Allow aiapy.calibrate.register to accept submaps

[nabobalis]

In GitLab by @wtbarnes on Jul 9, 2021, 05:09

Currently, register will throw an exception if you try to pass it anything but a full-disk image. This is because of the way register uses the underlying rotate method on sunpy.map.Map. We should figure out how to relax this assumption (maybe it does already work??).

A simple test would be to call register on a full disk map and then do a cutout, and then do the operation in reverse order (i.e. cutout, register) and compare the two. Where are they different? why? If implemented, this should become a test.

This would be a great feature to have as it would mean one can "prep" cutouts from the JSOC. As it stands right now, to do cross-channel alignment, one has to retrieve full-disk images.

[nabobalis]

In GitLab by @wtbarnes on Feb 7, 2023, 08:49

There is also the question of why update_pointing does not accept submaps either. Currently, this is because the reference pixel is computed assuming that the map has (4096,4096) pixels. However, there's no reason that this couldn't be then translated appropriately to the frame of the submap.