Use radius when mapping back to the sphere

[stes]

Unless I am overlooking something, the spherical space is currently not mapped back to the radius self.r when sampling from any of the existing distributions. I moved the normalization into its own function, and added the rescaling to self.r.

[zimmerrol]

Thanks for fixing this. Do you mind adding a docstring to the normalize function and adapting the von_mises_fisher function as well?