matplotlib-users Mailing List for matplotlib

Hi,

I have a strange (to me) problem with matplotlib version 0.91.2,
installed from source on a Debian Linux system.

I try to generate a plot which combines bitmap data (using "pcolor")
and some lines (using "plot") as in the appended script.  I switch off
the coordinate axes using the command axis("off") and I try to set the
displayed coordinate range using the command

    axis([x0-p, x1+p, y0-p, y1+p])

near the end of the script.

Problem: the output of this script, i.e. the file "out.eps" is
shifted.  To me it seems that the PostScript bounding box is wrong.

Some things I noticed:

1) It seems that the pixel data from the pcolor command is centred in
the boundign box.  If I plot the pcolor output at different locations,
the bounding box changes.  This seems wrong to me, because I expect
the bounding box to be determined by the above axis command.

2) If I comment out the axis("off") command, the output is no longer
shifted and looks perfect, except that it now has the unwanted
coordinate axes.

3) If I output an .png file, using the Agg backend, the problem does
not appear.  This does not help, since I need an .eps file.

My questions: What am I doing wrong?  How do I get an unshifted .eps
file without coordinate axes?  Any help would be very welcome.

Many thanks in advance,
Jochen

== script starts next line ===========================================
#! /usr/bin/env python

import matplotlib
matplotlib.use("PS")
from pylab import *

def create_hist():
    X = array([-2.3, -2.0, -1.7 ])
    Y = array([-0.3, -0.0, 0.3 ])
    hist = array([[ 0.2, 0.4 ], [ 0.6, 0.8 ]])
    return X, Y, hist

def generate_figure(fname, X, Y, hist, stamp=None):
    width = 4.5
    height = 1.6
    margin = 0.05
    padding = 0.05

    x0 = -3.3
    x1 = 4.0
    q = (x1-x0)/(width - 2*margin - 2*padding)
    y0 = -0.3
    y1 = y0 + (height - 2*margin - 2*padding)*q

    # create an appropriately sized figure
    rc('text', usetex=True)
    rc('font', family="serif", serif="Times", size=12.0)
    rc('xtick', labelsize=12)
    rc('ytick', labelsize=12)
    rc('figure.subplot', left=margin/width)
    rc('figure.subplot', right=(width-margin)/width)
    rc('figure.subplot', bottom=margin/height)
    rc('figure.subplot', top=(height-margin)/height)
    fig = figure(figsize=(width, height))
    ax = axes([margin/width, margin/height,
               (width-2*margin)/width, (height-2*margin)/height])
    p = q*padding

    # plot the density data
    mycmdata = {
        'red': ((0., 1.0, 1.0), (0.01, 0.95, 0.95), (1.0, 0.0, 0.0)),
        'green': ((0., 1.0, 1.0), (0.01, 0.95, 0.95), (1.0, 0.0, 0.0)),
        'blue': ((0., 1.0, 1.0), (0.01, 0.95, 0.95), (1.0, 0.0, 0.0)),
    }
    mycm = matplotlib.colors.LinearSegmentedColormap('mycm',  mycmdata)
    pcolor(X, Y, hist, cmap=mycm, shading='flat')

    # plot the corner
    plot([x0, x0, x1], [y1, y0, y0], "k-")
    plot([x0+0.6, x0+0.6, x1], [y1, y0+0.6, y0+0.6], "k-")

    # place the stamp
    if stamp is not None:
        text(x1-p, y1-p, stamp, va='top', ha='right')

    # save the result
    axis([x0-p, x1+p, y0-p, y1+p])
    axis("off")
    savefig(fname, facecolor="yellow")

X, Y, hist = create_hist()
generate_figure("out.eps", X, Y, hist, "$t=1$")

Hi Eric!

Are you sure you want a colorbar with lines and not a legend?


Well, I've got a couple of lines (~50) and thus the legend would become
confusing.
I wanted to have lines in the colorbar because I think it's more adequate to
have lines in the bar when there are lines in the plot.

A colorbar with lines only, matching a LineCollection, can be done with
>  the facilities in mpl's colorbar.py module; you might take a look at
> the code in that module to see how it is being done in the line contour
> case. Note that ColorbarBase has a method, add_lines, for this. I'm
> sorry I don't have time to be more specific now, though.


Thanks for the hint! I was quite simple, I made a call to
pylab.colorbar(filled=False)
which gave me an empty colorbar. Then I add lines to this instance:
colorbar.ColorbarBase.add_lines(cbar, myrange, linecl.get_colors(),
linecl.get_linewidth())

Thanks! Bernhard



>
> Eric
>
> Bernhard Voigt wrote:
> > Dear all!
> >
> > I'm folowing the line_collection2.py example to create a LineCollection
> > plot with a colorbar.
> > How can I force the colorbar to show distinct lines like it does in
> > contour plots for a LineCollection plot?
> >
> > Thanks! Bernhard
>

   Hi,

I've made a local installation of python2.5, numpy 1.0.4 and matplotlib
0.91.2.

I didn't see any pb during compilation, but when I try to show a figure, I
received SIGSEGV.

You will find attached to this mail the log of my installation, and a
backtrace of a segfault in gdb, I don't know if this helps.


Have you got any glue ?

-- 
LB

Thanks for the reply,

It was not a problem with the specific python I used.

In fact I have a problem because the Tcl/Tk libraries were in /usr/share/lib
whereas their include files were in /usr/include. In my case, I used the
debian package tkx8.3-dev which does not follow that idiom ( I don't know
why..)
As far as I understand, matplotlib asks Python for the path of its Tk
library, and guess the path of the include directory from this path by
adding '../../include/tk' + tk_ver.

I solved my problem by removing this old version of Tk and building a new
one.
But I still think matplotlib should look in the directories given at the top
of setupext.py (or in the setup.cfg file, which would be even better, I
don't like the fact of having to modify directly setupext.py).

Regards,

-- 
LB

Hi,

I'm new here so sorry in advance if this has already been addressed
and I missed it in my archive search...

I think there is a significant bug in plot_wireframe in matplotlib
where it incorrectly displays the Z axis values.  The code below
demonstrates the problem:


import scipy
import pylab as p
import matplotlib.axes3d as p3
from numpy import *

"""
# If you do a wire frame of the following, the graph is correct:
Z = scipy.array(
[[ 0.52,  0.00020],
 [ 0.45,  0.00018],
 [ 0.34,  0.00016]] )
"""

# but if you put negative signs in:
Z = scipy.array(
[[ -0.52,  -0.00020],
 [ -0.45,  -0.00018],
 [ -0.34,  -0.00016]] )

"""
 the graph displays:
[[ -0.62, -0.10020 ],
 [ -0.55, -0.10018 ],
 [ -0.44, -0.10016 ]]
"""

X, Y = meshgrid(arange(0, 3, 1.0), arange(0, 4, 1.0))

fig = p.figure()
ax = p3.Axes3D(fig)
ax.plot_wireframe(X, Y, Z)

ax.set_xlabel('X')
ax.set_ylabel('Y')
ax.set_zlabel('Z')

p.show()


I'm running Ubuntu 7.10 x64 with python 2.5.1-1ubuntu2 and
python-scipy 0.5.2-9ubuntu4 both installed from the .deb files.
I sent the above code to somebody with a 32bit Linux system
and they had the same problem.

Any help appreciated!

Cheers
Shane

You may want to confirm that the python use use to build matplotlib is 
the intended one (in your home).  As a first try, matplotlib's setup 
will import Tkinter and grab the include paths from there.

By this I mean when you type:

    python setup.py install

that "python" is the specific python use want to use.

Also check that Tkinter imports without errors in your custom Python.

Cheers,
Mike

BL wrote:
>     Hi,
> 
> I've got difficulties to make a local installation of matplotlib.
> I think it's due to my nonstandard installation of python and tk, which 
> both are in my home.
> 
> I've seen on the website that , for non standard installation, I should 
> complete the basedir dictionary defined in the setupext.py script.
> 
> But for the Tcl/Tk header files, this does not seems to work. Whatever I 
> put in this dictionary, the module.include_dirs corresponding to Tk 
> (line 825 of setupext.py) is always  equal to  ['/usr/share/include', 
> '/usr/share/include'].
> 
> Did I miss something obvious ?
> 
> I'm on a linux2 platform and I''m trying make a local installation of 
> python2.5 + matplotlib 0.91.2 in my home.
> 
> Regards,
> -- 
> LB
> 
> 
> ------------------------------------------------------------------------
> 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> 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 Guys

Is there an eays way to plot a number vs a label from a file?

Here is an example
I have a datafile like this:
#Aa     Bb      nA      nB      Surf            Bulk            segEnergy
Fe      Cr      19      1       -909.065612     -909.060586     -0.100520
Fe      Ni      19      1       -855.708731     -855.700865     -0.157320
Fe      Mo      19      1       -810.257339     -810.252167     -0.103440

I want to plot the last row vs the second row.

Now i have to do:
y,dummy = load('data.dat',usecols=(6,6),unpack=True)
xticks=xticks(arange(3),('Cr','Ni','Mo'))
plot(y)

Is there an easy way to do this?

And why does usecols=(6) not work?

-- 
Med Venlig Hilsen / Best Regards

Troels Kofoed Jacobsen