Workflow for JupyterLab if it lives in a different environment than the IPython kernel

[raffaelemancuso]

What is the supposed workflow to use if JupyterLab lives in a different environment that the IPython kernel that will use the PerspectiveWidget?

Do I need to install perspective-python in both environments, in order to get the JupyterLab extension in the JupyterLab environment?

For ipywidgets you install different packages in the two environments.

Best

[LarryHu0217]

For separate environments, treat the JupyterLab server and the notebook kernel separately.

With the current perspective-python release:

• Install perspective-python[jupyter] in the kernel environment used by the notebook.
• Install perspective-python in the environment that launches JupyterLab, because the prebuilt Perspective JupyterLab extension must be visible to that server.
• Select the kernel environment from JupyterLab after both environments are configured.

For example:

kernel environment

python -m pip install "perspective-python[jupyter]"

JupyterLab server environment

python -m pip install perspective-python

Then verify with:

python -m pip show perspective-python
jupyter labextension list

Installing the package only in the kernel environment may not make the extension available to a separate JupyterLab server. Matching compatible versions in both environments is recommended.