matplotlib-users — Discussion related to using matplotlib

[Matplotlib-users] Re: matplotlib.mlab.find() not quite like Matlab's

[<jou...@gm...>]

On Sun, 14 Nov 2004 18:08:59 +0200, Jouni K Sepp=E4nen
<jou...@gm...> wrote:
>
> Here's what I'm using to emulate Matlab's find; it only seems to work
> with Numeric, not numarray

Oops - actually it seems that after

from matplotlib.numerix import ravel, indices, compress

the function works whether numerix is using numarray or Numeric.

Making a special case of single-dimensional inputs is not very
elegant, but doing otherwise would break compatibility with the
existing mlab.find; and I guess there is no way in Python of emulating
Matlab's detection of the number of output arguments.

--=20
Jouni K Sepp=E4nen
http://www.iki.fi/jks

[Matplotlib-users] matplotlib.mlab.find() not quite like Matlab's

[<jou...@gm...>]

Hi,

In Matlab, the command

  [i,j] =3D find(mat >=3D 3)

causes i and j to hold the indices where the condition holds. (If
there is only one output argument, it will hold indices to the
flattened version of the condition.) Matplotlib's mlab.find() seems to
work for one-dimensional arrays only.

Here's what I'm using to emulate Matlab's find; it only seems to work
with Numeric, not numarray, and I have no idea whether this is an
efficient way to achieve the goal. I was going to suggest that the
utility be included in matplotlib, but perhaps it should then be
generalized to work with numarray as well. I wonder if anyone has any
ideas on how to do this? I'm a newcomer to matplotlib (and Numeric,
etc.), so please do point out if there is a simpler way to achieve the
effect of Matlab's find.

def find(condition):
    """
    Return the indices where condition is true.
    For arrays of N>=3D2 dimensions, returns a tuple T of N arrays
    such that the condition is true at indices (T[0][i],...,T[N-1][i]).
    """

    sh =3D condition.shape
    if len(sh) =3D=3D 1:
        return nonzero(condition)
    idx =3D indices(sh)
    cond =3D ravel(condition)
    return tuple([compress(cond, ravel(idx[i]))=20
                  for i in range(len(sh))])

--=20
Jouni K Sepp=E4nen
http://www.iki.fi/jks

[Matplotlib-users] Re:

[Dominique O. <Dom...@po...>]

> To: Nils Wagner <nw...@me...>
> Cc: SciPy Users List <sci...@sc...>,
> 	mat...@li...
> Subject: Re: [Matplotlib-users] Visualizing Sparsity Pattern of matrices
> From: John Hunter <jdh...@ac...>
> Date: Thu, 11 Nov 2004 09:16:03 -0600
> 
> 
>>>>>>"Nils" == Nils Wagner <nw...@me...> writes:
> 
> 
>     Nils> Hi all, Structure plots provide a quick visual check on the
>     Nils> sparsity pattern of the matrix.  A structure plot is a
>     Nils> rectangular array of dots; a dot is black if the
>     Nils> corresponding matrix element is nonzero otherwise it is
>     Nils> white.
> 
>     Nils> Is it possible to generate such plots with scipy or should
>     Nils> we switch over to matplotlib ?
> 
> Here's another implementation that uses images - likely to be much
> faster for very large matrices.

Hi,

As part of a programming environment for optimization in Python (soon to 
be released hopefully), I use Matplotlib for the graphics. I created 
several functions imitating Matlab's spy() using scatter(). My sparse 
matrices are represented in linked-list, compressed column or compressed 
row storage using the PySparse implementation
	http://people.web.psi.ch/geus/pyfemax/pysparse.html

Using scatter, I can plot the sparsity pattern of matrices with several 
thousands of lines and columns in a blink on my 1.7GHz P4 laptop. Using 
color maps, you can even color your dots according to the magnitude of 
the element they represent (a 2-dimensional "city plot" of a matrix).

Dominique

Re: [Matplotlib-users] test on hold

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

>>>>> "Dimitri" == Dimitri D'Or <dim...@fs...> writes:

    Dimitri> Hello, I would like to know if there is a mean for
    Dimitri> testing on the hold status of a figure, i.e. to know if
    Dimitri> hold is 'on' or 'off' ?

    Dimitri> In fact, I use frequently the ISHOLD function in Matlab
    Dimitri> and I can't find a similar one in Matplotlib.

A sin of omission.  I just added it to CVS.  If you don't have CVS
access, in the meantime, you can either use the ax._ishold variable to
inspect the axes hold status, or add the following

to matplotlib.axes.Axes:

    def ishold(self):
        'return the HOLD status of the axes'
        return self._hold

and to matplotlib.matlab

    def ishold(b=None):
       """
       Return the hold status of the current axes
       """
       return gca().ishold()    

Should cure what ails ya ...

JDH

Re: [Matplotlib-users] figure() does not create new figure.

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

>>>>> "Shin" == Shin  <sd...@em...> writes:

    Shin> Currently, figure() is the same as figure(1).  But, I think
    Shin> a better behavior of figure is: figure() is the same as
    Shin> figure(n+1) where n is the maximum handle number of opened
    Shin> figures. So, figure() always creates a new figure.  This is
    Shin> exactly MATLAB is doing.  Any comment?

Looks like a bug - thanks.

Is anybody relying on the fact the figure() currently returns the
current figure if it exists?  This is basically the role of gcf,
anyway, which creates a figure none exists and returns it, otherwise
it returns the current figure.

So figure should auto-increment, as you suggest.

JDH

Re: [Matplotlib-users] Any way to turn off interactive mode?

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

>>>>> "Shin" == Shin  <sd...@em...> writes:

    Shin> Thanks. Let me ask another question.  How can I get the
    Shin> current mode?  I like to recover mode after drawing
    Shin> everything in non-interactive mode.  From help document,
    Shin> interactive() seems not to give current mode.

>>> from matplotlib import interactive, is_interactive
>>> print is_interactive()
False

Sorry that these features aren't documented :-(

By the way, ipython already has all of this built-in.  In the pylab
mode you start in interactive mode.  But you can "run" a code snippent
using the run command, which turns off interactive mode for the
duration of the run, restoring the original interactive state when the
run finishes.

JDH

[Matplotlib-users] test on hold

[Dimitri D'O. <dim...@fs...>]

Hello,

I would like to know if there is a mean for testing on the hold status of a 
figure, i.e. to know if hold is 'on' or 'off' ?

In fact, I use frequently the ISHOLD function in Matlab and I can't find a 
similar one in Matplotlib.

Thank you,

Dimitri

[Matplotlib-users] figure() does not create new figure.

[Shin <sd...@em...>]

Currently, figure() is the same as figure(1).
But, I think a better behavior of figure is:
figure() is the same as figure(n+1) where n is the maximum handle number 
of opened figures. So, figure() always creates a new figure.
This is exactly MATLAB is doing.
Any comment?

Daehyok Shin
UNC-CH

Re: [Matplotlib-users] Any way to turn off interactive mode?

[Shin <sd...@em...>]

Thanks. Let me ask another question.
How can I get the current mode?
I like to recover mode after drawing everything in non-interactive mode.
 From help document, interactive() seems not to give current mode.

Daehyok Shin

John Hunter wrote:

> 
> from matplotlib import interactive
> from matplotlib.matlab import *
> 
> plot([1,2,3])
> interactive(False)  # turn off interactive mode
> xlabel('hi mom')
> ylabel('bye')
> title('all done')
> interactive(False)  # turn it back on
> draw()              # draw the canvas
> 
> JDH
> 
> 
> 
> 
> 
> 
> 
>     Shin> ------------------------------------------------------- This
>     Shin> SF.Net email is sponsored by: Sybase ASE Linux Express
>     Shin> Edition - download now for FREE LinuxWorld Reader's Choice
>     Shin> Award Winner for best database on Linux.
>     Shin> http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
>     Shin> _______________________________________________
>     Shin> Matplotlib-users mailing list
>     Shin> Mat...@li...
>     Shin> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 

Re: [Matplotlib-users] Any way to turn off interactive mode?

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

>>>>> "Shin" == Shin  <sd...@em...> writes:

    Shin> My default mode of matplotlib is interactive mode, but in
    Shin> some programs I like to turn off the interactive model
    Shin> temporarily so postpone drawing until I call show(), because
    Shin> of performance concern.  Any way for swith the mode in a
    Shin> script? Thanks in advance.

from matplotlib import interactive
from matplotlib.matlab import *

plot([1,2,3])
interactive(False)  # turn off interactive mode
xlabel('hi mom')
ylabel('bye')
title('all done')
interactive(False)  # turn it back on
draw()              # draw the canvas

JDH







    Shin> ------------------------------------------------------- This
    Shin> SF.Net email is sponsored by: Sybase ASE Linux Express
    Shin> Edition - download now for FREE LinuxWorld Reader's Choice
    Shin> Award Winner for best database on Linux.
    Shin> http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
    Shin> _______________________________________________
    Shin> Matplotlib-users mailing list
    Shin> Mat...@li...
    Shin> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Re: [Matplotlib-users] Visualizing Sparsity Pattern of matrices

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

>>>>> "Nils" == Nils Wagner <nw...@me...> writes:

    Nils> Hi all, Structure plots provide a quick visual check on the
    Nils> sparsity pattern of the matrix.  A structure plot is a
    Nils> rectangular array of dots; a dot is black if the
    Nils> corresponding matrix element is nonzero otherwise it is
    Nils> white.

    Nils> Is it possible to generate such plots with scipy or should
    Nils> we switch over to matplotlib ?

Here's another implementation that uses images - likely to be much
faster for very large matrices.

import random
from matplotlib.colors import LinearSegmentedColormap
from matplotlib.matlab import *


def spy2(Z):
    """
    SPY(Z) plots the sparsity pattern of the matrix S as an image
    """

    #binary colormap min white, max black
    cmapdata = {
         'red'  :  ((0., 1., 1.), (1., 0., 0.)),
         'green':  ((0., 1., 1.), (1., 0., 0.)),
         'blue' :  ((0., 1., 1.), (1., 0., 0.))
         }
    binary =  LinearSegmentedColormap('binary',  cmapdata, 2)

    Z = where(Z>0,1.,0.)
    imshow(transpose(Z), interpolation='nearest', cmap=binary)

def get_sparse_matrix(M,N,frac=0.1):
    'return a MxN sparse matrix with frac elements randomly filled'
    data = zeros((M,N))*0.
    for i in range(int(M*N*frac)):
        x = random.randint(0,M-1)
        y = random.randint(0,N-1)
        data[x,y] = rand()
    return data

data = get_sparse_matrix(50,60)
spy2(data)
show()

Re: [Matplotlib-users] Visualizing Sparsity Pattern of matrices

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

>>>>> "Nils" == Nils Wagner <nw...@me...> writes:

    Nils> Hi all, Structure plots provide a quick visual check on the
    Nils> sparsity pattern of the matrix.  A structure plot is a
    Nils> rectangular array of dots; a dot is black if the
    Nils> corresponding matrix element is nonzero otherwise it is
    Nils> white.

    Nils> Is it possible to generate such plots with scipy or should
    Nils> we switch over to matplotlib ?

A quick matplotlib implementation is below.  In matlab this function
is called "spy" and Alexander Schmolck requested this in an earlier
post.  The spy implementation uses plot markers which are fixed sizes
(in points).  For large matrices, you'll likely want to use a smaller
markersize.

Perhaps better would be to use a polygon collection setup so that the
marker sizes filled the boundaries of the matrix cell.  This would
take a little more work, and would also have a different call
signature that matlab's, since matlab also uses plots markers .  If
you have any thoughts on how you would like the implementation to
work, please share them...

JDH

from matplotlib.matlab import *

def get_xyz_where(Z, Cond):
    """
    Z and Cond are MxN matrices.  Z are data and Cond is a boolean
    matrix where some condition is satisfied.  Return value is x,y,z
    where x and y are the indices into Z and z are the values of Z at
    those indices.  x,y,z are 1D arrays

    This is a lot easier in numarray - is there a more elegant way to
    do this that works on both numeric and numarray?
    """
    
    M,N = Z.shape
    z = ravel(Z)
    ind = nonzero( ravel(Cond) )

    x = arange(M); x.shape = M,1
    X = repeat(x, N, 1)
    x = ravel(X)

    y = arange(N); y.shape = 1,N
    Y = repeat(y, M)
    y = ravel(Y)

    x = take(x, ind)
    y = take(y, ind)
    z = take(z, ind)
    return x,y,z


def spy(Z,  marker='s', markersize=10, **kwargs):
    """
    SPY(Z, **kwrags) plots the sparsity pattern of the matrix S.

    kwargs give the marker properties - see help(plot) for more
    information on marker properties


    """
    x,y,z = get_xyz_where(Z, Z>0)
    plot(x+0.5,y+0.5, linestyle='None', marker=marker,markersize=markersize, **kwargs)

M,N = 25,20
data = zeros((M,N))*0.
data[:,12] = rand(M)
data[5,:] = rand(N)
spy(data)
axis([0,M,0,N])
show()

[Matplotlib-users] Visualizing Sparsity Pattern of matrices

[Nils W. <nw...@me...>]

Hi all,

Structure plots provide a quick visual check on the sparsity pattern of 
the matrix.
A structure plot is a rectangular array of dots;
a dot is black if the corresponding matrix element is nonzero otherwise 
it is white.

Is it possible to generate such plots with scipy or should we switch 
over to matplotlib ?

Nils

Reference:
http://math.nist.gov/MatrixMarket/structureplots.html

 

Re: [Matplotlib-users] [Andy Beardmore <apb@star.le.ac.uk>] matplotlib 0.63

[Jochen V. <vo...@se...>]

Hello,

On Wed, 10 Nov 2004 16:26:47 +0000 (GMT) Andy Baerdmore wrote
> Anyway, the upshot is matplotlib runs but it is not finding the
> sans font and makes a poor substitute in its place. For exampe the output
> from simple_demo.py is : ...
I think that this might be a bug in Vittorio's Debian packages.
There was a problem with the matplotlib font loading code which
made it only find fonts installed under /usr/share/matplotlib.
Since Debian hast the fonts under /usr/share/fonts they were not
found.  Maybe this was not fixed in his packages?

You can try my alternative packages at http://seehuhn.de/debian/,
which hopefully work.  Download python-matplotlib_0.63.4-2.1_i386.deb
=66rom there and install it manually with e.g.

    dpkg -i python-matplotlib_0.63.4-2.1_i386.deb

I hope this helps,
Jochen
--=20
http://seehuhn.de/

[Matplotlib-users] Any way to turn off interactive mode?

[Shin <sd...@em...>]

My default mode of matplotlib is interactive mode,
but in some programs I like to turn off the interactive model temporarily
so postpone drawing until I call show(), because of performance concern.
Any way for swith the mode in a script? Thanks in advance.

Daehyok Shin
UNC-CH

[Matplotlib-users] Grid Labels on Polar Plots

[Carol L. <car...@sr...>]

Is it possible to control the text of the polar grid labels?

I would like to have 8 grid lines in theta, but label only 4 of them.  I 
also want to set my own labels, such as "NE", "NW", "SW" and "SE".

I would like to turn off the labels along r completely or replace them 
with my own set of labels.

The comments in polar_demo.py indicate that the properties of the 
gridlines and labels can be accessed directly from the polar axes.  How 
can I find out what keywords can be set for each of the attributes: 
thetagridlines, rgridlines, thatagridlabels and rgridlabels?
-- 
Ms. Carol A. Leger
SRI International			Phone: (650) 859-4114
333 Ravenswood Avenue G-273
Menlo Park, CA 94025                    e-mail: le...@sr...

Re: [Matplotlib-users] Plotting *lists of data* instead of *functions*...

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

>>>>> "seberino" == seberino  <seb...@sp...> writes:

    seberino> Stephen Thanks for the help.  The format of the data is
    seberino> ASCII files and/or Python arrays that contain (x, y)
    seberino> coordinates or (x, y, z) coordinates for 3D.

    seberino> For starters, what is best way to plot triples in this
    seberino> list?...

    seberino> [ (0, 0, 3.3), (0, 1, 4.4), (1, 0, 2.2), (1, 1, 2.34)]

    seberino> I want to duplicate color plot on screenshots page:
    seberino> http://matplotlib.sourceforge.net/screenshots/pcolor_demo_large.png

You have a 2x2 grid.  pcolor is a bit funny in that for an MxN grid it
only plots M-1 x N-1 rectangles since it doesn't know how to handle
the edges.  So it's a bit of a pathalogical case for pcolor.

Basically you need to transform your list of 3 tuples into 3 arrays
x,y,z, and then reshape the array.  Here I'll use imshow which
doesn't have the edge problem


    from matplotlib.matlab import *
    data = [ (0, 0, 3.3), (0, 1, 4.4), (1, 0, 2.2), (1, 1, 2.34)]
    x,y,z = zip(*data)
    z = array(z); z.shape = 2,2
    imshow(z, interpolation='nearest', extent=(0,1,0,1))
    show()

If you had a longer data sequence, say 5x5, you would use pcolor like

    from matplotlib.matlab import *
    data = ....
    x,y,z = zip(*data)
    x = array(x); x.shape = 5,5
    y = array(y); y.shape = 5,5
    z = array(z); z.shape = 5,5
    pcolor(x,y,z)
    show()


In short, there is nothing special about plotting "data" versus
"functions".  Both are simply cases of plotting 1-D or 2-D arrays as
far as matplotlib is concerned.  Your task is to get your data in to
the right array shape.

JDH

Re: [Matplotlib-users] Plotting *lists of data* instead of *functions*...

[<seb...@sp...>]

Stephen

Thanks for the help.
The format of the data is ASCII files and/or Python arrays
that contain (x, y) coordinates or (x, y, z) coordinates for 3D.

For starters, what is best way to plot triples in this list?...

[ (0, 0, 3.3), (0, 1, 4.4), (1, 0, 2.2), (1, 1, 2.34)]

I want to duplicate color plot on screenshots page:
  http://matplotlib.sourceforge.net/screenshots/pcolor_demo_large.png

Chris



On Wed, Nov 10, 2004 at 12:21:00PM -0800, Stephen Walton wrote:
> On Wed, 2004-11-10 at 10:06 -0800, seb...@sp... wrote:
> > The examples on the screenshots page are great and very helpful.
> > I hope to never use any other plotting app again except Matplotlib.
> >
> > What is easiest/cleanest way to modify examples to accept
> > precalculated /data/ rather than use a mathematical /function/???
>
> You have asked a _huge_ question.  What format are these data in? ASCII
> files?  HDF files?  FITS images?  JPEG images?
>
> --
> Stephen Walton, Professor of Physics and Astronomy,
> California State University, Northridge
> ste...@cs...
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by:
> Sybase ASE Linux Express Edition - download now for FREE
> LinuxWorld Reader's Choice Award Winner for best database on Linux.
> http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

--
_______________________________________

Christian Seberino, Ph.D.
SPAWAR Systems Center San Diego
Code 2872
49258 Mills Street, Room 158
San Diego, CA 92152-5385
U.S.A.

Phone: (619) 553-9973
Fax  : (619) 553-6521
Email: seb...@sp...
_______________________________________

Re: [Matplotlib-users] Plotting *lists of data* instead of *functions*...

[Stephen W. <ste...@cs...>]

On Wed, 2004-11-10 at 10:06 -0800, seb...@sp... wrote:
> The examples on the screenshots page are great and very helpful.
> I hope to never use any other plotting app again except Matplotlib.
> 
> What is easiest/cleanest way to modify examples to accept
> precalculated /data/ rather than use a mathematical /function/???

You have asked a _huge_ question.  What format are these data in? ASCII
files?  HDF files?  FITS images?  JPEG images?

-- 
Stephen Walton, Professor of Physics and Astronomy,
California State University, Northridge
ste...@cs...

Re: [Matplotlib-users] [Andy Beardmore <apb@star.le.ac.uk>] matplotlib 0.63

[Paul B. <ba...@st...>]

Andy Beardmore wrote:

>I finally bit the bullet and upgraded my laptop from mandrake 9.1 to
>Ubuntu 4.01. I have been wanting a debian based distro on it for ages...
>
>In the process I upgraded matplotlib from 0.61.x  to 0.63.4.
>I installed the python-matplotlib deb package linked to from the
>matplotlib website, and finally managed to solve all the dependencies
>in order to get matplotlib to work - there was a numarray
>extension package that needed to be installed in addition to the main
>numarray one.
>
>Anyway, the upshot is matplotlib runs but it is not finding the
>sans font and makes a poor substitute in its place. For exampe the output
>from simple_demo.py is :
>
>examples/simple_plot.py
>/usr/lib/python2.3/site-packages/matplotlib/backends/backend_gtk.py:1077:
>DeprecationWarning:
>  self.append_space()
>/usr/lib/python2.3/site-packages/matplotlib/backends/backend_gtk.py:1077:
>GtkWarning: mixing deprecated and non-deprecated GtkToolbar API is not
>allowed
>  self.append_space()
>/usr/lib/python2.3/site-packages/matplotlib/backends/backend_gtk.py:1080:
>DeprecationWarning:
>  self.append_widget(self.message, '', '')
>/usr/lib/python2.3/site-packages/matplotlib/backends/backend_gtk.py:1080:
>GtkWarning: mixing deprecated and non-deprecated GtkToolbar API is not
>allowed
>  self.append_widget(self.message, '', '')
>Could not match sans-serif, normal, normal.  Returning
>/usr/share/matplotlib/cmmi10.ttf
>Could not match sans-serif, normal, normal.  Returning
>/usr/share/matplotlib/cmmi10.ttf
>Could not match sans-serif, normal, normal.  Returning
>/usr/share/matplotlib/cmmi10.ttf
>
>
>Is there any  way I can force it to use a better looking font ?
>Or even find out what has happened to the sans-serif font?
>  
>
Have you tried setting the TTFPATH environment variable to the paths 
where your other TTF fonts are located?  It appears that the default 
font paths that matplotlib uses are not the ones used by Ubuntu.  It 
might also be useful to print the FontManager.ttfdict attribute to see 
what fonts have actually been found.

 -- Paul

-- 
Paul Barrett, PhD      Space Telescope Science Institute
Phone: 410-338-4475    ESS/Science Software Branch
FAX:   410-338-4767    Baltimore, MD 21218

[Matplotlib-users] Plotting *lists of data* instead of *functions*...

[<seb...@sp...>]

The examples on the screenshots page are great and very helpful.
I hope to never use any other plotting app again except Matplotlib.

What is easiest/cleanest way to modify examples to accept
precalculated /data/ rather than use a mathematical /function/???

e.g.

x   = arange(xmin, xmax, dx)
y   = arange(ymin, ymax, dy)
X,Y = meshgrid(x, y)
Z   = my_function(X, Y)
im  = imshow(Z, extent=(xmin, xmax, ymin, ymax))

How supply "Z" data myself rather than define my_function for Z?

====================

Also,

plot(t, my_function(t))

Same question here for this 2D plot.  (I have the 2D data already calculated.)

thanks!

Chris



--
_______________________________________

Christian Seberino, Ph.D.
SPAWAR Systems Center San Diego
Code 2872
49258 Mills Street, Room 158
San Diego, CA 92152-5385
U.S.A.

Phone: (619) 553-9973
Fax  : (619) 553-6521
Email: seb...@sp...
_______________________________________

Re: [Matplotlib-users] Subplots.

[<na...@te...>]

John Hunter wrote:

> I haven't seen this one before.  Intermittent errors are the hardest
> to track down.  What GUI is PSPad based on.  Is any of the information
> at http://matplotlib.sourceforge.net/faq.html#FREEZE helpful?

PSPad is based on native toolkit, I think. It is not open source,
though it is freeware. Information on the FAQ didn't help, but I
tried running some of the examples and my scripts on every IDE,
and got basically the same results. I think the DOS box message
error is the most helpful of them, but not that much.

Double-clicking from Windows Explorer: Microsoft Visual C++ Runtime
Error - abnormal programa termination

DOS command line:
   Fatal Python Error: PyEval_RestoreThread: NULL state

Interpreter running in a DOS box: shows no error, but gives me
a Microsoft Visual C++ error as above when I quit the DOS box.

IDLE: shows no error, but gives me a Microsoft Visual C++ error
as above.

Pyshell: BSODs.

I noticed, however, that it only happens when I save the figure,
and _never_ when I show them in a windows. And, in every case, the
picture is saved with no problems before the error occurs. I am
using the previous binary version of matplotlib (I'm downloading
the newest version right now and will test as soon as possible),
with Python-2.3.4. I double-checked to see if I was using the
right versions (it could have happen), but I don't think that this
is the problem. If there is any other information that could be
helpful, I can find it here.

> You can use
> axes to make the axes any size you want.  The syntax is

I'm sorry to ask so simple questions. The main reason is that the
matlab interface is so simple that I rarely need to do more than
what it provides. This will help a lot, thanks!

---
José Alexandre Nalon
na...@te...

Re: [Matplotlib-users] Subplots.

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

>>>>> "Jos=E9" =3D=3D Jos=E9 Alexandre Nalon <na...@te...> writes:

    Jos=E9> Hello, Matplotlib has been helping me a lot with my graphic
    Jos=E9> needs. I am still surprised by the looking of the
    Jos=E9> pictures. Many thanks for the great software.

    Jos=E9> I'm having some issues, though. Sometimes I get error
    Jos=E9> messages, usually an error in KERNEL32.DLL on Windows ME,
    Jos=E9> and on Windows only, I don't get this behaviour in
    Jos=E9> Linux. It's not as bad as it may seem, as every script runs
    Jos=E9> completely, the pictures are saved and, besides the message
    Jos=E9> box informing the error, nothing weird happens. I don't know
    Jos=E9> how to reproduce the errors - when I run from the DOS prompt
    Jos=E9> or from IDLE, I get the messages. When I run from my IDE (I
    Jos=E9> use PSPad) I usually don't get error messages, with the same
    Jos=E9> scripts. Any hint to what I can be doing wrong, or how to
    Jos=E9> find out what is happening?

I haven't seen this one before.  Intermittent errors are the hardest
to track down.  What GUI is PSPad based on.  Is any of the information
at http://matplotlib.sourceforge.net/faq.html#FREEZE helpful?

    Jos=E9> Also, I'm in need of some help. I must draw six subplots,
    Jos=E9> one below the other (subplot(6, ...)), but the way things
    Jos=E9> are coming out, the plots are to thin, and, although the
    Jos=E9> picture looks good, I thought that if I could make each
    Jos=E9> subplot a little bigger, that would help a lot. Is there any
    Jos=E9> way this can be done?

subplot is simply a thin wrapper to axes -
http://matplotlib.sf.net/matplotlib.matlab.html#-axes .  You can use
axes to make the axes any size you want.  The syntax is

  # left, bottom, width, height
  ax1 =3D axes([0.125, 0.1, 0.7, .8])

where all values are fractions of the total figure size.  See
http://matplotlib.sourceforge.net/examples/ganged_plots.py for an
example where no space is left between the axes, and xlabels are put
only on the bottom.

    Jos=E9> In other plots, I need to index the subplots (label them
    Jos=E9> '(a)', '(b)', ... for reference in text). I was using xlabel
    Jos=E9> to do that, but when I have more than two subplots, the
    Jos=E9> xlabel is shadowed by the following subplot. Is there any
    Jos=E9> way to make the space between the plots bigger, so the
    Jos=E9> xlabels can be shown, or is there any other (better) way to
    Jos=E9> do that?

You can use the text command to place text anywhere in the figure you
want - http://matplotlib.sourceforge.net/matplotlib.matlab.html#-text.
You can place text in data coordinates

  text(.5, 12, 'hi mom')

in which case the text will "move" visually if you pan and zoom the
axes, or in axes coordinates (0,0) is lower left and 1,1 is upper
right, in which case the text will remain stationary with respect to
changes in the axes limits

  text(0.05, 0.9, 'hi mom', transform=3Dgca().transAxes)

See http://matplotlib.sourceforge.net/examples/alignment_test.py for
lots of examples showing text placement and alignment.

You can also place text outside the axes using the text command
  # to the left and above the axes box
  text(-0.1, 1.05, 'hi mom', transform=3Dgca().transAxes)

Hope this helps,
JDH

Re: [Matplotlib-users] Legend for a single line

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

>>>>> "Dominique" == Dominique Orban <Dom...@po...> writes:

    Dominique> However, if I plot only one line, the legend appears
    Dominique> vertically:

    >>>> plot( [1,2,3], [4,5,6] ) legend( ( 'line1' ), 'lower right' )
    >>>> show()

    Dominique> Finally,

    >>>> plot( [1,2,3], [4,5,6] ) legend( ( 'line1', ), 'lower right'
    >>>> ) show()

    Dominique> (note the comma after 'line1') produces the horizontal
    Dominique> text.

    Dominique> This isn't a big deal, but I am not sure where in
    Dominique> legend.py I should fix that.

This isn't a legend bug exactly, but perhaps legend could detect this
common error and warn you.  Legend expects a sequence of legend
labels.  When you pass it

'my label' or equivalently ('my label') you are simply passing it a
string which is a length 8 sequence of characters.  When you pass it
('my label', ) you are passing a length 1 tuple with a string as the
first element.

Admittedly legend could be smarter, and just "do what you mean".

JDH

Re: [Matplotlib-users] pycxx

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

>>>>> "Greg" == Greg Novak <no...@uc...> writes:

    Greg> I upgraded to Fedora Core 3 and hence spent some time
    Greg> rebuiding all of the Python modules I use frequently...

Makes you wish for a good python package manager ...

    Greg> In order to compile matplotlib, I had to install pycxx as a
    Greg> separate package.  I noticed that there was a cxx directory
    Greg> in the matplotlib tree, but it didn't seem to be using it.


    Greg> I got it to compile and it seems to be working.  This
    Greg> message is just for someone's information to point out the
    Greg> dependency, or the fact that the stuff in the cxx directory
    Greg> doesn't seem to be satisfying it.

You shouldn't need a separate pycxx - all the cxx requirements ship
with matplotlib, and the paths to them are hardwired in the setup.py
file.  Try this ( I know you already got it working but it would be
helpful for me if there is a problem to know the source of it)

 1) get a new copy of matplotlib (perhaps you got an incomplete
    download) -
    http://aleron.dl.sourceforge.net/sourceforge/matplotlib/matplotlib-0.64.tar.gz


 2) unpack it in a new directory so that you have a clean build tree

 3) rm -rf your site-packages/matplotlib tree

 4) python setup.py build.  If the build fails, please capture the
    standard output and standard error from the build process and send
    it to me.


Thanks,
JDH