import matplotlib.numerix as num
> from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg,
> NavigationToolbar2TkAgg
>
> root = Tk.Tk()
> a =[ [0,1,2,3,4,5,6,\
> 90,91,92,93,94,95,96,15,30,45,\
> 60,75,21,36,51,66,81,\
> 8,9,10,11,12,13,14,98,99,\
> 100,101,102,103,104,\
> 23,38,53,68,83,29,44,59,74,89],\
> [16,17,18,19,20,76,77,78,79,80,35,50,65,\
> 31,46,61,24,25,26,27,28,\
> 84,85,86,87,88,39,54,69,43,58,73],\
> [32,33,34,62,63,64,47,49,40,41,42,\
> 70,71,72,55,57],\
> [48,56] ]
>
> ioff()
> f = figure(figsize=(3,3))
> sp = f.add_subplot(111)
> canvas = FigureCanvasTkAgg(f, master=root)
> canvas.get_tk_widget().pack(side=Tk.TOP, fill=Tk.BOTH, expand=1)
>
> def displayData(maxCount=20, delayTimeMS=20, count=0):
>     fx = count % len(a)
>     a[fx].sort()
>     ra = []
>     cnt = 0
>     v = a[fx][cnt]
>     for i in range(15*15):
>         if i == v:
>             ra.append(1)
>             cnt += 1
>             if cnt < len(a[fx]):
>                 v = a[fx][cnt]
>
>         else:
>             ra.append(0)
>
>     z = num.reshape(ra, (15,15))
>     sp.pcolor(transpose(z), cmap=cm.gray, norm=normalize())
>     canvas.draw()
>     count += 1
>     if count < maxCount:
>         root.after(delayTimeMS, displayData, maxCount, delayTimeMS,
> count)
>
> displayData(maxCount=100, delayTimeMS=20)
> root.mainloop()
>
>
>
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real 
> users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

[Matplotlib-users] Re: Matplotlib-users digest, Vol 1 #521 - 2 msgs

[Joachim B H. <cj...@st...>]

John Hunter
<jdh...@ac...> writes:

> Perhaps Joachim would be better off using a gtk timer to handle his
> animation.  Basically, matplotlib tries to provide a GUI neutral way
> to do animation (eg anim.py) but if you want to do something
> semi-sophisticated like alter the timing between draws, you should
> use the GUI specific functionality for this, eg gtk timers.

Thanks for the comments and example, John. I'll study it in detail
later.

For the moment I don't really want to do anything more advanced than
anim.py. The sleep() was only there to make the problem obviously
visible, but it was also present without it.

Longer term, I'd like to be able to pass in a 2d array of y values,
and get a window like the one show() presents, only with a "play"
button for animation plus a slider or similar to navigate to a
specific frame. I don't think this should be too hard, but I don't
know python yet so it may take some time.

The context is that I want to visualise time evolving partial
differential equations. (In general 3d complex-valued ones, but that's
a bit too much information to put in one plot.) So I have results from
the solver, for each time step, no real-time plotting or anything like
that.

Cheers, 
Joachim.

[Matplotlib-users] slowdown when using pcolor and TkAgg; help?

[Russell E. O. <ro...@ce...>]

I have appended some code which draws a varying pcolor plot in a Tkinter 
window. The plot shows two 5x5 white rings which shrink to 3x3 and then 
1x3, then cycles through this repeatedly.

Unfortunately, it slows down with each cycle!!!

Are we doing something wrong? Or can we work around the problem somehow?

(By the way, I did not write the code, though I cleaned up some portions 
of it. I realize the data generation can be cleaned up, but it's not 
slowing things down).

-- Russell

P.S. At risk of distracting from the main problem: I am not entirely 
sure I'm commanding a redraw in the correct fashion. I'm fuzzy on the 
difference between show and plot. Using canvas.show() and canvas.plot() 
both seem to work. Just calling plot() doesn't. Any hints on this or 
pointers to documentation would be appreciated.

Also, is pcolor documented anywhere?

----- the code -----

import Tkinter as Tk
from matplotlib.pylab import *
import matplotlib.numerix as num
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg, 
NavigationToolbar2TkAgg

root = Tk.Tk()
a =[ [0,1,2,3,4,5,6,\
90,91,92,93,94,95,96,15,30,45,\
60,75,21,36,51,66,81,\
8,9,10,11,12,13,14,98,99,\
100,101,102,103,104,\
23,38,53,68,83,29,44,59,74,89],\
[16,17,18,19,20,76,77,78,79,80,35,50,65,\
31,46,61,24,25,26,27,28,\
84,85,86,87,88,39,54,69,43,58,73],\
[32,33,34,62,63,64,47,49,40,41,42,\
70,71,72,55,57],\
[48,56] ]

ioff()
f = figure(figsize=(3,3))
sp = f.add_subplot(111)
canvas = FigureCanvasTkAgg(f, master=root)
canvas.get_tk_widget().pack(side=Tk.TOP, fill=Tk.BOTH, expand=1)

def displayData(maxCount=20, delayTimeMS=20, count=0):
    fx = count % len(a)
    a[fx].sort()
    ra = []
    cnt = 0
    v = a[fx][cnt]
    for i in range(15*15):
        if i == v:
            ra.append(1)
            cnt += 1
            if cnt < len(a[fx]):
                v = a[fx][cnt]

        else:
            ra.append(0)

    z = num.reshape(ra, (15,15))
    sp.pcolor(transpose(z), cmap=cm.gray, norm=normalize())
    canvas.draw()
    count += 1
    if count < maxCount:
        root.after(delayTimeMS, displayData, maxCount, delayTimeMS, 
count)

displayData(maxCount=100, delayTimeMS=20)
root.mainloop()

[Matplotlib-users] Re: Parts of plot lagging from previous frame in animation

[Joachim B H. <cj...@st...>]

Joachim B Haga <cj...@st...> 
writes:

>> You may have something there. I tested different backends:
>>                                                       
>> These show the problem:           While these work fine:
>> 
>>   GtkAgg                            WX                  
>>   QtAgg                             WXAgg               
>>   CairoAgg                          GTK                 
>>   GtkCairo (yes, really)            TkAgg               

Forget this list. I just discovered that -d<nonexistentbackend> 
silently defaults to GtkAgg, at least on my install (debian). So the
"problem" list probably boils down to GtkAgg. Maybe you knew that.

Steve Chaplin <ste...@ya...> writes:
> I've just fixed it in CVS, it should appear in the next matplotlib
> version. Or try installing from CVS if you need the fix now, and would
> like to test it.

Excellent! I'll wait; or, if you want me to test, tell me which file and
version it is so I can patch my install with the relevant changes.

Cheers,
Joachim.

[Matplotlib-users] Toggle Logplot

[<an...@tn...>]

Hi

After finally trying out Matplotlib, I am really impressed, and will stick to
it. And hear comes the first "would be nice to have" feature request:

o instead of secret keys for toggling the grid or the log-scaling of the axis,
  it would be nice to have buttons for this.

o toggling logplot on one or both axes would be nice, too, maybe with 'l' and
  'L', or with buttons

  
Thanks, Martin

-- 
Martin Lüthi                 an...@tn...

Re: [Matplotlib-users] colorbar with matshow plot?

[Fernando P. <Fer...@co...>]

John Hunter wrote:
>>>>>>"James" == James Boyle <bo...@ll...> writes:
> 
> 
>     James> How would one get a colorbar() with a matshow plot?  That
>     James> is without altering the matshow code in pylab.,py.
> 
> A good question.  The naive answer is: add the gci._current line to
> the matshow code below the imshow line
> 
>     im = ax.imshow(*args,**kw)
>     gci._current = im
> 
> 
> and then do
> 
> In [1]: matshow(rand(12,12))
> In [2]: colorbar()
> 
> The problem, of course, is that colorbar resizes the current axes,
> which defeats the purpose of matshow.  
> 
> What one needs is a colorbar kwarg to the matshow function that makes
> extra room for the colorbar and still preserves the aspect ratio.
> This would require some arithmetic and extra work.  Any takers?

Well, I've been wanting this (I mentioned it to you in SF in passing) for a 
while.  But I don't know the sizing code well enough to be able to write it up 
quickly.  I'm just not terribly comfortable yet with the overall internal 
design of mpl.  If nobody else does it I probably will, because I need it. 
But it will likely be _weeks_ before I even have a look.

best,

f

Re: [Matplotlib-users] rectangle fill patterns

[John H. <jdh...@ac...>]

>>>>> "john" == john  <joh...@jp...> writes:

    john> Hello, I am trying to set a bar (a patched series of
    john> rectangles) with a fill pattern instead of

    john> just a solid color.  Is there an easy way to do this in
    john> matplotlib?

    john> I am thinking of something like Qt's QBrush which has cross,
    john> vertical, dense, etc. patterns.

There is no support for this currently -- it wouldn't bee too hard to
add for backends that support this kind of thing.  Basically, we need
to specify the API for it, and add support to backends.  I have been
wanting to add gradient fills for patches (eg polygons, rectangles)
and it would be good to do both at once.

I know Ted has mentioned a desire to plot ellipses avoiding the
discretizations caused by manually computing the vertices.  When the
migration to proper path drawing in mpl is complete, this would be
possible with splines or by exposing a backend draw_ellipse method.
For simplicity of design, I'm inclined to the former, but if you have
any suggestions here, let me know.
 
    john> Also, could you please tell me how to plot a rotated
    john> ellipse?

 There is no built-in support for this, but you can do it "the hard
 way" (example below).

I'm in the process of rethinking and refactoring the way mpl does
transformations, so if you have suggestions on how this could be done
better from an interface standpoint, please let me know.  The code
below requires mpl CVS (but will be included in the next release,
possibly today)

    from matplotlib.agg import trans_affine
    from matplotlib.patches import Polygon
    from matplotlib.numerix import sin, cos, pi, arange
    import pylab

    rx = 2
    ry = 1
    angle = arange(0.0, 2*pi, 0.1)
    xs = rx*cos(angle)
    ys = ry*sin(angle)
    a = 2*pi*10./360.  # 10 degrees
    trans = trans_affine(cos(a), sin(a), -sin(a), cos(a), 0.0, 0.0)
    verts = [tuple(trans.transform(x,y)) for x,y in zip(xs,ys)]
    poly = Polygon(verts)

    ax = pylab.subplot(111)
    ax.add_patch(poly)
    ax.autoscale_view()
    pylab.show()

JDH

Re: [Matplotlib-users] colorbar with matshow plot?

[John H. <jdh...@ac...>]

>>>>> "James" == James Boyle <bo...@ll...> writes:

    James> How would one get a colorbar() with a matshow plot?  That
    James> is without altering the matshow code in pylab.,py.

A good question.  The naive answer is: add the gci._current line to
the matshow code below the imshow line

    im = ax.imshow(*args,**kw)
    gci._current = im


and then do

In [1]: matshow(rand(12,12))
In [2]: colorbar()

The problem, of course, is that colorbar resizes the current axes,
which defeats the purpose of matshow.  

What one needs is a colorbar kwarg to the matshow function that makes
extra room for the colorbar and still preserves the aspect ratio.
This would require some arithmetic and extra work.  Any takers?

Perry and I have been discussing some ideas to make this kind of
layout easier.  The problem is that the figure size is picked first,
and the axes are fractions of that.  In addition to the current
functionality, it would be nice to have a figure container that
resized itself, and axes that were in physical sizes.  Eg, something
like


    fig = FigureDynamic()
    ax = AxesSized(5,5)    # inches
    cax = AxesSized(.5,5)  # inches
    fig.hbox.add(ax)
    fig.hbox.add(cax)
    X = rand(12,12)
    ax.imshow(X)  # aspect ratio is correct since axwidth=axheight
    fig.colorbar(cax)

When you add the axes, the figure window would resize itself to
accommodate them, which is the inverse of what happens currently --
the axes physical dimensions resize themselves according to the figure
physical size.  I think both approaches are useful.

JDH

[Matplotlib-users] colorbar with matshow plot?

[James B. <bo...@ll...>]

How would one get a colorbar() with a matshow plot?

That is without altering the matshow code in pylab.,py.

--Jim

[Matplotlib-users] Another upgrade to the OS-X package for matplotlib 0.72.1

[Chris B. <Chr...@no...>]

Hi all,

A user discovered that I had built only Numeric support, and not 
numarray support, into my OS-X matplotlib package. I've fixed that, and 
you can find the new version at:

http://www.pythonmac.org/packages/matplotlib-0.72.1-py2.3-macosx10.3.zip

It supports Numeric and/or numarray, and the Tk, Wx, and Agg back-ends. 
Note that TK and Wx are a bit weird, they both can crash python when you 
close the window. I'm only using the Agg back-end myself at the moment, 
so I'm not going to figure this out, but it would be great if someone did!

-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...

[Matplotlib-users] rectangle fill patterns

[john <joh...@jp...>]

Hello,

 

I am trying to set a bar (a patched series of rectangles) with a fill
pattern instead of 

just a solid color.  Is there an easy way to do this in matplotlib?

I am thinking of something like Qt's QBrush which has cross, vertical,
dense, etc. patterns.

 

Also, could you please tell me how to plot a rotated ellipse?

 

Thanks, 

John Schwartz (JPL, Monte project)

Re: [Matplotlib-users] Still On Grid Problems...

[John H. <jdh...@ac...>]

>>>>> "andrea" == andrea gavana <and...@ti...> writes:

    andrea> Hello John and NG,
    >> I answered this question before at
    >> 
    >> http://sourceforge.net/mailarchive/message.php?msg_id=11115023
    >> 

    andrea> Probably I did not make myself clear.

    >> In a nutshell, do not manually bind 'g' to toggle the grid.
    >> This already happens by default in matplotlib 0.72.

    andrea> I am NOT manually binding the "g" key! I have NO
    andrea> key_press_event on my routines.  Similarly, the examples:

    andrea> - embedding_in_wx.py - embedding_in_wx2.py -
    andrea> wxcursor_demo.py


OK, I see. Sorry for the confusion.  The default keybindings are made
in the backend FigureManager classes.  All of the examples you
referenced embed matplotlib in a wx app.  The ones that use the
FigureManagerWX (eg embedding_in_wx.py) have the default key bindings;
the ones that don't use FigureManagerWX (eg embedding_in_wx2.py), do
not.  If you don't use the manager, you have to do the bindings you
want yourself.  Eg, in embedding_in_wx2.py

        self.axes.plot(t,s)
        self.canvas = FigureCanvas(self, -1, self.figure)

        def toggle_grid(event):
            self.axes.grid()
            self.canvas.draw()
        self.canvas.mpl_connect('key_press_event', toggle_grid)

Hope this helps,

JDH


 

Re: [Matplotlib-users] Still On Grid Problems...

[<and...@ti...>]

Hello John and NG,

>I answered this question before at
>
>  http://sourceforge.net/mailarchive/message.php?msg_id=3D11115023
>

Probably I did not make myself clear.

>In a nutshell, do not manually bind 'g' to toggle the grid.  This
>already happens by default in matplotlib 0.72.  

I am NOT manually binding the "g" key! I have NO key_press_event on my ro=
utines.
Similarly, the examples:

- embedding_in_wx.py
- embedding_in_wx2.py
- wxcursor_demo.py

The is NOT any key_press_event. So there is no manually binding of char
events (in my understanding). Moreover, I can click many times in the WX
window (as you suggested), but the binding of the "g" char is never activ=
ated
on the following demos:

- embedding_in_wx2.py
- wxcursor_demo.py

For the embedding_in_wx.py demo, you are right, John. You have to click
on the WX window in order to get the "g" behavior, that works. 
I don't really know why you need to click on the window to activate the
char binding... probably Robin Dunn may have an answer to this problem.

Thanks to you all.

Andrea.



 

[Matplotlib-users] another question for colorbar

[Humufr <hu...@ya...>]

I have this script:

from pylab import *
from numarray import *

a = arange(10)

scatter(a,a,c=a,s=60,cmap=cm.winter)
colorbar()
show()

but instead to have the brighter point in green I want to have them in 
blue. To obtain it I did a stupid things and I invert the map:

from pylab import *
from numarray import *

a = arange(10)

scatter(a,a,c=-a,s=60,cmap=cm.winter)
colorbar()
show()

but the colormap scale is in good direction green to blue but negative. 
Do you know how can I correct this perhaps it's not possible and the 
anly way to do this is to create a new color map?

Thanks,

N.

[Matplotlib-users] horizontal colorbar pb

[Humufr <hu...@ya...>]

Hi,

so the scripts to see my problem with the colorbar are:

script 1:

from pylab import *
from numarray import *

a = arange(10)

scatter(a,a,c=a,s=60,cmap=cm.winter)
colorbar(orientation='horizontal')
show()


script 2:

from pylab import *
from numarray import *

a = arange(10)

scatter(a,a,c=a,s=60,cmap=cm.winter)
colorbar()
show()


The first script must have (I think) a scale for the colorbar between 0 
and 10 like it is in the vertical mode but it's only a guess and I 
probably misunderstood something else.

Thanks,

    N.

[Matplotlib-users] Could not fire up pylab in Japanese Win XP

[Natsu <nat...@ya...>]

Hi,
After resolving the font_manager.py issue, I could fire up pylab on
Japanese Win 2K. However, on my Win XP Japanese PCs, there seems
some trouble remaining.

The following result was obtained with Enthogut python 2.3.3 on XP
professional Version 2002 SP1, Japanese.
On other PC with XP SP2, Japanese, I got a similar error.
So I'm afraid it is repeatable to any Japanese Win XP.

Maybe, my WinXPs are not fresh and clean.  What do you think?
I'm ready to test the special debugging build which was mentioned in the
list earlier, if the binaries are sent to me.

--
Natsu

>python
Enthought Edition build 1056
Python 2.3.3 (#51, Feb 16 2004, 04:07:52) [MSC v.1200 32 bit (Intel)] on
win32
Type "help", "copyright", "credits" or "license" for more information.
>>> ^Z

>python simple_plot.py --verbose-helpful
matplotlib data path C:\Python23\share\matplotlib
loaded rc file C:\Python23\share\matplotlib\.matplotlibrc
matplotlib version 0.72.1
verbose.level helpful
interactive is False
platform is win32
numerix Numeric 23.3
font search path ['C:\\Python23\\share\\matplotlib']

Then a message box appears telling that python.exe is quitting.
The box has a link and a "debug" button.

The "debug" button tells that
The instruciton "0x1e07460e" refers to memory at "0x00000000".
It could not be "read". (in Japanese)

The link gudes to an other message box which tells that the error
signature is as follows:

AppName: python.exe	 AppVer: 0.0.0.0	 ModName: python23.dll
ModVer: 2.3.3150.1012	 Offset: 0007460e


>python -v
--- snip ---
>>> import pylab
--- snip ---
import matplotlib.lines # precompiled from
C:\Python23\lib\site-packages\matplotlib\lines.pyc
# C:\Python23\lib\site-packages\matplotlib\patches.pyc matches
C:\Python23\lib\site-packages\matplotlib\patches.py
import matplotlib.patches # precompiled from
C:\Python23\lib\site-packages\matplotlib\patches.pyc
# C:\Python23\lib\site-packages\matplotlib\path.pyc matches
C:\Python23\lib\site-packages\matplotlib\path.py
import matplotlib.path # precompiled from
C:\Python23\lib\site-packages\matplotlib\path.pyc
# C:\Python23\lib\site-packages\matplotlib\ticker.pyc matches
C:\Python23\lib\site-packages\matplotlib\ticker.py
import matplotlib.ticker # precompiled from
C:\Python23\lib\site-packages\matplotlib\ticker.pyc
# C:\Python23\lib\site-packages\matplotlib\font_manager.pyc matches
C:\Python23\lib\site-packages\matplotlib\font_manager.py
import matplotlib.font_manager # precompiled from
C:\Python23\lib\site-packages\matplotlib\font_manager.pyc
# C:\Python23\lib\glob.pyc matches C:\Python23\lib\glob.py
import glob # precompiled from C:\Python23\lib\glob.pyc
# C:\Python23\lib\fnmatch.pyc matches C:\Python23\lib\fnmatch.py
import fnmatch # precompiled from C:\Python23\lib\fnmatch.pyc
# C:\Python23\lib\site-packages\matplotlib\afm.pyc matches
C:\Python23\lib\site-packages\matplotlib\afm.py
import matplotlib.afm # precompiled from
C:\Python23\lib\site-packages\matplotlib\afm.pyc
import matplotlib.ft2font # dynamically loaded from
C:\Python23\lib\site-packages\matplotlib\ft2font.pyd
# C:\Python23\lib\posixpath.pyc matches C:\Python23\lib\posixpath.py
import posixpath # precompiled from C:\Python23\lib\posixpath.pyc
import _winreg # dynamically loaded from C:\Python23\DLLs\_winreg.pyd
import cPickle # builtin

Then the message box appears.

Re: [Matplotlib-users] Re: Japanese Windows win32 API / _winreg / font_manager.py bug workaround

[John H. <jdh...@ac...>]

>>>>> "Natsu" == Natsu  <nat...@ya...> writes:

    >> try: key, direc, any = _winreg.EnumValue( local, j) if not
    >> os.path.dirname(direc): direc = os.path.join(directory, direc)
    >> direc = os.path.abspath(direc).lower() if direc[-4:] ==
    >> '.'+fontext: items[direc] = 1 except EnvironmentError: pass

Thanks for the patch -- this will be in the next release.

    Natsu> The idea is taken from a Japanese blog.

    Natsu>   http://somethingred.dip.jp/blog/239

    Natsu> The author of the Blog seems to have some idea to patch
    Natsu> _winreg, also.

    Natsu>   http://somethingred.dip.jp/blog/253 -- Natsu Mizutani

Perhaps you can contact the winreg maintainers and let them know about
this...

Thanks!
JDH

Re: [Matplotlib-users] Still On Grid Problems...

[John H. <jdh...@ac...>]

>>>>> "andrea" == andrea gavana <and...@ti...> writes:

    andrea> - embedding_in_wx.py

    andrea> The "g" key works fine, the grid is shown/hidden when you
    andrea> hit this key.  I'm having the same problem in my
    andrea> application, I can not show the grid for some reason, and
    andrea> it seems to me that these 2 problems are related.

    andrea> Do you have any suggestion/pointer?


I answered this question before at

  http://sourceforge.net/mailarchive/message.php?msg_id=11115023

In a nutshell, do not manually bind 'g' to toggle the grid.  This
already happens by default in matplotlib 0.72.  

In WX* you have to click on the figure before any of the keypress
functionality is activated (does anyone know how to fix this?)


JDH

Re: [Matplotlib-users] Bug in loglog plot, matplotlib 0.71 (Mac)

[John H. <jdh...@ac...>]

>>>>> "Eliot" == Eliot R Smith <es...@in...> writes:

    Eliot> The following test program dies with an error when used
    Eliot> with the arguments shown -- but cutting back the argument
    Eliot> lists by even one element allows it to work OK.  So does
    Eliot> taking out the hold(False).  This is with Mac OS X, Apple
    Eliot> Python 2.3, Matplotlib version 0.71, __revision__

See the release notes for matplotlib-0.72, specifically

 log plot enhancements

  Lots of work making log plots just work. You can toggle log y axes
  with the l ("ell") keypress -- nonpositive data are simply ignored
  and no longer raise. log plots should be a lot faster and more
  robust

Hope this helps!
JDH

Re: [Matplotlib-users] Re: Matplotlib-users digest, Vol 1 #521 - 2 msgs

[John H. <jdh...@ac...>]

>>>>> "Steve" == Steve Chaplin <ste...@ya...> writes:
    Steve> I ran the script with GTKCairo and it looks fine to me.

    Steve> Perhaps you are using GTKAgg which looks like it has the
    Steve> problem. My guess is that its caused by using
    Steve> gtk.idle_add() which results in asynchronous expose_event
    Steve> updates.

I looked at this briefly and am a little confused by something.  The
base class FigureCanvasGTK.expose_event has this in the doc string

        """Expose_event for all GTK backends
        Should not be overridden.

but then the derived class 

  class FigureCanvasGTKAgg(FigureCanvasGTK, FigureCanvasAgg):

does in fact override it.

I inserted a debug print statement into
FigureCanvasGTKAgg.expose_event and it is not called until the loop is
over.  I think this is because the call to sleep is preventing the
idle handler from ever getting engaged.

Perhaps Joachim would be better off using a gtk timer to handle his
animation.  Basically, matplotlib tries to provide a GUI neutral way
to do animation (eg anim.py) but if you want to do something
semi-sophisticated like alter the timing between draws, you should use
the GUI specific functionality for this, eg gtk timers.

Here is an example

# animate increasing terms in fourier expansion of y=x
import pygtk
pygtk.require('2.0')
import gtk
import matplotlib
matplotlib.use('GTKAgg')

from matplotlib.numerix import arange, pi, sin, zeros, Float
import pylab


class Animator:
    
    def __init__(self, samples=1000, max_k=30):
        self.samples = samples
        self.max_k   = max_k
        self.fig = pylab.figure()
        self.ax = self.fig.add_subplot(111)
        x = arange(0.0, 1.0, 1.0/self.samples)

        self.s = zeros((self.max_k,samples), Float)


        for k in range(1,self.max_k):
            self.s[k] = self.s[k-1]+(-1)**(k+1)*sin(pi*x*k)/k


        self.line, = self.ax.plot(x, x*1.3, linewidth=1.0)
        self.ax.grid(True)
        self.k = 1

    def draw(self, *args):
        
        self.line.set_ydata(2./pi*self.s[self.k])
        self.ax.set_title('k = '+str(self.k))
        pylab.draw()
        self.k += 1
        return self.k<self.max_k


anim = Animator()
gtk.timeout_add(100,anim.draw)
pylab.show()

Re: [Matplotlib-users] The plot symbol 'x' fails?

[John H. <jdh...@ac...>]

>>>>> "Shidai" == Shidai Liu <shi...@gm...> writes:

    >>  Greetings all, trivial problem, don't know if it's my machine
    >> problem.
    >> 
    >> <<< plot([1,2,3],'gx')
    >> 
    >> fails to plot the canvas.
    >> 
    >> Py2.3 + Pylab .70

This will be fixed in the next release -- thanks.

JDH

[Matplotlib-users] minor issue with matplotlib 0.72.1 (Mac)

[Eliot R. S. <es...@in...>]

The following test program dies when line style  'bx-' is requested
in the 4th plotset.  It works ok with 'b+-'    (so fixing this is not 
urgent...)

It worked fine with 0.71.

This is with Mac OS X, Apple Python 2.3,  Matplotlib version 0.72.1
--Eliot Smith, Indiana University

#---------------begin test program
from pylab import *

a=[0,20,40,60,80]
b1=[1,2,3,4,5]
b2=[6,5,4,3,2]
b3=[1,3,5,7,9]
b4=[8,6,4,2,1]
b5=[7,6,5,6,7]

plot(a,b1,'bo-', a,b2,'rs:', a,b3,'g^--',a,b4,'bx-',\
    a,b5,'yv-')
# fails with 'bx-' in 4th plotset.  Works ok with 'b+-'
# message is
#Traceback (most recent call last):
#  File "../testmplbug2.py", line 15, in ?
#    savefig('test2.png')
#  File "/Library/Python/2.3/pylab.py", line 763, in savefig
#
#  File "/platlib/matplotlib/figure.py", line 455, in savefig
#  File "/platlib/matplotlib/backends/backend_tkagg.py", line 161, in 
print_figure
#  File "/Library/Python/2.3/matplotlib/backends/backend_agg.py", line 
371, in print_figure
#    self.draw()
#  File "/Library/Python/2.3/matplotlib/backends/backend_agg.py", line 
320, in draw
#    self.figure.draw(self.renderer)
#  File "/platlib/matplotlib/figure.py", line 338, in draw
#  File "/platlib/matplotlib/axes.py", line 1296, in draw
#  File "/platlib/matplotlib/lines.py", line 294, in draw
#  File "/platlib/matplotlib/lines.py", line 1044, in _draw_x
#  File "/Library/Python/2.3/matplotlib/backends/backend_agg.py", line 
156, in draw_line
#    self._renderer.draw_lines(gc, x, y)
#IndexError: Unexpected SeqBase<T> length.

savefig('test2.png')
#--------------------------end test program

Re: [Matplotlib-users] polyfit implementation

[<and...@ti...>]

Hello Jean-Michel,

>Anyway it seems that my polyfit troubles come from computation precision=

>issues. Indeed my polynomial may be of too high order than data actually=

>
>need so that some coefficients are very low (eg. 1e-20). I made some 
>experiments that show the results depend on the use of Numeric/numarray:=

>
>* Numeric
>I must use polyfit(x,y,N) with Float64 for x because it crashes with 
>Float32, but results are very inaccurate
>
>* numarray
>I can use either Float64 or Float32, but only Float32 give good results

>(may someone have an explanation...)
>
>Compared to Matlab, polynomial fitting with Python is not as good in 
>this case. For people who are interested in testing my data I provided

>an attached text file of x and y, the polynomial order is 6.

That does not depend on Numeric, numarray or Matlab. Is your data. If you=

try, in Matlab, to see how polyfit works (type polyfit), you will see wit=
h
a simple trial that your data are bad conditioned. As an example, if x,
y are the 2 rows of your data:

x =3D x(:);
y =3D y(:);

% Construct the Vardermonde Matrix
V(:,n+1) =3D ones(length(x),1);
for j =3D n:-1:1
   V(:,j) =3D x.*V(:,j+1);
end

[Q,R] =3D qr(V,0);
p =3D R\(Q'*y);
Warning: Matrix is close to singular or badly scaled.
         Results may be inaccurate. RCOND =3D 2.575966e-026.

In fact, try to take a look at the condition number of your matrix R:

condest(R)

ans =3D

     3.89823249817041e+025

That's too high. Neither Matlab, Python or whatever software will give yo=
u
a result on which you can rely. Maybe at a first glance Matlab seems to
be more powerful (and, in general, this is the case), but be aware that
you should not trust on results that are affected by so bad conditioning
number/numerical errors.
Try to reduce the number of points (some of them are too close), or try
a non-linear regression (as lsqnonlin), even if you should not need such
a tool in order to do the job.

HTH.

Andrea.

Re: [Matplotlib-users] polyfit implementation

[Jean-Michel P. <jea...@ar...>]

jea...@ar... wrote:
> I'm having some troubles while fitting data with high order polynomials
> (typically order 6). I looked at the polyfit function in
> 'matplotlib/mlab.py' and I must say I don't understand why the
> least-square problem isn't solved using the matrix multiplication
> instead of the * multiplication. Unless I misunderstood something,
> shouldn't the following line:

Obviously I misunderstood something since quick trials gave me:

 >>> array([[0,1],[1,0]]) * array([[2],[3]])
array([[0, 2],
        [3, 0]])

 >>> matrixmultiply(array([[0,1],[1,0]]), array([[2],[3]]))
array([[3],
        [2]])

as the polyfit function works well...

Anyway it seems that my polyfit troubles come from computation precision 
issues. Indeed my polynomial may be of too high order than data actually 
need so that some coefficients are very low (eg. 1e-20). I made some 
experiments that show the results depend on the use of Numeric/numarray:

* Numeric
I must use polyfit(x,y,N) with Float64 for x because it crashes with 
Float32, but results are very inaccurate

* numarray
I can use either Float64 or Float32, but only Float32 give good results 
(may someone have an explanation...)

Compared to Matlab, polynomial fitting with Python is not as good in 
this case. For people who are interested in testing my data I provided 
an attached text file of x and y, the polynomial order is 6.



JM. Philippe

[Matplotlib-users] Re: Japanese Windows win32 API / _winreg / font_manager.py bug workaround

[Natsu <nat...@ya...>]

Hi,
here I propose a minimum patch to fontmanager.py from 0.72.1
which I've confirmed to work on Japanese Win2k.

$ diff font_manager.org.py font_manager.py

113,118c113,121
<                 key, direc, any = _winreg.EnumValue( local, j)
<                 if not os.path.dirname(direc):
<                     direc = os.path.join(directory, direc)
<                 direc = os.path.abspath(direc).lower()
<                 if direc[-4:] == '.'+fontext:
<                     items[direc] = 1
---

>                 try:
>                     key, direc, any = _winreg.EnumValue( local, j)
>                     if not os.path.dirname(direc):
>                         direc = os.path.join(directory, direc)
>                     direc = os.path.abspath(direc).lower()
>                     if direc[-4:] == '.'+fontext:
>                         items[direc] = 1
>                 except EnvironmentError:
>                     pass


The idea is taken from a Japanese blog.

  http://somethingred.dip.jp/blog/239

The author of the Blog seems to have some idea to patch _winreg, also.

  http://somethingred.dip.jp/blog/253
--
Natsu Mizutani