utils.py

import numpy as np
from qiskit.quantum_info import SparsePauliOp

def compute_cut_size(graph, bitstring):
    """
    Get the cut size of the partition of ``graph`` described by the given
    ``bitstring``.
    """
    cut_sz = 0
    for (u, v) in graph.edges:
        if bitstring[u] != bitstring[v]:
            cut_sz += 1
    return cut_sz

def get_ising_energies(
        operator: SparsePauliOp, 
        states: np.array
    ):
    """
    Get the energies of the given Ising ``operator`` that correspond to the
    given ``states``.
    """
    # Unroll Hamiltonian data into NumPy arrays
    paulis = np.array([list(ops) for ops, _ in operator.label_iter()]) != "I"
    coeffs = operator.coeffs.real
    
    # Vectorized energies computation
    energies = (-1) ** (states @ paulis.T) @ coeffs
    return energies

def expected_energy(
        hamiltonian: SparsePauliOp,
        measurements: np.array
):
    """
    Compute the expected energy of the given ``hamiltonian`` with respect to
    the observed ``measurement``.

    The latter is assumed to by a NumPy records array with fields ``states``
    --describing the observed bit-strings as an integer array-- and ``counts``,
    describing the corresponding observed frequency of each state.
    """
    energies = get_ising_energies(hamiltonian, measurements["states"])
    return np.dot(energies, measurements["counts"]) / measurements["counts"].sum()