Add more state preparation kernels

[amccaskey]

We need to start adding more kernels to the stateprep module in Solvers. A good first one is the hardware efficient ansatz.

These should be implemented in C++ and bound to Python. See the UCCSD implementation as a reference.

For HWE, here is the signatures we want

__qpu__ void hwe(cudaq::qview<> qubits, std::size_t numLayers, const std::vector<double>& parameters); 

// cnotCoupling is a 1D array where CNOTs are applied on qubits i and i+1. 
__qpu__ void hwe(cudaq::qview<> qubits, std::size_t numLayers, const std::vector<double>& parameters, const std::vector<std::size_t>& cnotCoupling);