High-eccentricity targets from JPL-H as MPC_COMET

[Fingel]

Makes passing on elements to SLALIB easier when dealing with high eccentricity targets from JPL horizons. Fixes #1180

[jchate6]

I think we might want to approach this another way.
I think the logic is the same, but it might make sense to move it into the model.
I'm picturing moving this to within save() or somewhere in full_clean() like model.clean() so that we perform this on all ingested moving objects.

The problem is that when we try to request an observation for a high eccentricity object we get the wrong pointings if the scheme is incorrect. Doesn't matter where that came from.

[Fingel]

I think we might want to approach this another way. I think the logic is the same, but it might make sense to move it into the model. I'm picturing moving this to within save() or somewhere in full_clean() like model.clean() so that we perform this on all ingested moving objects.

This would remove the user's ability to make that choice, though. Would this not be surprising behavior (target type changes transparently during save() from something other than the parameters it was passed) for people using observatories that, for example, don't rely on SLALIB weirdness? My gut tells me this should remain explicit instead of introducing magic. But maybe it doesn't matter?

[Fingel]

Or maybe this logic really belongs at the LCO/OCS facility level, if what we are really working around is SLALIB? And then we can otherwise leave targets alone.

[jchate6]

I think we have updated the OCS to do this a little better, but this is also an orbital scheme, not a classification or anything. @talister Do you think there is a problem enforcing this change universally? At what level would you expect this to happen?

[talister]

Or maybe this logic really belongs at the LCO/OCS facility level, if what we are really working around is SLALIB? And then we can otherwise leave targets alone.

It's not "working around SLALIB", it's that the maths fundamentally doesn't work in high eccentricity cases and you need to do something different. This is indicated by the scheme which switches (inside SLALIB or whatever else you would use) to a different method of solving Kepler's equation (background on solving this from Bill Gray/find_orb).
Kepler's equation is one of those deceptively simple things but which actually contains a lot of complexity, diverges into chaos (the actual mathematical definition) in certain regimes (high eccentricity) and that excites mathematicians (but see this) and is still being worked on (arXiv paper from 4 months ago)