Bugfix: viscosity outputs

[jcoulter12]

Summary:

In careful testing, two issues with the viscosity output were identified.

• The unit conversion to Pa.s was initially incorrect
• In some places, wavevectors in the viscosity calculation were not properly folded into the WS cell

Tremendous thanks to @MSimoncelli for helping to uncover and debug this issue.

To my knowledge, there are no published results using this output, so we hope the impact of this bug is minimal. However, this is a critical bugfix for anyone hoping to use the viscosity output.

In addition,

• This PR includes a serious overhaul of the viscosity calculation, merging many functions which were previously duplicated for electrons and phonons. Further work should be done on this, but the current refactoring is already a significant structural improvement that will help to avoid future errors.
• Scalar products with the special eigenvectors are now printed by the relaxons solvers, which are important sanity checks.
• The quantities related to the real-space solvers, as in Simoncelli PRX 2020 are now output to file -- though, for now these will remain undocumented, as they are still subject to some testing and should be used cautiously.
• A script in phoebe/scripts/ now exists to help run git actions tests locally for debugging and updating the test reference files.