This PR adds a whole host of new features including:
Massively updated documentation
Added a new utility function option: naive
Made optGP, a function for optimizing the GP hyperparameters, a class method
Added ability to pass GP hyperparameter optimization priors
Added new function to find maximum a posteriori (MAP) estimate given a trained GP
Added new example, tests for MAP functions
Added ability to ignore initial training set after 0th iteration (sometimes useful, apparently?)
Added many new tests for GP optimization, finding next point, etc
Added ability to toggle whether or not to fit for the GP kernel amplitude (sometimes can be very numerically unstable when the amp is included for large dim cases, even with regularization)
This PR adds a whole host of new features including:
naive