Thanks for the clarification but I’m not sure I understand what you’re trying to do: are you trying to fit a GP model with MCMC?
But it sounds like you already have posterior samples. So, are you trying to do posterior retrodictive checks (i.e check whether the data generated by your model are in the realm of the possible compared to your observed data)?
In the latter case, I don’t think specifying a whole PyMC model just to do this (and not also MCMC) is worth it. But tell me if I’m mistaken, I’m not sure I understood your goal here