Improve how point rotation works for large dip angles

This is where things start to get weird,

in the SVF plotting,

Also, for SVF calculation, when the horizon line intersects a ray extending from the origin more than once, it breaks things. it might be possible for a discretized segment to assume that a sky view differential can be calculated in parts, i.e. cos2(a) - cos2(b)

Start by locating the SVF plotting and calculation entry points, then reproduce the behavior with large dip angles and a horizon line intersecting a ray more than once. Done means point rotation remains correct and the affected SVF calculation handles the intersecting segments consistently.

Written by the indexing model from the issue text.