plot_random_samples for GAIAmodel

[j-faria]

For the GAIAmodel, I'm wondering if res.plot_random_samples should produce a plot like this:

(from this paper)

What do you think @ThomasBaycroft ?

[j-faria]

Or something like this from the Octofitter documentation

[ThomasBaycroft]

The top version should be possible, but I think the odds of seeing anything interesting are low, as typical planet signals are much smaller than the paralactic motion, and in cases where there is a big signal it would have a high SNR and so not much spread would be visible in the posteriors. It could maybe be useful for very distant systems?

For the second option, this requires removing the parallax and proper-motion. This means that for each posterior the data would look slightly different. One potential could be to just plot the data from the maximum-likelihood model, then plot the orbits from others with the knowledge that there may be slight discrepancies with the data?