Beef up GaussianProcessSurrogate

[AdrianSosic]

Rough agenda:

#743 (Cleanup)

• Clean up GP code

#746 (GP Components)

• Extract current GP setting into preset
• Add EDBO / SMOOTHED_EDBO preset
• Generalize kernel machinery into component machinery using generics
• Add configurability for mean / likelihood
• Turn current mean / noise factories into serializable classes using new generic mechanism
• Add low-level support for gpytorch GP components

#747 (Index Kernel)

• Add BayBE IndexKernel class
• Add BayBE PositiveIndexKernel class

#763 (Kernel Arithmetic)

• improved + and * operations for Kernels

#748 (Active Dimensions)

• Enable active dimension control for kernels
• Absorb hardcoded IndexKernel logic into kernel factory
• Implement deprecation mechanism for breaking change

#776 (Kernel Factory Validation)

• Ensure kernel factories check if they support their given parameters

#752 (CHEN Preset)

• Add CHEN preset

#757 (BoTorch Preset)

• Add BOTORCH preset

Must TODO

• Add decorator for extending presets to apply mechanism they don't cover themselves (e.g. transfer learning)
• Stratification for multitask likelihoods
• Decide on new BayBE defaults (i.e. do we use BOTORCH, CHEN, etc, or some heuristic on top)
• Documentation

Optional TODO

• Add HVARFNER preset
• Enable serialization for gpytorch GP components
• Add preset Setting attribute

To be added but out of scope

• Make optimizer configurable