Polygon fix to better handle colinear points

[eepeterson]

Description

This PR fixes issues with coplanar points in the openmc.model.Polygon class. As far as I am aware, QHull cannot be forced to generate specific edges (or follow constraints that would enforce that). As a result, my workaround has been to split edges of the desired polygon in half if the first triangulation does not produce an edge on that segment of the polygon. This produces colinear points that are not included in the scipy.spatial.ConvexHull object when testing for the convexity of a particular grouping of triangles inside the polygon. However, they do show up in the coplanar attribute and therefore we can say that if the number of test points is equal to the number of vertices in the convex hull plus the coplanar points then the grouping of triangles is still convex. This combined with increasing the number of attempts at triangulation I believe will provide a more robust Polygon functionality. I also altered the algorithm for grouping simplices by starting with ones with the smallest neighbor lists since that has anecdotally resulted in smaller numbers of convex regions comprising a polygon. Although I'm sure it's not universal, it's likely better than grabbing a simplex at random.

Fixes #2931 pinging @MicahGale since he raised the issue

Checklist

• I have performed a self-review of my own code
  - [x] I have run clang-format (version 15) on any C++ source files (if applicable)
• I have followed the style guidelines for Python source files (if applicable)
  - [x] I have made corresponding changes to the documentation (if applicable)
• I have added tests that prove my fix is effective or that my feature works (if applicable)

[MicahGale]

Should the examples from #2931, and from slack be added as test cases?

[paulromano]

Changes here look good to me. @eepeterson if it's easy enough to add a test case based on the example that brought this up, that would be great.