[IPython-User] IPython Machine Image: Best packages and extensions for Ubuntu?
Thu Jul 5 20:51:17 CDT 2012
On Thu, Jul 5, 2012 at 6:34 PM, Carl Smith <firstname.lastname@example.org> wrote:
> NotebookCloud currently uses Ubuntu 12.04 with IPython and it's
> dependencies. I'm about to upgrade the AMI to use a new version of
> IPython and want to add a bunch of useful packages in the process.
> Within reason, the footprint isn't that important, there's plenty of
> room, but I do want to keep it tidy and documented, not just chuck
> everything in.
I'd replied privately to you, so here is my take on this for the public record:
sudo apt-get install ipython ipython-notebook python-scipy \
python-matplotlib python-rpy2 \
python-sympy cython python-networkx python-pexpect python-nose \
python-mpi4py python-h5py python-tables \
python-pandas python-sklearn python-scikits.statsmodels \
python-pip python-setuptools python-sphinx python-pygments \
sudo apt-get build-dep python python-scipy python-matplotlib cython
for p in "ipython oct2py scikits-image pandas statsmodels networkx"
sudo pip install -u $p
That should give you a fully updated VM with all the major tools for
data analysis and scientific computing, as well as supporting the main
new extensions (Cython, R, Octave) out of the box, and updates to a
few packages that have made important releases since 12.04 came out.
I think the above would make for a fantastic baseline platform to have
> I'm also happy to bundle IPython notebooks, extensions, magics and so
> on, as well as non-Python stuff that's nice to have.
It might be nice to put in the starting directory a few of the
notebooks we ship in our examples:
I wouldn't put them all, since it will just make things too busy, but
a few key ones would be nice to have for beginners (people can always
delete the ones they don't want, which is persistent as long as they
don't terminate the instance and only stop it).
At first cut, this is what I would leave in from the examples
directory (note the svg/m4v files are necessary so the 'tour' notebook
displays correctly, since it shows them):
If you want things tidier you could trim a few more from there,
ultimately it's your call.
Thanks again for this, it's great work.
More information about the IPython-User