Docker image for data-science

This is a customization of the docker jupyter/scipy-notebook image with additional packages and configurations.

Running the image

The container can be run with an helper script or manually.

Use helper scripts

Windows 10

Open a PowerShell prompt and use runme.ps1 script.
Get help with: Get-Help .\runme.ps1.

Linux/Mac OS X

In a terminal use the runme.sh script.
Get help with: ./runme.sh -h.

Run without helper scripts

Assuming you want to share the host directory <myhostdirectory>. The image can be run manually with:

$ docker run -ti -p 8888:8888 -v <myhostdirectory>:/home/jovyan/work <imagename>

Then open the webpage.

On Linux/Mac it may be useful to set the user and group ids so that ownership of files between docker container and host is preserved:

$ docker run -ti --user $(id -u):$(id -g) --add-group users \ 
         -p 8888:8888 -v <myhostdirectory>:/home/jovyan/work \ 
        <imagename>

Important Notes:

Differently from the use of the helper scripts, the change to the working directory /home/jovyan/work is not performed automatically.
It is recommended not to share the host directory with a sub-directory of the container home to avoid interference of the host directory with the container environment.

Creating a derived image

TODO

Adding and modifying packages or configurations

A python package can be added modifying the environment.yml file.
A post install script post-install.sh is run after installing the python packages, provide additional configuration/installation steps in this script.

List of installed software

Build and run the image, the list of available packages will be appended to this README file.

andreadotti / jupyter-geneva-ds