Acq anom warnings don't use dither adjusted position to check for out-of-box case.

[jeanconn]

This code

https://github.com/sot/mica/blob/master/mica/stats/acq_stats.py#L465

Doesn't use the OBC attitude when determining the offset from the expected star position.

I think using 'dy' and 'dz' from the acq stats database would be more reliable.

As written, it is very close, but if we used 'dy', 'dz' we wouldn't see the "non-standard" warning for 18302 as slot 5 would be within 160 + 16 in Y (it is at -175 from expected position with dy, but -195 using the yang-yang_obs test).

[jeanconn]

Note that we still want dy and dz, not cdy/cdz. In this case the cdy/cdz have back-propagated one-shot corrections, and those aren't needed just to see if the star was really in the box that was commanded.