Development environment with:

• python 3
• R
• octave
• scrapy suite
• scrapinghub suite
• jupyter notebook (aka ipython notebook)
• scipy & sklearn & matplotlib
• neurolab
• TensorFlow (no GPU support*)
• pandas
• geopandas & mplleaflet
• If you have a GPU, follow rules here to install TensorFlow with GPU support instead:

https://www.tensorflow.org/install/install_linux#determine_which_tensorflow_to_install

Install/Run

Build instructions:

docker build -t kalessin/martindev .

Alternatively:

docker pull kalessin/martindev:latest

Run instructions:

docker run -ti kalessin/martindev /bin/bash

Running Jupyter notebook

If want to run jupyter, don't forget to map the port to host machine:

docker run -ti -v <host notebooks folder>:/app/jupyter -p 127.0.0.1:8888:8888 kalessin/martindev /bin/bash

And once logged in container console, run jupyter notebook with following options:

$ jupyter notebook --no-browser --allow-root --ip=0.0.0.0

So you can access from host by pointing browser to 127.0.0.1:8888

Running X applications

If some application running in the container needs access to host X server, first run following commands in host machine running X server:

XSOCK=/tmp/.X11-unix
XAUTH=/tmp/.docker.xauth
touch $XAUTH
xauth nlist :0 | sed -e 's/^..../ffff/' | xauth -f $XAUTH nmerge -

Those steps make available the host X socket and authorization credentials for applications running in container.

Then run a new container as follows:

docker run -ti -v $XSOCK:$XSOCK:rw -v $XAUTH:$XAUTH:rw -e DISPLAY -e XAUTHORITY=${XAUTH} kalessin/martindev /bin/bash

It is not the safest way to grant applications running in container access X server as root, but it is the simplest one. For more elaborated alternatives, check http://wiki.ros.org/docker/Tutorials/GUI