Greetings, I have tried installing matplotlib from
> source, and also from debian packages, and get the same
It is easy to get the wrong debian package. Did you do this:
Add the following to /etc/apt/sources.list
And then run:
> sudo apt-get update
> sudo apt-get install python-matplotlib python-matplotlib-doc
If you have a successful install, you should be able to do
>>> import matplotlib
>>> print matplotlib.__version__
and get 0.82
>>>> import pylab
> No module named pygtk
> PyGTK version 1.99.16 or greater is required to run the
> GTK Matplotlib backends
Note, matplotlib uses GTKAgg as the default backend, but I if you are
having trouble, I suggest trying the TkAgg, QtAgg or WXAgg backends,
which usew Tkinter, pyqt, and wxpython. I suspect all of these are
available in your debian distribution. See
http://matplotlib.sf.net/matplotlibrc for more information -- this is
where you choose your backend. To make matters a little confusing,
between matplotlib 0.82 and 0.83 we changed the default location for
the rc file.
Make sure you name it accordingly.
> After sucessfully installing glib and atk using
> checkinstall, I tried installing pygtk:
> checking for atk >= 1.0.0... Package atk was not found in
> the pkg-config search path.
I wonder if you also need all the -dev packages, eg gtk2-dev and so
Before you go and install a bunch of gtk stuff from src, you may want
to pursue getting the debian stuff working for you. I know from
experience that it is easy to get a system out of whack if you have
some gtk stuff in /usr, more recent versions in /usr/local, and if you
are not very careful you are soon in a world of pain and considering a
clean reinstall. Especially if you are not an experienced linux user,
as you indicated below.
> Help, I need to get away from windows!
I use ubuntu linux -- if you are just getting started with linux and
haven't fully committed to a distribution I highly recommend it. It's
an easy install, has good hardware support, a very modern desktop, and
so on. It also has a good live/boot CD so you can test it out without
committing to it.
I maintain my own ubuntu debian packages for matplotlib and other
scientific computing packages for use on my own computers -- most of
the stuff I do typically requires having CVS versions of everything so
using the standard debs is not an option for me. If you end up going
with ubuntu, let me know and I'll send you the repository information.