Replace OpenMOC comparisons with RandomRay runs

[pshriwise]

OpenMOC, while nifty, has been giving our notebook CI a lot of trouble. Now that we have an MOC-like solver in OpenMC itself, it's probably time to update these comparison runs to use random ray instead.

The openmoc package is imported in the following notebooks:

• mgxs-part-ii.ipynb
• mgxs-part-iii.ipynb

[pshriwise]

..though I'm realizing now that we may want to review the notebooks as there are correction factors that OpenMOC can use that might be relevant in these examples.

[jtramm]

After a quick look through, the main thing that jumps out to me is the use of transport corrected XS data. It should be pretty trivial to have the random ray solver use transport data instead of total if it's available in the mgxs.h5 file, so should just be a few lines of change we'd need to make. We could also add in Geoff's transport stabilization technique as well, should be easy to apply, and helps prevent stability issues arising from transport XS data.