Patrick Ingraham says to use 2021-09-08, seq_nums 526-610 for this.
Task is to measure for each image the dx, dy, d_boresight_rotation_angle for the fitted wcs vs the raw one.
For that set: "that set slews to a target, then centers it. So you get several images for a certain object. If you take the first image associated with each object then you'll measure pointing error + whatever else. If you take the last image of each object, you should get only the "whatever else""
So the deliverable here is the seq_num, dx, dy, d_theta table, and Patrick will do the slicing. Code for making that table will live on this branch, but not be merged, so that it's there for when we have a new list dataIds.