calculating the mean of an orientation with respect to symmetry

[argerlt]

Description of the change

Currently a draft, but this would be an up-to-date solution to #434 and #437
EDIT: complete, ready for review. Description below also updated for clarity.

As originally discussed, mean is calculated for transforms with symmetry using a variation of the 'flip-flop' method:

1) reduce everything with respect to symmetry.
2) find a rough mean (allow weighted averages if desired)
3) convert (mis)orientations to the point closest to the rough mean
4) recalculate the mean, again with weighted averages if desired.

In the process of this simple fix, it turned out ORIX had some inconsistent handling of fundamental zones and reduction. This update makes it match exactly with the original paper previously mentioned as the souce for OrientationRegion.from_symmetry.

Other things fixed along the way:

1. added some unit tests for uncovered lines
2. fixed the error with Orientation.dot_outer discussed in Orientation.dot_outer() scambles axis ordering and can give the wrong answer #673
3. Added a lot of clarifying language
4. Detangled the previous logic in orientation_region.get_point_groups and added the logical extension discussed in the above paper.
5. fixed the loss of proper/improper discussed in Misorientation.reduce() tosses proper/improper information #670

Progress of the PR

• Docstrings for all functions
• Unit tests with pytest for all lines
• Clean code style by running black via pre-commit

For reviewers

• The PR title is short, concise, and will make sense 1 year later.
• New functions are imported in corresponding __init__.py.
• New features, API changes, and deprecations are mentioned in the unreleased
  section in CHANGELOG.rst.
• Contributor(s) are listed correctly in __credits__ in orix/__init__.py and in
  .zenodo.json.

[hakonanes]

This is good, it's the same approach as taken in MTEX (I believe some call it the "flip-flop" algorithm).

[argerlt]

pre-commit.ci autofix

[argerlt]

Almost done, but the new failing test is a real error and I need to run some tests to see if #671 is the correct solution.

[argerlt]

@hakonanes, if you get a chance, check out this updated docstring for Misorientation.reduce().

Nevermind, heavily edited.

[argerlt]

pre-commit.ci autofix