GaussianProcess.GIBBS

[jschueller]

The name ot.GaussianProcess.GIBBS is confusing, because this is not a traditional Gibbs sampler of a Gaussian vector Y with mean 0 and covariance matrix C. It implements the idea, proposed by Galli and Gao (2001) of using a Gibbs sampler to sample from an auxiliary Gaussian vector X with mean 0 and covariance matrix C^{-1}, and then setting Y = C X. What we do is use a smart implementation of that idea given in Ayano and Emery (2020) (thanks for the ref @regislebrun) - see my post below. So maybe we could rename this ot.GaussianProcess.GALLIGAOGIBBS? The doc should also be updated to explain what the method does, or at least provide a reference.

from #2121

[regislebrun]

@jschueller Good idea, let's adopt ot.GaussianProcess.GALLIGAOGIBBS