Inbuilt likelihoods for fitting Gaussian processes

[bobverity]

I have a couple of example on my local machine of likelihoods for fitting 1D and 2D Gaussian processes to data. This is a pretty common problem, and it's also one where the posterior can (in rare situations) end up being multi-modal if there are multiple plausible smoothing levels, which motivates the use of drj. Wondering if we want to bring these likelihoods inside the package so they are available to user, and write vignettes showing how to use them? More generally, wondering if we want a folder full of example likelihood/prior functions available to user?

[pwinskill]

+1 I thought the same thing for our simple examples