Feature: Use JupyterLab for multiple notebooks sharing a single kernel

[nulinspiratie]

JupyterLab allows you to connect a notebook to an existing kernel.
This allows you to only load the measurements that you need, instead of one massive Measurements notebook.

One potential issue is that existing widgets might not work any longer

[nulinspiratie]

Tested it, and most of it seems to work nicely. However, at the moment there is one major issue preventing us from using jupyter lab.
The issue is that ipywidgets only work for the original notebook. So any secondary notebook connected to the kernel of the original notebook is not able to display any widgets.

At the moment, this is a dealbreaker because interactive matplotlib uses ipywidgets in jupyterlab

See issue jupyter-widgets/ipywidgets#2628

[nulinspiratie]

I managed to get it working with a custom extension https://github.com/nulinspiratie/shared_kernel
However, it's quite difficult to install at the moment, and it's lacking important features