Feature description

JupyterLab 4 includes a new Extension Manager API. The default implementation is a PyPI Extension Manager which exclusively pulls extensions from PyPI based on framework tags and installs them using pip. There is currently no official conda-based implementation of the Extension Manager, although it is very simple in principle.

Only a subset of conda-store capabilities would be needed for such an extension manager, because it would only modify an existing environment (the one in which jupyterlab is running) rather than creating new environments.

Value and/or benefit

• Users would benefit from having an extension manager which installs extension packages using conda rather than pip, by reducing the risk of breaking dependencies by mixing up pip and conda.
• Users who are limited to conda-store (with pip installations limited by system administrator) would gain a way of installing custom jupyterlab extension in a familiar way (using the same interface as in default JupyterLab, but a different backend)

Anything else?

No response