Need uncertainties in kbmod output (including Parsl workflow incorporation)

[ColinOrionChandler]

For orbit fitting and linking purposes especially we need to output uncertainties on position, velocity, and (if available) time.

[ColinOrionChandler]

There's a script in kbmod-wf, here.

[ColinOrionChandler]

We may want to consider carrying this out adjacent to the search phase since it will have the WorkUnit in memory and the GPU populated for that matter. Connected to Issue #834 .

[wilsonbb]

I spent about an hour putting together a draft PR of a simple workflow for re-running an existing resampled WorkUnit and generating uncertainties for its results based on Dino's work in DEEP: dirac-institute/kbmod-wf#53

I can flesh out the workflow for full processing later, but seems nice to have around for these sorts of one-off use cases (which will surely happen again).

It sidesteps the issue of #847 for now, but should be enough to generate some uncertainties for the existing results so that those will be ready for orbitfit and layup.