matplotlib-users Mailing List for matplotlib

Hi,

I am trying to generate postscript output with matplotlib. I installed
matplotlib binaries from pythonmac on two different macs (both dual G5s,
one with panther and the other with out-of-box tiger). 

My test script

import matplotlib
matplotlib.use('Agg')
from pylab import *
plot(arange(5),arange(5),'k-')
title("some title")
xlabel('some unit label')
ylabel('some different unit label')
savefig("filename.ps")

creates a file that seems to be ok but is 4.4 M in size and seems to be
broken somehow (I am not able to open it with any standard ps tools):
<http://www.helsinki.fi/~tjrinne/matplotlib/filename.ps>. On both macs,
savefig("filename.png") works ok, i.e., the matplotlib seems to be
operational but there is something wrong about postscript output. 

On a third mac (panther, everything installed several months ago), the
script happens to be working as expected and the 144K file opens ok. I have
no idea what might be the difference between this fully working matplotlib
installation and my new ones that don't output correct ps.  (The working
mac has an US keyboard while the two macs are international versions with
Finnish keyboards, this should not affect things like this?)

I also tried binaries from <http://sda.iu.edu/projects.html>, same results
(i.e., a big broken file as output).


Any ideas how to track this down or what might be going on? 



Cheers,

Teemu Rinne
Dept Psych
Univ Helsinki

On Fri, 2 Sep 2005, Alan Isaac wrote:> matplotlib's __init__.py is apparently looking
> at os.environ['HOME'], but the default XP setting
> for that is '%USERPROFILE%', which then gets passed
> (*without* expansion) on by get_home even though it's
> expansion is needed.  So import pylab does not work.

Patch to matplotlib/__init__.py is attached (unified diff).
Cheers,
Alan Isaac

		Hello,

I did a mistake with scatter function and I remark a strange things.
The size change even when I pass a bad argument. It's seems that the 
default size change if the size is equal to None or to something stupid. 
It's not important but I prefere to tell you this comportment.

M = 30
x = (rand(M)*scale-scale/2) * .4
y = (rand(M)*scale-scale/2) * .4

scatter (x,y)


M = 30
x = (rand(M)*scale-scale/2) * .4
y = (rand(M)*scale-scale/2) * .4
scatter (x,y,s='s')

N.

>>>>> "Tim" == Tim Leslie <ti...@cs...> writes:

    Tim> Thanks for that, it worked just fine on a standard breezy
    Tim> install. How closely does your repo track CVS? If I see an
    Tim> email saying "fixed in cvs", how long should I expect it to
    Tim> take for a new package to be available? Not that it
    Tim> particularly matters, I'm just curious.

I don't have a policy -- I tend to update the repo on an "as needed"
basis, where as-needed means one of the apps I deploy at work needs
the latest-greatest features or bug-fixes.  I have recently started
naming CVS versions with tags like

  0.84.cvs.2

If you need a feature from CVS and it is not yet in the repo, you can
always email me and I can update.  It takes just a minute or two to
update the repo.

I just did this and ticked the version num to 0.84.cvs.3, which is
current with CVS as of 30 seconds ago.

So feel free to update one more time.

JDH

On Wed, 07 Sep 2005, John Hunter <jdh...@ac...> wrote...

> >>>>> "Tim" == Tim Leslie <ti...@cs...> writes:
> 
>     Noel> Is the debian package for matplotlib broken??  I found the
>     Noel> following error when trying to install the debian package
>     Noel> for matplotlib as per the instructions at the bottom of
>     Noel> http://matplotlib.sourceforge.net/installing.html.
>     >>  What debian variant are you using.  If by chance ubuntu, I
>     >> could point you to my own repository which usually has the
>     >> latest mpl from CVS.
> 
>     Tim> I'd be quite interested in being pointed at these
>     Tim> repositories please John.
> 
> A word of warning.  These quick and dirty packages and they don't do
> all the right things (like gpg authentication) and aren't "debian
> compliant" (sorry too little time).  However, I have used them on a
> number of ubuntu hoary systems (including a custom livecd with all the
> scientific python goodies) and haven't had any trouble.

Thanks for that, it worked just fine on a standard breezy install. How 
closely does your repo track CVS? If I see an email saying "fixed in cvs", 
how long should I expect it to take for a new package to be available? Not 
that it particularly matters, I'm just curious.

Cheers,

Timl

> 
> Included below is my sources.list (universe + multiverse + my repo).  
> 
> > sudo apt-get update
> > sudo apt-get install python-matplotlib-jdh
> 
> I have other packages in there too: scipy CVS (scipy-jdh), ipython CVS
> (ipython-jdh) and so on.
> 
> JDH
> 
> # /etc/apt/sources.list
> #deb cdrom:[Ubuntu 5.04 _Hoary Hedgehog_ - Release i386 (20050407)]/ hoary main restricted
> 
> 
> deb http://archive.ubuntu.com/ubuntu hoary main restricted
> deb-src http://archive.ubuntu.com/ubuntu hoary main restricted
> 
> ## Major bug fix updates produced after the final release of the
> ## distribution.
> deb http://archive.ubuntu.com/ubuntu hoary-updates main restricted
> deb-src http://archive.ubuntu.com/ubuntu hoary-updates main restricted
> 
> ## Uncomment the following two lines to add software from the 'universe'
> ## repository.
> ## N.B. software from this repository is ENTIRELY UNSUPPORTED by the Ubuntu
> ## team, and may not be under a free licence. Please satisfy yourself as to
> ## your rights to use the software. Also, please note that software in
> ## universe WILL NOT receive any review or updates from the Ubuntu security
> ## team.
> deb http://archive.ubuntu.com/ubuntu hoary universe
> deb-src http://archive.ubuntu.com/ubuntu hoary universe
> 
> deb http://security.ubuntu.com/ubuntu hoary-security main restricted
> deb-src http://security.ubuntu.com/ubuntu hoary-security main restricted
> 
> deb http://security.ubuntu.com/ubuntu hoary-security universe
> deb-src http://security.ubuntu.com/ubuntu hoary-security universe
> 
> deb http://archive.ubuntu.com/ubuntu hoary multiverse
> deb-src http://archive.ubuntu.com/ubuntu hoary multiverse
> 
> ## Backports
> deb http://ubuntu-backports.mirrormax.net/ hoary-backports main universe multiverse restricted
> deb http://ubuntu-backports.mirrormax.net/ hoary-extras main universe multiverse restricted
> 
> #deb file:///home/jdhunter/debs binary/
> 
> deb http://peds-pc311.bsd.uchicago.edu binary/
> 
> 
> 
> 
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
`-

>>>>> "Tim" == Tim Leslie <ti...@cs...> writes:

    Noel> Is the debian package for matplotlib broken??  I found the
    Noel> following error when trying to install the debian package
    Noel> for matplotlib as per the instructions at the bottom of
    Noel> http://matplotlib.sourceforge.net/installing.html.
    >>  What debian variant are you using.  If by chance ubuntu, I
    >> could point you to my own repository which usually has the
    >> latest mpl from CVS.

    Tim> I'd be quite interested in being pointed at these
    Tim> repositories please John.

A word of warning.  These quick and dirty packages and they don't do
all the right things (like gpg authentication) and aren't "debian
compliant" (sorry too little time).  However, I have used them on a
number of ubuntu hoary systems (including a custom livecd with all the
scientific python goodies) and haven't had any trouble.

Included below is my sources.list (universe + multiverse + my repo).  

> sudo apt-get update
> sudo apt-get install python-matplotlib-jdh

I have other packages in there too: scipy CVS (scipy-jdh), ipython CVS
(ipython-jdh) and so on.

JDH

# /etc/apt/sources.list
#deb cdrom:[Ubuntu 5.04 _Hoary Hedgehog_ - Release i386 (20050407)]/ hoary main restricted


deb http://archive.ubuntu.com/ubuntu hoary main restricted
deb-src http://archive.ubuntu.com/ubuntu hoary main restricted

## Major bug fix updates produced after the final release of the
## distribution.
deb http://archive.ubuntu.com/ubuntu hoary-updates main restricted
deb-src http://archive.ubuntu.com/ubuntu hoary-updates main restricted

## Uncomment the following two lines to add software from the 'universe'
## repository.
## N.B. software from this repository is ENTIRELY UNSUPPORTED by the Ubuntu
## team, and may not be under a free licence. Please satisfy yourself as to
## your rights to use the software. Also, please note that software in
## universe WILL NOT receive any review or updates from the Ubuntu security
## team.
deb http://archive.ubuntu.com/ubuntu hoary universe
deb-src http://archive.ubuntu.com/ubuntu hoary universe

deb http://security.ubuntu.com/ubuntu hoary-security main restricted
deb-src http://security.ubuntu.com/ubuntu hoary-security main restricted

deb http://security.ubuntu.com/ubuntu hoary-security universe
deb-src http://security.ubuntu.com/ubuntu hoary-security universe

deb http://archive.ubuntu.com/ubuntu hoary multiverse
deb-src http://archive.ubuntu.com/ubuntu hoary multiverse

## Backports
deb http://ubuntu-backports.mirrormax.net/ hoary-backports main universe multiverse restricted
deb http://ubuntu-backports.mirrormax.net/ hoary-extras main universe multiverse restricted

#deb file:///home/jdhunter/debs binary/

deb http://peds-pc311.bsd.uchicago.edu binary/

On Wed, 07 Sep 2005, John Hunter <jdh...@ac...> wrote...

> >>>>> "Noel" == Noel O'Boyle <noe...@ma...> writes:
> 
>     Noel> Is the debian package for matplotlib broken??  I found the
>     Noel> following error when trying to install the debian package
>     Noel> for matplotlib as per the instructions at the bottom of
>     Noel> http://matplotlib.sourceforge.net/installing.html.
> 
> What debian variant are you using.  If by chance ubuntu, I could point
> you to my own repository which usually has the latest mpl from CVS.

I'd be quite interested in being pointed at these repositories please 
John.

Cheers,

Timl

> 
> JDH
> 
> 
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
> 
`-

Debian Sarge. What are the dependencies of the Ubuntu package you have
created?

I've also tried to compile it with Tk support, and it compiles fine, but
it won't import the library as it is looking for gtk (but in setup.py I
told it to use Tk!).

... it seems that I need to compile pygtk myself now.

Regards,
Noel

On Wed, 2005-09-07 at 09:16 -0500, John Hunter wrote:
> >>>>> "Noel" == Noel O'Boyle <noe...@ma...> writes:
> 
>     Noel> Is the debian package for matplotlib broken??  I found the
>     Noel> following error when trying to install the debian package
>     Noel> for matplotlib as per the instructions at the bottom of
>     Noel> http://matplotlib.sourceforge.net/installing.html.
> 
> What debian variant are you using.  If by chance ubuntu, I could point
> you to my own repository which usually has the latest mpl from CVS.
> 
> JDH
> 
> 
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

>>>>> "Noel" == Noel O'Boyle <noe...@ma...> writes:

    Noel> Is the debian package for matplotlib broken??  I found the
    Noel> following error when trying to install the debian package
    Noel> for matplotlib as per the instructions at the bottom of
    Noel> http://matplotlib.sourceforge.net/installing.html.

What debian variant are you using.  If by chance ubuntu, I could point
you to my own repository which usually has the latest mpl from CVS.

JDH

Noel O'Boyle wrote:
> Is the debian package for matplotlib broken??
> 
> I found the following error when trying to install the debian package
> for matplotlib as per the instructions at the bottom of
> http://matplotlib.sourceforge.net/installing.html.

[snipped]

> 
> I then tried apt-get install python2.3-matplotlib:
> "The following packages have unmet dependencies:
>   python2.3-matplotlib: Depends: libcairo1 but it is not installable
> E: Broken packages"
> 
> I then tried apt-get install libcairo1:
> "Package libcairo1 is not available, but is referred to by another
> package.
> This may mean that the package is missing, has been obsoleted, or
> is only available from another source
> E: Package libcairo1 has no installation candidate"

Debian sid libraries are in transition. You can get older version of packages
via http://snapshot.debian.net

> 
> Regards,
> Noel
> 

Regards,
ST
--

Ok, I'll do that. Just give me some time.

Cheers
Christian

On 7 Sep 2005, at 14:54, John Hunter wrote:

>>>>>> "Christian" == Christian Meesters <mee...@un...> writes:
>
>     Christian> Well, the solution IS easy and pretty straightforward,
>     Christian> but I didn't realise that for a while ...
>
> How about a little wiki entry explaining how to plot with masked data,
> including and example, etc...
>
>   http://www.scipy.org/wikis/topical_software/MatplotlibCookbook
>
> Thanks,
> JDH
>

On Wednesday 07 September 2005 05:19 am, Christian Meesters wrote:
> Hi
>
> It's me one more time ...
> Does anybody see what's wrong in my script? I've attached the final
> plot as an .png file for clarity. I expected the plot to show a uniform
> label font, but in the upper insert you see a non-uniform labelling,
> why? And why do I get a bigger '1xe4'? Why do I get it at all?

The 1xe4 is telling you that your yaxis spans 0-90,000, not 0-9. This label is 
called the offset, you can access it by the yaxis' get_offset method, and 
then change the label's font. (The offset should respect the ticklabel 
fontsize, would you please file a bugreport at the sourceforge site? I don't 
have time to work on it right now.)

Alternatively, you can use OldScalarFormatter, which does not create that 
label. Check out newscalarformatter_demo.py in the examples.


-- 

Darren

>>>>> "Christian" == Christian Kristukat <c.k...@ho...> writes:

    Christian> Thanks. I did not understand the meaning of 'extent'.

    Christian> Another thing I could not find which is probably
    Christian> possible as well: How can I display a label for the
    Christian> colorbar (not ticklabel)?


colorbar returns an Axes instance for the colorbar, so you can use any
of the Axes methods described at
http://matplotlib.sf.net/matplotlib.axes.html#Axes

    cax = colorbar()
    cax.set_title('A title')
    cax.set_ylabel('ylabel')
    cax.set_xlabel('xlabel')

or if your prefer "setp" style args

    cax = colorbar()
    setp(cax, title='A title', ylabel='ylabel', xlabel='xlabel')


JDH

>>>>> "Christian" == Christian Meesters <mee...@un...> writes:

    Christian> Well, the solution IS easy and pretty straightforward,
    Christian> but I didn't realise that for a while ...

How about a little wiki entry explaining how to plot with masked data,
including and example, etc...

  http://www.scipy.org/wikis/topical_software/MatplotlibCookbook

Thanks,
JDH

Is the debian package for matplotlib broken??

I found the following error when trying to install the debian package
for matplotlib as per the instructions at the bottom of
http://matplotlib.sourceforge.net/installing.html.

apt-get install python-matplotlib python-matplotlib-doc
"The following packages have unmet dependencies:
  python-matplotlib: Depends: python2.3-matplotlib (= 0.82-1) but it is
not going to be installed
E: Broken packages"

I then tried apt-get install python2.3-matplotlib:
"The following packages have unmet dependencies:
  python2.3-matplotlib: Depends: libcairo1 but it is not installable
E: Broken packages"

I then tried apt-get install libcairo1:
"Package libcairo1 is not available, but is referred to by another
package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source
E: Package libcairo1 has no installation candidate"

Regards,
Noel

Hi

It's me one more time ...
Does anybody see what's wrong in my script? I've attached the final 
plot as an .png file for clarity. I expected the plot to show a uniform 
label font, but in the upper insert you see a non-uniform labelling, 
why? And why do I get a bigger '1xe4'? Why do I get it at all?
I think getting an answer on these question would help me a lot. 
Perhaps a pointer on the tutorial is already sufficient, but I didn't 
answers there. I'm sorry if I overlooked something.

TIA
Christian

Here's the relevant part of my script:

     pylab.title('Beam intensity dependence of the ES slit position', 
fontproperties=font)
     pylab.plot(position_data,intensity_data,'k-')
     pylab.xlabel('shift [}\mu\rm{m] from closure', fontproperties=font)
     pylab.ylabel('intensity [counts/min]', fontproperties=font)
     pylab.xticks(fontproperties=font)
     pylab.yticks(fontproperties=font)
     #plotting peak broadness
     a = pylab.axes([.65,.6,.2,.2])
     pylab.plot(position_data,sigmas,'k-')
     a.set_xlim(position_data[0],position_data[-1])
     pylab.title('Peak broadness', fontproperties=font)
     pylab.xlabel(r'$\rm{shift [}\mu\rm{m] from closure}$', 
fontproperties=font)
     pylab.ylabel(r'$\sigma \rm{ [channels]}$', fontproperties=font)
     for labelx, labely in zip(a.get_xticklabels(),a.get_yticklabels()):
                 labelx.set_fontproperties(font)
                 labely.set_fontproperties(font)
     #plotting peak centers
     b = pylab.axes([.65,.2,.2,.2])
     pylab.plot(position_data,mus,'k-')
     b.set_xlim(position_data[0],position_data[-1])
     pylab.title('Peak center position', fontproperties=font)
     pylab.ylabel(r'$\rm{beam center [channels]}$', fontproperties=font)
     pylab.xlabel(r'$\rm{shift [}\mu\rm{m] from closure}$', 
fontproperties=font)
     for labelx, labely in zip(b.get_xticklabels(),b.get_yticklabels()):
                 labelx.set_fontproperties(font)
                 labely.set_fontproperties(font)

     pylab.savefig("ES_calibration.%s" % plot_format)

Thanks, Arnd & Christian.

Combining your answers and the example on masked plots brought me to 
this solution (pseudo code):

y_values = M.array(y_values)
y_values = M.masked_where(y_values < threshold , y_values) #mask values 
below a certain threshold

pylab.plot(x_values, y_values,'k-')
a.set_xlim(x_values[0], x_values[-1]) #for otherwise the range of 
x_values gets truncated

Well, the solution IS easy and pretty straightforward, but I didn't 
realise that for a while ...

Cheers,
Christian

John Hunter wrote:
>>>>>>"Jo=C3=A3o" =3D=3D Jo=C3=A3o Lu=C3=ADs Silva <rom...@ya...> w=
rites:
>=20
>=20
>     Jo=C3=A3o> Christian Kristukat wrote:
>     >> Hi, is it possible to have an image map (imshow) with real axes
>     >> like those created by 'contour' when supplying a xy meshgrid?
>     >> The pcolor command unfortunately doesn't do interpolation.
>     >> Regards, Christian
>=20
> You can use set the coordinates of the image data using the extent
> kwarg.  Among other things, this lets you plot image data with other
> data (eg line data).  See examples/image_demo2.py. =20

Thanks. I did not understand the meaning of 'extent'.

Another thing I could not find which is probably possible as well:
How can I display a label for the colorbar (not ticklabel)?

Regards, Christian

John Hunter wrote:
>>>>>>"Jo=C3=A3o" =3D=3D Jo=C3=A3o Lu=C3=ADs Silva <rom...@ya...> w=
rites:
>=20
>=20
>     Jo=C3=A3o> Christian Kristukat wrote:
>     >> Hi, is it possible to have an image map (imshow) with real axes
>     >> like those created by 'contour' when supplying a xy meshgrid?
>     >> The pcolor command unfortunately doesn't do interpolation.
>     >> Regards, Christian
>=20
> You can use set the coordinates of the image data using the extent
> kwarg.  Among other things, this lets you plot image data with other
> data (eg line data).  See examples/image_demo2.py. =20

Thanks. I did not understand the meaning of 'extent'.

Another thing I could not find which is probably possible as well:
How can I display a label for the colorbar (not ticklabel)?

Regards, Christian