matplotlib-users Mailing List for matplotlib

Jouni K. Seppänen skrev:
> 
> In Unix shells ' is the better quoting character because all sorts of
> things have special meaning within " characters... but I changed it to
> use subprocess.Popen instead, so we shouldn't need to worry about shell
> quoting at all.
> 
> Jörgen: Thanks for your report, which I think is the first one from a
> Windows user using usetex with the pdf backend. Can you check that the
> latest version on the trunk works (either by updating from svn or by
> applying the attached patch)? Have you run into any other problems? What
> TeX distribution are you using?
> 
>

Jouni,

I'll try it out tomorrow.

/Jörgen

Michael Droettboom <md...@st...> writes:

> Jörgen Stenarson wrote:
>> I tried to use usetex to generate my pdf figures but I got a crash 
>> when saving the figure, log attached. I traced the crash to 
>> find_tex_file(), apparently ' can not be used to quote filenames in 
>> the windows shell it has to be ". In my patch I just changed it to 
>> always use " I don't know if that works on other systems.
> Someone who knows about usetex should review and apply this patch.

In Unix shells ' is the better quoting character because all sorts of
things have special meaning within " characters... but I changed it to
use subprocess.Popen instead, so we shouldn't need to worry about shell
quoting at all.

Jörgen: Thanks for your report, which I think is the first one from a
Windows user using usetex with the pdf backend. Can you check that the
latest version on the trunk works (either by updating from svn or by
applying the attached patch)? Have you run into any other problems? What
TeX distribution are you using?

-- 
Jouni K. Seppänen
http://www.iki.fi/jks

Dear Jeff & ALL,

Attached is the latest version of my Basemap embedded in wxPython
sample application. I have added a check menu option that allows one
to toggle the overlay of the Blue Marble image on and off the Basemap
figure. Everything works well -- except that the Blue Marble image is
plotted upside down! I could not figure out the cause of this, say,
rather bizarre behaviour. Any hints?

Thanks in advance for any assistance you can provide.

Best regards,

PS This version has another known bug, to be eventually fixed -- if
the user has plotted a point coordinate file, the points are erased if
the Blue Marble overlay is requested because the PlotMap() routine
calls ax.cla() at the start.

-- 
Dr. Mauro J. Cavalcanti
Ecoinformatics Studio
P.O. Box 46521, CEP 20551-970
Rio de Janeiro, RJ, BRASIL
E-mail: mau...@gm...
Web: http://studio.infobio.net
Linux Registered User #473524 * Ubuntu User #22717
"Life is complex. It consists of real and imaginary parts."

Jörgen Stenarson wrote:
> Michael Droettboom skrev:
>>> put the pfm/pfb files it somewhere else and have matplotlib use it?
>> I believe Nimbus Roman is just a clone of Times that is included with 
>> Ghostscript.
>>
>> http://www.tug.dk/FontCatalogue/nimbus/
>>
>> If you have Times or Times New Roman installed, that's probably a 
>> reasonable substitute.
>>
>
> I tried to use usetex to generate my pdf figures but I got a crash 
> when saving the figure, log attached. I traced the crash to 
> find_tex_file(), apparently ' can not be used to quote filenames in 
> the windows shell it has to be ". In my patch I just changed it to 
> always use " I don't know if that works on other systems.
Someone who knows about usetex should review and apply this patch.

But just for clarification -- the font lookup issues we were talking 
about yesterday only apply when usetex is False.  When usetex is True, 
all the font lookup happens with LaTeX, so you can't directly specify a 
font by name.  Any particular reason why you're using usetex over the 
default?

Cheers,
Mike

Michael Droettboom skrev:
>> put the pfm/pfb files it somewhere else and have matplotlib use it?
> I believe Nimbus Roman is just a clone of Times that is included with 
> Ghostscript.
> 
> http://www.tug.dk/FontCatalogue/nimbus/
> 
> If you have Times or Times New Roman installed, that's probably a 
> reasonable substitute.
> 

I tried to use usetex to generate my pdf figures but I got a crash when 
saving the figure, log attached. I traced the crash to find_tex_file(), 
apparently ' can not be used to quote filenames in the windows shell it 
has to be ". In my patch I just changed it to always use " I don't know 
if that works on other systems.

/Jörgen

On Thu, Dec 11, 2008 at 10:19 AM, TP <par...@fr...> wrote:
> Hi,
>
> I use matplotlib 0.91.2.
> When I plot the same Line2D on two subplots, it disappears: execute the
> following script:
>
> #######################
> from pylab import *
> ion()
> f = figure()
> s = f.add_subplot("211")
> curve = matplotlib.lines.Line2D([0,1],[0,1],color='m')
> s.add_line( curve )
> s2 = f.add_subplot("212")
> draw()
> raw_input('press a key to delete a line to second subplot')
> s2.add_line( curve )
> s2.lines[-1].figure.canvas.draw()
> print s2.lines[-1].get_visible()
> raw_input('press a key to quit')
> #######################
>
> What is the reason for this behavior?

The lines get their transforms from the axes they reside in -- you
can't share a line on multiple axes

On Thu, Dec 11, 2008 at 7:56 AM, TP <par...@fr...> wrote:
> TP wrote:
>
>> I have a question about the behavior of "del()" Python built-in.
>
> Ok, del only removes a name from the local namespace.
> I have found an old answer of John, below. It seems that a better solution
> is to use the remove method of a line instance:
>
> http://osdir.com/ml/python.matplotlib.general/2005-05/msg00045.html
>
> Until now, I used "del" which works well when the complete object hierarchy
> is given...

The line is stored in axes.lines list, so simply deleting it will not
work because there is a reference to it in the list.  You need to
remove it from the list, as you inidicate.  See also

http://matplotlib.sourceforge.net/users/artists.html

Hi,

I use matplotlib 0.91.2.
When I plot the same Line2D on two subplots, it disappears: execute the
following script:

#######################
from pylab import *
ion()
f = figure()
s = f.add_subplot("211")
curve = matplotlib.lines.Line2D([0,1],[0,1],color='m')
s.add_line( curve )
s2 = f.add_subplot("212")
draw()
raw_input('press a key to delete a line to second subplot')
s2.add_line( curve )
s2.lines[-1].figure.canvas.draw()
print s2.lines[-1].get_visible()
raw_input('press a key to quit')
#######################

What is the reason for this behavior?

Thanks

Julien

PS: thanks to Friedrich Hagedorn for his correction of my signature.

-- 
python -c "print ''.join([chr(154 - ord(c)) for c in '*9(9&(18%.9&1\
+,\'Z4(55l4('])"

"When a distinguished but elderly scientist states that something is
possible, he is almost certainly right. When he states that something is
impossible, he is very probably wrong." (first law of AC Clarke)

TP wrote:

> I have a question about the behavior of "del()" Python built-in.

Ok, del only removes a name from the local namespace.
I have found an old answer of John, below. It seems that a better solution
is to use the remove method of a line instance:

http://osdir.com/ml/python.matplotlib.general/2005-05/msg00045.html

Until now, I used "del" which works well when the complete object hierarchy
is given...

Thanks

Julien

-- 
python -c "print ''.join([chr(154 - ord(c)) for c in '*9(9&(18%.9&1+,\'Z
(55l4('])"

"When a distinguished but elderly scientist states that something is
possible, he is almost certainly right. When he states that something is
impossible, he is very probably wrong." (first law of AC Clarke)

Hi everybody,

I have a question about the behavior of "del()" Python built-in.
In the following example, when I use del() on the copy of a line, it does
not delete it, whereas with the original line, it works. Why? I do not
understand, because the id is the same for the copy and the original:

#######################
from pylab import *
import Tkinter as Tk
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg
ion()
root = Tk.Tk()
f = Figure( figsize = (8,7) )
veryplot = FigureCanvasTkAgg( f
      , master = root )
veryplot.get_tk_widget().pack( side = Tk.LEFT
      , expand = Tk.YES
      , fill = Tk.BOTH )
b = f.add_subplot( 211 )
t = arange(0.01, 5.0, 0.01)
s1 = sin(2*pi*t)
b.plot( t, s1 )
b.plot( t, s1+1 )
print b.lines
raw_input('press a key to delete a line with a copy')
line_copy = b.lines[-1]
print "original id=", id( b.lines[-1] ), "copy id", id( line_copy )
del( line_copy ) # or b.lines.pop()
b.figure.canvas.draw()
print b.lines
raw_input('press a key to delete a line directly')
print "original id=", id( b.lines[-1] )
del( b.lines[-1] ) # or b.lines.pop()
b.figure.canvas.draw()
print b.lines
raw_input('press a key to quit')
#######################

Thanks in advance,

Julien


-- 
python -c "print ''.join([chr(154 - ord(c)) for c in '*9(9&(18%.9&1+,\'Z
(55l4('])"

"When a distinguished but elderly scientist states that something is
possible, he is almost certainly right. When he states that something is
impossible, he is very probably wrong." (first law of AC Clarke)

Zane Selvans wrote:
> It seems like there ought to be an easy way to associate labels with the 
> various groups of patches generated by a call to hist() that uses a list 
> of arrays, setting label=["a", "list", "of", "strings"] for instance, 
> instead of having to go in and label one patch from each returned list 
> of patches afterward.  Would make legend creation easier anyway.

Good point! Has to be added. Thanks.

mm

> Just a suggestion,
> Zane
> 
> -- 
> Zane Selvans
> Amateur Earthling
> za...@id...
> 303/815-6866
> http://zaneselvans.org
> PGP Key: 55E0815F
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 

Hi,

I noticed that there is a problem rendering the plot here:

http://matplotlib.sourceforge.net/users/screenshots.html#basemap-demo

It might be something to do with the location of the data required by
the plotmap.py demo script (line 14).

Cheers,
Scott

On Wed, Dec 10, 2008 at 5:20 PM, pierre garrigues <
pie...@gm...> wrote:

>
> # create the initial line
> x = np.linspace(-3, 3, 1000)
> line, = plt.plot(x, np.sin(x), animated=True)
>
> # save the clean slate background -- everything but the animated line
> # is drawn and saved in the pixel buffer background
> background = canvas.copy_from_bbox(ax.bbox)
>

It looks to me that you need to force a figure draw *before* doing
copy_from_bbox.  Eg::

    fig.canvas.draw()
    background = canvas.copy_from_bbox(ax.bbox)

The reason you are seeing a difference between ipython "run" and ipython
copy-and-paste is that in "run" mode ipython turns interactive drawing off
(for details see http://matplotlib.sourceforge.net/users/shell.html) so the
draw event is never called, and the cached renderer, which is triggering
your exception, is never set.

This is speculation, as I haven't tested, so please answer back if this
fixes your problem and if so I will update the matplotlib animation recipe
which also suffers from this problem (which is not exposed unless you are
running in interactive mode, which you are).  It is on my list of things to
do to write a proper animation chapter for the user's guide....

JDH

On Wed, Dec 10, 2008 at 5:01 PM, Ryan Wagner <rw...@vn...> wrote:

>  Just upgraded to 98.4. Can reproduce this on two XP32 boxes. Any
> thoughts?
>

This is a bug in the win32 installer, which has gtkagg set as the default
backend.  You can fix this by setting TkAgg in the backend, which is what we
*should* have done in the installer.  You need to change the line

  backend : GTKAgg

to

  backend : TkAgg

in your matplotlibrc config file.  Instructions for editing the rc file can
be found at http://matplotlib.sourceforge.net/users/customizing.html.  You
can also read about "what is a backend?" here -
http://matplotlib.sourceforge.net/faq/installing_faq.html#id1

We will be putting out new installers that fix this problem ASAP, so those
of you who on windows who have not yet installed 98.4 may prefer to wait
until we announce the fixed installers are up.  This problem does not affect
the src tarball or the os x installers.

Sorry for the trouble!
JDH