Related to pyvista/pyvista#1257

There might be a way to fail sooner if for example PyVista throws (ValueError?) when the requested backend is not available/does not exist.

Finally, I'll remind that although ipyvtk is fully included in our environment and codebase, it's still experimental. From the README.md:

This is an early prototype of creating a Jupyter interface to VTK. This toolkit is a proof of concept and a more polished tool will be available as ipyvtk in the future.

A simple way to address this is to bump the requirements to pyvista>=0.30.0 altogether.

Probably same treatment for pyvistaqt>=0.4.0

pip-pre is green, it's a good sign :)

This should be updated after #8741 is merged because plotter.iren.FindPokedRenderer is not valid anymore starting from pyvista 0.30.0.

@GuillaumeFavelier can you rebase or merge with upstream/main? Looks like PyVista 0.30 is out so we need this PR:

https://pypi.org/project/pyvista/0.30.0/

Done in c141be4

Next step is to use https://github.com/pyvista/pyvistaqt/zipball/8eb40405b31ef815527a2620bc395cb9c5548377

We also might have to deal with:

DeprecationWarning: Use of the "cond" and "rcond" keywords are deprecated and will be removed in future versions of SciPy. Use "atol" and "rtol" keywords instead

https://github.com/mne-tools/mne-python/runs/2545597094#step:13:4933

@larsoner should I just update test_dipole_fitting_fixed to use rtol in:

C = linalg.pinvh(J, rcond=1e-14)

I am fixing these in #9386

Good to go from your end @GuillaumeFavelier ? If so I'll merge once CIs come back happy other than pip-pre, and I'll double check that those just fail non-viz tests

It's ready for me. I'll stay on the lookout for future adjustements if necessary.