Testing_rsbeams_decoherence runs fine for me and mostly looks alright. Though looking back at it I think I wrote this as more of a regression test. I did not document it very well for an example.

In example-NLI-construction it's probably better just to remove the code for NonlinearInsert defined in the notebook. It isn't being used and there is a link to the nonlinear module in the notebook description at the top.

If we really want to show the code in the notebook this code block be replaced with nonlinear.NonlinearInsert??