StabilizerState.expectation_value should accept SparsePauliOp

[kevinsung]

What should we add?

Currently, it only accepts Pauli. See

qiskit/qiskit/quantum_info/states/stabilizerstate.py

Lines 257 to 320 in e28f30e

	def expectation_value(self, oper: Pauli, qargs: None | list = None) -> complex:
	"""Compute the expectation value of a Pauli operator.
	Args:
	oper (Pauli): a Pauli operator to evaluate expval.
	qargs (None or list): subsystems to apply the operator on.
	Returns:
	complex: the expectation value (only 0 or 1 or -1 or i or -i).
	Raises:
	QiskitError: if oper is not a Pauli operator.
	"""
	if not isinstance(oper, Pauli):
	raise QiskitError("Operator for expectation value is not a Pauli operator.")
	num_qubits = self.clifford.num_qubits
	if qargs is None:
	qubits = range(num_qubits)
	else:
	qubits = qargs
	# Construct Pauli on num_qubits
	pauli = Pauli(num_qubits * "I")
	phase = 0
	pauli_phase = (-1j) ** oper.phase if oper.phase else 1
	for pos, qubit in enumerate(qubits):
	pauli.x[qubit] = oper.x[pos]
	pauli.z[qubit] = oper.z[pos]
	phase += pauli.x[qubit] & pauli.z[qubit]
	# Check if there is a stabilizer that anti-commutes with an odd number of qubits
	# If so the expectation value is 0
	for p in range(num_qubits):
	num_anti = 0
	num_anti += np.count_nonzero(pauli.z & self.clifford.stab_x[p])
	num_anti += np.count_nonzero(pauli.x & self.clifford.stab_z[p])
	if num_anti % 2 == 1:
	return 0
	# Otherwise pauli is (-1)^a prod_j S_j^b_j for Clifford stabilizers
	# If pauli anti-commutes with D_j then b_j = 1.
	# Multiply pauli by stabilizers with anti-commuting destabilizers
	pauli_z = (pauli.z).copy() # Make a copy of pauli.z
	for p in range(num_qubits):
	# Check if destabilizer anti-commutes
	num_anti = 0
	num_anti += np.count_nonzero(pauli.z & self.clifford.destab_x[p])
	num_anti += np.count_nonzero(pauli.x & self.clifford.destab_z[p])
	if num_anti % 2 == 0:
	continue
	# If anti-commutes multiply Pauli by stabilizer
	phase += 2 * self.clifford.stab_phase[p]
	phase += np.count_nonzero(self.clifford.stab_z[p] & self.clifford.stab_x[p])
	phase += 2 * np.count_nonzero(pauli_z & self.clifford.stab_x[p])
	pauli_z = pauli_z ^ self.clifford.stab_z[p]
	# For valid stabilizers, `phase` can only be 0 (= 1) or 2 (= -1) at this point.
	if phase % 4 != 0:
	return -pauli_phase
	return pauli_phase