JupyterLab¶

Jupyter Startup¶

The container uses conda to manage environments. Jupyter is automatically started from jupyter_env.

All other dependencies for working in jupyter lab are installed in worker_env. You can extend this environment at runtime, or create your own environments using conda or venv.

The default is to choose worker_env as your Kernel after starting JupyterLab and creating a new notebook.

Also have a look at the excellent Jupyter Docs.

Takeaway:

• There is a base environment prepared with the name worker_env. You can select this environment from the list of known kernels (e.g. from the top right corner in a JupyterLab notebook).

• The Jupyter server is installed in a separate environment named jupyter_env.

Creating a custom environment¶

There are many ways to do this. One example is to use venv.

1. Create a new empty environment somewhere from within a notebook cell

!python -m venv ./wikidata_venv

2. Install packages to the newly created venv

./wikidata_venv/bin/python -m pip install qwikidata ipykernel pandas

3. Make the Kernel available in JupyterLab

In order to select the Kernel on the top-right corner in a Jupyter notebook, install (link) ipykernel to the venv.

./wikidata_venv/bin/python -m ipykernel install --user --name=qwikidata

See further information and alternatives under Task Guide: Updating packages and custom envs.