LAT pointing model and ffp updates

[skhrg]

No description provided.

[mhasself]

I only looked at pointing_model.py for now -- please remove all the reformatting of code you didn't change.

Also please don't hardcode the model parameters. Those should default to zero (and that would ideally correspond to the trivial pointing model).

[mhasself]

Basically I'd like to converge this a bit better with the SAT model and params ... and take an opportunity to decide how we want to think about those (xi, eta) "offsets".

[elle-shaw]

Saianeesh and I are on board with switching to the (-xi, -el) formalism, which would appear in the code with adding tildes in front of the xieta rotations.

I suggest reconstructing the parameters as follows:
A. {az, el, cr}_offset --> enc_offset_{az, el, cr} to match SAT formalism.
q_enc = quat.rotation_lonlat(-1* (az + enc_offset_az), el + enc_offset_el)

B. rx_{xi, eta}_offset --> cr_center_{xi, eta}0 or cr_rot_{xi, eta}0: These will be the (xi, eta) point in the LATR focal plane centered coordinates which appear fixed under corotator rotation. This parameter is most similar to the fp_rot_{xi, eta}0 parameters in the SAT model, I believe. Once the xieta rotation is applied and the corotator roll is applied, this is NOT rotated back to regular LATR focal plane centered coordinates, and we remain in "corotator-centered" coordinates.
q_cr_center = ~quat.rotation_xieta(cr_center_xi0, cr_center_eta0)

C. tel_{xi, eta}_offset --> el_axis_center_{xi, eta}0 or el_axis_rot_{xi, eta}0:
These will be the (xi, eta) point in the CR-centered focal plane coordinates that appear fixed when the elevation is rotated about it's axis, that is nominally supposed to align with the corotator axis.
Once the xieta rotation is applied and the elevation roll is applied, the xieta rotation is NOT undone, and we remain in "El/Tel-centered" coordinates.
q_el_center = ~quat.rotation_xieta(el_center_xi0, el_center_eta0)

D. mir_{xi, eta}_offset --> mir_center_{xi, eta}0 or mir_rot_{xi, eta}0:
These will be the (xi, eta) point in the El/Tel-centered focal plane coordinates that account for any extra shift as a result of the mirror misalignment. I suppose it is also what sops up any final collimation issues.
After this rotation, it is considered that we are in final "boresight-centered" coordinates and then the latlon rotation is applied.
q_mir_center = ~quat.rotation_xieta(mir_center_xi0, mir_center_eta0)

[mhasself]

Exciting times! Just minor comments.

[mhasself]

I don't understand this last description.

[skhrg]

Let me know if the latest update makes more sense. Will try to make diagrams for all these things soon, but image I attach you to M2 and rigidly tilt the M1+M2 system as a unit, you will see one point stay fixed.

In reality its more complicated than that, M1 and M2 tilt independently about their internal origins but that adds up to an effective offset of the system on sky which can be reproduced by a joint tilt about some effective origin.