Minor improvements and changes of orientation/PC refinement before v0.5 release

[hakonanes]

We now have orientation and/or projection/pattern center (PC) refinement in from #387!

There are some things we must improve and some things we can improve before a v0.5 release, which this issue keeps track of:

• Update the pattern matching notebook with refinement after dictionary indexing.
• Don't make a deepcopy of the entire signal and rescale the intensities, rescaling should be done per pattern in the refinement function if possible, or elsewhere. Internal improvements to the EBSD orientation/projection center refinement code #405
• Update the API reference. Internal improvements to the EBSD orientation/projection center refinement code #405
• Perhaps move parts of the refinement module utilities elsewhere and use them other places, e.g. in dictionary generation.
• Provide the user with more information about the refinement parameters, perhaps provide the optimization results from scipy, and other optimization details.
• Simplify tests and shorten test times. Internal improvements to the EBSD orientation/projection center refinement code #405
• Allow to refine multiphase crystal maps #411

The PR number behind each task is the relevant PR where this is done or planned to be done.

[hakonanes]

I've started this work in my fork https://github.com/hakonanes/kikuchipy/tree/internal-improvements-refinement. I hope to keep it to one PR, we'll see if that is doable.

[friedkitteh]

Great! Let me know if I can be of assistance.

[hakonanes]

Will do, thanks @friedkitteh!

[hakonanes]

All except the last point about allowing to refine one phase from a multiphase crystal map (converted into its own issue) has now been implemented. RTD cannot build the latest documentation yet (a fix is on the way in #410), but the refinement examples in the pattern matching notebook can be viewed here: https://kikuchipy--410.org.readthedocs.build/en/410/user_guide/pattern_matching.html.