Optimization constraints

[lkwagner]

Make an adaptor object that wraps parameters derivatives.

init should be a representation of what parameters should be set equal or constant.

There should be three functions:

linearize_parameters:
input: dictionary of wave function parameters
output: one-dimensional numpy array of parameters

unlinearize_parameters:
input: 1-D array
output: dictionary of parameters

linearize_gradient:
input: dictionary of wave function parameter derivatives
output: one-dimensional array of reduced parameter derivatives.