Feature request: directional offsets

[aleberti]

Not sure if it is implemented already, if so please ignore :) it would be nice to have a feature to compute directional offsets from one coordinate like in https://docs.astropy.org/en/stable/_modules/astropy/coordinates/sky_coordinate.html#SkyCoord.directional_offset_by (which uses offset_by as defined in https://docs.astropy.org/en/stable/_modules/astropy/coordinates/angles/utils.html#offset_by).

[attipaci]

Hi @aleberti,

This has of course crossed my mind. However, the astropy implementation is a little too simplistic for my taste. Sky coordinates are defined on a sphere, whereas offsets are flat space coordinates. Doing it properly requires a choice of a spherical projection (see e.g. the FITS WCS papers by Calabretta & Greisen 2002). I guess astropy defaults to using the Samson-Flamsteed (SFL) projection, but a proper implementation should allow for other projections also...

WCSLIB by Calabretta provides a base C implementation that SuperNOVAS could build on top, in principle. But that would be a new dependency, which will complicate the build and portability, for sure. I'll look into it, but I also expect that this may be a while before that integration can take place, if at all...

However, you can use WCSLIB along with SuperNOVAS yourself to implement the offset coordinates in your application. Arguably, keeping astrometric position calculations (SuperNOVAS) and spherical projections around a reference coordinate (WCSLIB) separate is the way to go anyways...

PS. I'll convert this from an Issue to a Discussion item, if you don't mind...