allow for radial dependence of alignment strength

[duncandc]

code up radial dependence for alignment strength in satellite alignment models.

The hybrid model should also have a radial dependence for the major_to_radial parameter that controls the alignment vector.

I suspect power laws are the first thing to try here.