Jupyter Notebook Lab Workflow

Setting Up Jupyter Notebook/Lab

Installation

• Install with conda
• Install Jupyter Notebook/Lab Extensions
  • jupyter toc extension
  • jupytext extension
  • jupyterlab vim bindings
  • jupyterlab celltags

Add Access to Conda Environment Kernels

Install nb_conda extension to add conda environment to kernels.

Manually Adding Conda Environments as Kernels

See IPython Kernel Install Documentation. Essentially, the conda environment needs ipykernel installed and to call

source activate <myenv>
python -m ipykernel install --user --name <myenv> --display-name "Python [<myenv>]"

where <myenv> should be replaced with the Conda environment name.

Configuring for Version Control

Use jupytext extension

Connecting across SSH

To access jupyter notebooks on a remote machine follow the instructions here.

Notebook Recipe

Install extensions https://github.com/ipython-contrib/jupyter_contrib_nbextensions

• jupyter-vim bindings

To format the output of the notebook

Working over SSH

With an SSH connection, we can connect to a Jupyter Notebook or Lab running on a remote server. This consists of three steps:

1. SSH to remote and run jupyter with the --no-browser option

ssh <user_name>@<remote>
jupyter lab --no-browser --port=8889

2. Open a SSH tunnel to remote for the specified port number

ssh -N -L localhost:8889:localhost:8889 <user_name>@<remote>

3. Open up localhost:8889 in the browser of your choice. Or copy the URL in the terminal from step 1.

Other Notes

Develop Notebook in notebook/develop folder with file names of the form [Date]-[2-4 word description].

Final Notebooks in notebook/publish folder for final analysis