Implement the AbstractGPs internal API

[rossviljoen]

Would it make sense to implement the internal API for cov & var and remove the definitions for FiniteBLR?

It would add some currently missing methods like cov(f, x, y)

[willtebbutt]

Hmmmm. Do you have a particular use case for cov(f, x, y) with a BLR? I don't have any super strong objections, I'm more just intriuged as to where it's useful.

[rossviljoen]

I think I came across it when writing tests, so I don't have a practical use case in mind.

The major benefit would be that BLRs could be more easily substituted in downstream things that require a GP and might want cov(f, x, y) for some reason.

[willtebbutt]

Hmm yeah, why no. I can't see what the harm would be -- lets do it.