Installing Custom Python Packages In Fabric

Thoughts on the best way to go about managing Langchain, llamaindex and all the massive amount of related dependencies in the Fabric notebook environment? I’m really starting to like the latter with the various flavors and approaches to RAG indexing. Although maybe starting to worry it is becoming a Frankenstein like Langchain already is.

Or best to just pivot to and stick with Semantic Kernel, Cognitive Search, etc with Fabric?

My gut tells me the second…

The configuration for packages has moved into a new "Environments" area where, much like the Environments for Runbooks you can configure a number of settings areas and then direct your notebook to use that Environment