Pointing workflow

[tterasaki]

This is a PR for the pointing workflow which I suggested last week.

There four new things.

1. sotodlib.coords.mapbased_pointing.py
2. sotodlib.site_pipeline.make_mapbased_pointing
3. sotodlib.site_pipeline.update_pointing
4. sotodlib.site_pipeline.combine_focal_planes

The first one includes a function which make wafer centered single-detector maps and a function to reconstruct the rough pointing result from the map.

The second one is the site_pipeline, which use the sotodlib.coords.mapbased_pointing.py. Note that the main function of it has tune_by_tod argument, which call the third function, sotodlib.site_pipeline.update_pointing, and execute the tod based tuning with the map-based prior.

The third one is the tod-based fitting, which I demonstrated before. It is super quick.
So once we get focal plane, we will only use this function by every single planet observation.
As it outputs a result file by each observation, it will be usefull for the pointing's elevation dependence study.

The forth one is a function to combine multiple focal_plane resultsets and make a focal_plane_combined.hdf and focal_plane_combined.hdf, which is the final product. The finarize_focal_plane will use this resultset and give a feedback to the third function.

The figure below is the product for SATP3 run12, and I have checked the offset problem of map-domain is resolved by the tod-domain tuning.

I am a bit selfish and maybe it is breaking the initial design of site_pipeline. But I do think it is a realistic design for site_pipeline.

[elle-shaw]

I will begin to make changes to this code to include true Az/El locations to correspond with each xi-eta detector hit. Before I do that I would like to close out this PR and start fresh.

[elle-shaw]

Nomenclature notes differing between Tomoki's initial PR and now:

1. sotodlib.coords.mapbased_pointing.py ----------------> sotodlib.coords.brightsrc_pointing
2. sotodlib.site_pipeline.make_mapbased_pointing ------> sotodlib.site_pipeline.get_brightsrc_pointing_step1
3. sotodlib.site_pipeline.update_pointing -----------------> sotodlib.site_pipeline.get_brightsrc_pointing_step2
4. sotodlib.site_pipeline.combine_focal_planes -----------> Not part of this PR

[mhasself]

Thanks for pushing this through. I have some documentation and standardization requests. @msilvafe or @mmccrackan please do sign off on changes to flagman.py and processes.py.

[msilvafe]

Thanks for pushing this through. I have some documentation and standardization requests. @msilvafe or @mmccrackan please do sign off on changes to flagman.py and processes.py.

I'm ok with the changes to flagman but I think the ReduceFlags process could be absorbed into CombineFlags which already handles arbitrary combinations of union and intersection so you just need to add in handling for the new except method.

[elle-shaw]

My latest changes to get_brightsrc_pointing_step2 save the median the az, el, and roll of the SAT boresight from the time when each detector observed the source object. Changes to finalize_focal_plane and `fp_containers' is so the az, el, and roll data is passed correctly into a FocalPlane object so the data can be accessed when fed into the pointing model solving code.

[skhrg]

Only reviewing the focal plane changes for now. Looks good to me, just one minor question.