Binary stars do not have correct motions

[emilyhunt]

Although the spatial position of binary stars is treated correctly (see here), this is not currently added to their motions. This could be fixed by:

1. Ensuring that true XYZ velocities are added to the binaries based on their orbit
2. Propagating proper motion uncertainties onto unresolved binaries correctly, such as with astromet
3. Optionally also adding binary motion to radial velocities, which would allow for simulation of the true intrinsic radial velocity dispersion of a cluster (very cool!).

[emilyhunt]

The branch fix-binary-motions now has a fix for part 1 of this issue.

Part 2 is annoying to solve right now as the viewing angle of the transformed binary system is not known (i.e., after it's transformed to the location of the cluster). However, this is required by astromet. This could either be fixed by somehow propagating the rotation matrices (hard) or just by empirically measuring angles between the stars in the binary systems after all transforms have completed (easier).