rjw57 / jupyter-container

Containerised scientific Python workstation

This container (based on ipython/scipystack) provides a containerised research environment suitable for scientific use with a particular emphasis on Computer Vision.

In addition to the packages pulled in via scipystack, this container also pulls in a recent version of the Jupyter notebook server and installs OpenCV 3 for both Python 3 and Python 2. The opencv_contrib modules are also installed.

The container exposes a single port which the Jupyter notebook server runs on. Launch the container with an incantation similar to the following:

$ docker run -i -p 8888:8888/tcp rjw57/jupyter

By default the Jupyter notebook runs as a user which is given sudo access inside the container. NOTE THAT HAVING ROOT IN THE CONTAINER SHOULD BE VIEWED AS HAVING ROOT ON THE HOST. The user logon name and id can be configured by setting the USER and USER_UID environment variables. For example, to launch the container for the current user:

$ docker run -i -p 8888:8888/tcp -e USER -e USER_UID=$(id -u) rjw57/jupyter

Some niceties of configuration include:

• A fresh version of OpenCV.
• Versions of pip for Python 2 and Python 3 are available via pip2 and pip3.
• pip is configured so that pip install will install the package into the user's home directory.
• The user's profile is configured to put ~/.local/bin on the path so packages installed via pip which have command-line scripts will work as expected.

Virtual machine using Vagrant

If you have Vagrant installed, the provided Vagrant file can be used to provision a virtual machine running the container.