let MinimumEigenOptimizer allow new primitive-based algorithms

Author: t-imamichi

qiskit_optimization/algorithms/minimum_eigen_optimizer.py

@@ -1,6 +1,6 @@
 # This code is part of Qiskit.
 #
-# (C) Copyright IBM 2020, 2021.
+# (C) Copyright IBM 2020, 2022.
 #
 # This code is licensed under the Apache License, Version 2.0. You may
 # obtain a copy of this license in the LICENSE.txt file in the root directory
@@ -11,24 +11,38 @@
 # that they have been altered from the originals.
 
 """A wrapper for minimum eigen solvers to be used within the optimization module."""
-from typing import Optional, Union, List, cast
+from typing import List, Optional, Union, cast
 
 import numpy as np
+from qiskit.algorithms.minimum_eigen_solvers import MinimumEigensolver as LegacyMinimumEigensolver
+from qiskit.algorithms.minimum_eigen_solvers import (
+    MinimumEigensolverResult as LegacyMinimumEigensolverResult,
+)
+from qiskit.algorithms.minimum_eigensolvers import (
+    NumPyMinimumEigensolver,
+    NumPyMinimumEigensolverResult,
+    SamplingMinimumEigensolver,
+    SamplingMinimumEigensolverResult,
+)
+from qiskit.opflow import OperatorBase, PauliOp, PauliSumOp
 
-from qiskit.algorithms import MinimumEigensolver, MinimumEigensolverResult
-from qiskit.opflow import OperatorBase
+from ..converters.quadratic_program_to_qubo import QuadraticProgramConverter, QuadraticProgramToQubo
+from ..deprecation import DeprecatedType, warn_deprecated
+from ..exceptions import QiskitOptimizationError
+from ..problems.quadratic_program import QuadraticProgram, Variable
 from .optimization_algorithm import (
-    OptimizationResultStatus,
     OptimizationAlgorithm,
     OptimizationResult,
+    OptimizationResultStatus,
     SolutionSample,
 )
-from ..exceptions import QiskitOptimizationError
-from ..converters.quadratic_program_to_qubo import (
-    QuadraticProgramToQubo,
-    QuadraticProgramConverter,
-)
-from ..problems.quadratic_program import QuadraticProgram, Variable
+
+MinimumEigensolver = Union[
+    SamplingMinimumEigensolver, NumPyMinimumEigensolver, LegacyMinimumEigensolver
+]
+MinimumEigensolverResult = Union[
+    SamplingMinimumEigensolverResult, NumPyMinimumEigensolverResult, LegacyMinimumEigensolverResult
+]
 
 
 class MinimumEigenOptimizationResult(OptimizationResult):
@@ -137,11 +151,18 @@ def __init__(
             TypeError: When one of converters has an invalid type.
             QiskitOptimizationError: When the minimum eigensolver does not return an eigenstate.
         """
-
+        if isinstance(min_eigen_solver, LegacyMinimumEigensolver):
+            warn_deprecated(
+                "0.5.0",
+                DeprecatedType.ARGUMENT,
+                f"min_eigen_solver as {LegacyMinimumEigensolver.__name__}",
+                new_name=f"min_eigen_solver as {SamplingMinimumEigensolver.__name__} "
+                f"or {NumPyMinimumEigensolver.__name__}",
+            )
         if not min_eigen_solver.supports_aux_operators():
             raise QiskitOptimizationError(
                 "Given MinimumEigensolver does not return the eigenstate "
-                + "and is not supported by the MinimumEigenOptimizer."
+                "and is not supported by the MinimumEigenOptimizer."
             )
         self._min_eigen_solver = min_eigen_solver
         self._penalty = penalty
@@ -206,6 +227,9 @@ def _solve_internal(
         # only try to solve non-empty Ising Hamiltonians
         eigen_result: Optional[MinimumEigensolverResult] = None
         if operator.num_qubits > 0:
+            # NumPyEigensolver does not accept PauliOp but PauliSumOp
+            if isinstance(operator, PauliOp):
+                operator = PauliSumOp.from_list([(operator.primitive.to_label(), operator.coeff)])
             # approximate ground state of operator using min eigen solver
             eigen_result = self._min_eigen_solver.compute_minimum_eigenvalue(operator)
             # analyze results

qiskit_optimization/algorithms/optimization_algorithm.py

@@ -15,14 +15,16 @@
 from abc import ABC, abstractmethod
 from dataclasses import dataclass
 from enum import Enum
-from typing import List, Union, Any, Optional, Dict, Type, Tuple, cast
+from typing import Any, Dict, List, Optional, Tuple, Type, Union, cast
 from warnings import warn
 
 import numpy as np
+from qiskit.opflow import DictStateFn, StateFn
+from qiskit.quantum_info import Statevector
+from qiskit.result import QuasiDistribution
 
-from qiskit.opflow import StateFn, DictStateFn
+from ..converters.quadratic_program_to_qubo import QuadraticProgramConverter, QuadraticProgramToQubo
 from ..exceptions import QiskitOptimizationError
-from ..converters.quadratic_program_to_qubo import QuadraticProgramToQubo, QuadraticProgramConverter
 from ..problems.quadratic_program import QuadraticProgram, Variable
 
 
@@ -518,7 +520,7 @@ def _interpret_samples(
 
     @staticmethod
     def _eigenvector_to_solutions(
-        eigenvector: Union[dict, np.ndarray, StateFn],
+        eigenvector: Union[QuasiDistribution, Statevector, dict, np.ndarray, StateFn],
         qubo: QuadraticProgram,
         min_probability: float = 1e-6,
     ) -> List[SolutionSample]:
@@ -566,7 +568,25 @@ def generate_solution(bitstr, qubo, probability):
             )
 
         solutions = []
-        if isinstance(eigenvector, dict):
+        if isinstance(eigenvector, QuasiDistribution):
+            probabilities = eigenvector.binary_probabilities()
+            # iterate over all samples
+            for bitstr, sampling_probability in probabilities.items():
+                # add the bitstring, if the sampling probability exceeds the threshold
+                if sampling_probability >= min_probability:
+                    solutions.append(generate_solution(bitstr, qubo, sampling_probability))
+
+        elif isinstance(eigenvector, Statevector):
+            probabilities = eigenvector.probabilities()
+            num_qubits = eigenvector.num_qubits
+            # iterate over all states and their sampling probabilities
+            for i, sampling_probability in enumerate(probabilities):
+                # add the i-th state if the sampling probability exceeds the threshold
+                if sampling_probability >= min_probability:
+                    bitstr = f"{i:b}".rjust(num_qubits, "0")
+                    solutions.append(generate_solution(bitstr, qubo, sampling_probability))
+
+        elif isinstance(eigenvector, dict):
             # When eigenvector is a dict, square the values since the values are normalized.
             # See https://github.com/Qiskit/qiskit-terra/pull/5496 for more details.
             probabilities = {bitstr: val**2 for (bitstr, val) in eigenvector.items()}
@@ -579,7 +599,6 @@ def generate_solution(bitstr, qubo, probability):
         elif isinstance(eigenvector, np.ndarray):
             num_qubits = int(np.log2(eigenvector.size))
             probabilities = np.abs(eigenvector * eigenvector.conj())
-
             # iterate over all states and their sampling probabilities
             for i, sampling_probability in enumerate(probabilities):
                 # add the i-th state if the sampling probability exceeds the threshold
@@ -588,6 +607,8 @@ def generate_solution(bitstr, qubo, probability):
                     solutions.append(generate_solution(bitstr, qubo, sampling_probability))
 
         else:
-            raise TypeError("Unsupported format of eigenvector. Provide a dict or numpy.ndarray.")
-
+            raise TypeError(
+                f"Eigenvector should be QuasiDistribution, Statevector, dict or numpy.ndarray. "
+                f"But, it was {type(eigenvector)}."
+            )
         return solutions