Adding Kernels to gpp_covariance.cpp and python wrappers

[bwbellmath]

My collaborator added some covariance functions (using polynomial exponential and matern kernels) to the gpp_covariance.cpp and gpp_covariance.hpp and the python wrapper covariance.py and also in domain.py

however when attempting to build moe with these changes -- cmake succeeds meaning the C++ code seems good, but I don't understand how the python code gets compiled and placed in the appropriate place. Therefore I don't know how to build moe so that I can use these new covariance functions through the python interface.

Could some detail be added to the documentation regarding adding custom functions via C+++ with wrappers?