Intro

The notebook images used by JupyterHub are based on Docker and later converted to Singularity for use on the compute cluster. The base images are inherited by the others to create feature-rich environments. This aim is to get users up and going quickly, while reducing the need for custom kernels and duplication across teams.

References such as jupyter/base-notebook refer to the community Jupyter Docker Stacks. See their documentation for more details about the base images.

For requesting changes, see documentation.

Selecting an Image

The notebook images and their base components are described below. This will give you an idea on what's available and what you can test on the various images.

base-notebook

The base-notebook is exposed on Jupyterhub as the python3-notebook and is used as the platform for all other images. This is where all common tooling is installed.

base-notebook includes

• Everything in jupyter/base-notebook
• Shells (bash, tcsh)
• Editors (vim, emacs)
• Other packages such as git, tmux.
• Slurm and it's required packages

base-cuda-notebook

The base-cuda-notebook isn't exposed via Jupyterhub but is used as the platform for all other NVIDIA GPU-support images. This is where all NVIDIA packages get installed and inherited by the other images.

base-cuda-notebook includes

• Everything in base-notebook
• CUDA runtime
• CUDNN8 runtime
• cudatoolkit - required for CUDA Array Interface

ai-notebook

ai-notebook includes

• Everything in the tensorflow-cuda-notebook
• dask (dask-jobqueue, dask-labextension)
• imutils
• keras
• pandas
• plotly
• widgetsnbextension

all-spark-notebook

all-spark-notebook includes

• Everything in the base-notebook and jupyter/all-spark-notebook
• Notebook from jupyter includes
  
  • Python
  • Scala
  • R
  • Spark

clas12-notebook

clas12-notebook sets up Modulefiles from clas12 and includes

• Everything in the pyroot-notebook
• Awkward Array (awkward)
  
• Metakernel
• TensorFlow
• Uproot
• vectors

datascience-notebook

datascience-notebook includes

• Everything in the base-notebook and jupyter/datascience-notebook
• Notebook from jupyter includes
  
  • Scipy
  • R

epsci-notebook

epsci-notebook includes

• Everything in the base-cuda-notebook, plus jupyter/scipy-notebook
• ipywidgets
• logger
• mlflow
• papermill
• pymc3
• tensorflow-probability

ijava-notebook

ijava-notebook includes

• Everything in the base-notebook
• Packages for openjdk and ijava

python3-notebook

python3-notebook includes

• Everything in the base-notebook

pyroot-notebook

pyroot-notebook includes

• Everything in the base-notebook
• Packages for ROOT, PyROOT
  
• Metakernel
• Uproot

scipy-notebook

scipy-notebook includes

• Everything in the base-notebook and jupyter/scipy-notebook

tensorflow-notebook

tensorflow-notebook includes

• Everything in the base-notebook and jupyter/tensorflow-notebook

tensorflow-cuda-notebook

tensorflow-cuda-notebook includes

• Everything in the base-cuda-notebook
• TensorFlow
• PyTorch (torch, torchaudio, torchvision)
• Matplotlib
• Scikit-learn (sklearn)
• seaborn
• Uproot

theory-notebook

theory-notebook includes

• Everything in the base-cuda-notebook, plus jupyter/scipy-notebook
• Packages for gfortran, lhapdf
• Python modules include, but not limited to
  
  • jupyterlab-latex
  • cupy-cuda111
  • fortran-magic
  • hiplot
  • matplotlib
  • mysql-connector-python
  • tensorflow
  • PyTorch (torch, torchaudio, torchvision)
  • vegas

beta

Used for testing upcoming releases of the theory-notebook.

development-notebook (admin)

Used by the admin to test various notebook releases. Might be sent to specific users to test their requested updates before the notebook is moved into production.