City of Cape Town's Datascience Docker Images

This repository contains cascading image tags used for Datascience at the City of Cape Town by the Data Science branch.

Overview

The current structure of these image tags (and the repo):

└── base
    └── drivers
        └── python_minimal
            |   ├── jupyter-ide
            |   └── python
            |       └── jupyter-k8s
            └── r
                └── rstudio

Overview of the tags:

• base - based on Ubuntu 22.04. Contains various CLI utilities such as vim, htop, etc.
• drivers - Installs various drivers, include ODBC as well as Selenium.
• python_minimal - Installs Python 3 and a limited selection of python packages (see here).
• python - Installs many Python packages (see here).
• jupyter-k8s - Installs and runs Jupyterlab on top of python
• r - Installs R base, ubuntu system packages for common R packages, renv, pak, rspm and remotes.
• rstudio - Installs Jupyterlab, Rstudio, and VSCode on top of r. Automatically launches to jupyterlab. Use renv to manage R packages on a per-project basis.
• jupyter-ide - Makes various IDEs (PyCharm, VSCode) available on top of python_minimal.

Usage

These images are intended to be used either as part of a Jupyterhub setup, or as the base for airflow Kubernetes operator jobs.

To enter an image and poke around (useful for checking whether a dependency is statisfied), use the docker exec or kubectl exec command to start a bash shell, e.g.

docker exec -it -rm cityofcapetown/datascience:python bash

To launch a jupyter-based web server from jupyter-ide, jupyter-k8s or rstudio enter something like the following:

docker run -it --rm -p 8888:8888 cityofcapetown/datascience:{jupyter-ide,jupyter-k8s,rstudio}