Add periodic kernel support

[jr-leary7]

i'm intuiting a bit, but i think the periodic kernel could be appropriate for tissues where there are repeating spatial patterns e.g., the pancreas, colon, etc. instead of well-defined, non-repeating patterns like the brain. the periodic kernel is defined as:

$$ k(x, x^\prime) = \exp\left(-\frac{2\sin^2\left(\frac{\pi||x - x^\prime||}{p}\right)}{\ell^2}\right) $$

where $p$ is the period of the kernel and $\ell$ is the length-scale of the approximate GP. this can be easily pre-computed in R and then passed to Stan.

good values of $p$ depend on the resolution of the spatial dataset i.e., roughly equal to the typical inter-spot distance or perhaps a small multiple of it. ergo, the default should be maybe $50-100$, though the user should really set this parameter since spatial resolution varies wildly between technologies.