This is the Python Data Science pack image made for Workspaces. The image runs a Jupyterhub server that has VSCode pre-installed with all of the essential and most used packages. The images also comes pre installed with peak-sdk.
Note: The image doesn't support PyCharm unlike previous Python workspace image versions.
This image uses 3.11.5-slim-bookworm as its base which is maintained by the Docker Community.
Debian GNU/Linux 12 (bookworm)
Linux Kernel 5.10.186-179.751.amzn2.x86_64
Python 3.11.5
cweijan.vscode-database-client2 6.6.3
aws-cli 2.7.4
curl 7.88.1
git 2.39.2
jq 1.6
nano 7.2
vim 9.0
fish 3.6.0
zsh 5.9
R 4.3.1
node 18.17.1
docker 24.0.6
htop 3.2.0
pandoc 2.17.1.1
less 590
latex 3.141592653-2.6-1.40.24
vscode-cli 1.82.2
ipywidgets 8.1.0
jupyter-server-proxy 4.0.0
jupyterhub 4.0.2
jupyterlab-git 0.42.0
jupyterlab-lsp 4.2.0
jupyterlab 3.6.5
jupyterlab_widgets 3.0.8
jupytext 1.15.1
lckr-jupyterlab-variableinspector 3.0.9
mypy-ls 0.5.1
nbconvert 7.8.0
notebook 6.5.5
pyls-black 0.4.7
pyls-flake 80.4.0
pyls-isort 0.2.2
pyls-mypy 0.1.8
python-lsp-black 1.3.0
python-lsp-server[all] 1.4.1
virtualenv 20.24.5
peak-sdk 1.0.0
jupysql 0.10.1
The image supports creating remote tunnels. For simplicity the image comes in handy with some node/bash scripts which can be used to initialise, start, stop, and restart the remote tunnels. More info can be found here.
To build the image locally, run the docker build command and pass in the required build arguments:
docker build . -t workflow-python-ds-pack-2.1.0-base-python-3.11.5
To use the image, select it when configuring the Workspace. If you need to install additional dependencies or add some use case specific environment variables, it can be easily extended.