A python scientific oriented container
EDIT:
A more recent release of this idea inside my work is available now here. Also if you are interested in Ipython and Notebooks, you may read a post i wrote on my blog about these topics.
OLD
Worked out reading: http://docs.continuum.io/anaconda/install.html#linux-install
Two versions available:
- full_anaconda: all anaconda packages installed
- light_anaconda: only few selected packages already installed
Note: the light version is also available in the official docker hub. https://hub.docker.com/u/cineca The image has been flattened to limit the overall size, following this guide: http://www.centurylinklabs.com/optimizing-docker-images/?hvid=4wO7Yt Starting from a 1.8 image, the result is less than 1GB (see light/flat.sh script)
From your favourite shell do:
# clone this repository
$ git clone https://github.com/pdonorio/docker_anaconda.git
# go for the light version
$ cd docker_anaconda/light
# build the image if not exists and launch
# this will open a bash shell inside the container
$ ./build_and_run.sh
# from the container, load environment
$ source activate scientific
Now you are able to use python/ipython from the command line
After following all the previous commands do (from /opt dir):
$ bash start_notebook.sh
Then, from your local browser you can access the notebook, at the following address:
if you are using linux: http://127.0.0.1:8888
if you are using boot2docker (windows/mac): http://192.168.59.103:8888