Visibility objects

[samaloney]

Need a high level object to hold visibility data more specifically set/bags of visibility for time/s and energy/ies range. Often want to create images / maps from these and to make a proper map need nearly all the same information as a map.

So the table below tries to separate these into two objects a BasicVisibility and a Visibility with the idea BasicVisibility can be turned into image but not map and Visibility can be tuned into images or full Sunpy Map e.g.

basic = BasicVisibility(...)
img = vis_to_image(basic, shape=101*u.pixel, pixel_size=1*u.arcsec) ✅ 
map = vis_to_map(basic, shape=101*u.pixel, pixel_size=1*u.arcsec) 🚫 

vis = Visibility(...)
img = vis_to_image(vis, shape=101*u.pixel, pixel_size=1*u.arcsec) ✅ 
map = vis_to_map(vis, shape=101*u.pixel, pixel_size=1*u.arcsec) ✅ 

Also currently maps (NDCube) are for a specific time and wavelength often have visibilities from multiple times and energies should these be in one or separate visibility objects? Often the time ranges are very short so could share same WCS info although it not 100% correct

This table lists all of the available properties from Map and some extra ones from visibility so we can try to see what is really needed and how to implement. Also need to take into account existing data e.g. STIX, RHESSI and maybe HXT (can even get this data) for STIX isc == dectector but for RHESSI this isn' the case?

Also what about other visibility specific isc, sigamp, ... are these generic can we have better names

	Basic Visibility	Visibility	Map	Comments
u	✅	✅	❌
v	✅	✅	❌
vis	✅	✅	❌	obsvis SSW
sigamp	✅	✅	❌	better name
phase_center	✅	✅	❌
bottom_left_coord	❌	❌	✅
carrington_latitude	❌	❌	✅
carrington_longitude	❌	❌	✅
center	❌	❌	✅
cmap	❓	❓	✅
coordinate_frame	❌	✅	✅
coordinate_system	❌	✅	✅
data	❓	❓	✅	u, v, obsivs as one object
date	✅	✅	✅	maybe prefer the 3 below
date_average	✅	✅	✅
date_end	✅	✅	✅
date_start	✅	✅	✅
detector	✅	✅	✅	ISC v DET for maps usually singular
dimensions	❓	✅	✅	of the visualities?
dsun	❓	❓	✅
dtype	❓	❓	✅
exposure_time	✅	✅	✅
fits_header	❓	❓	✅
heliographic_latitude	❌	❓	✅
heliographic_longitude	❌	❓	✅
instrument	✅	✅	✅
latex_name	✅	✅	✅
mask	❓	❓	✅	a way to exclude data
measurement	✅	✅	✅
meta	✅	✅	✅
name	✅	✅	✅
ndim	❓	❓	✅
nickname	✅	✅	✅
observatory	✅	✅	✅
observer_coordinate	❌	✅	✅
processing_level	✅	✅	✅
psf	❌	❌	✅
quantity	✅	✅	✅
reference_coordinate	❌	✅	✅	xyoffset in SSW
reference_pixel	❌	✅	✅	defaults to array center in SSW
rotation_matrix	❌	✅	✅
rsun_meters	❌	❓	✅
rsun_obs	❌	❓	✅
scale	❌	❌	✅
spatial_units	❌	❌	✅
timeunit	✅	✅	✅
top_right_coord	❌	❌	✅
uncertainty	❓	❓	✅
unit	✅	✅	✅
wavelength	✅	✅	✅
waveunit	✅	✅	✅
wcs	❌	❌	✅	Vis object has all but cedlt and ndims

I guess there are implicitly sets of visibilities for a given time and energy ranges so I guess a good question is as to what the smallest object could be? For this package a I think a set or bag of visualities can be used to form an image and then thinking of NDCube I guess these could be a Collection of Sequences where each sequences would be one time many energies (as the WSC is the same)?