make type annotations compatible with numpy 2

cirq-superstaq/cirq_superstaq/ops/qubit_gates.py

@@ -66,7 +66,7 @@ def __init__(self, theta: cirq.TParamVal) -> None:
     def _num_qubits_(self) -> int:
         return 2
 
-    def _unitary_(self) -> npt.NDArray[np.complex_] | None:
+    def _unitary_(self) -> npt.NDArray[np.complex128] | None:
         if self._is_parameterized_():
             return None
         return np.array(
@@ -126,7 +126,7 @@ def _resolve_parameters_(self, resolver: cirq.ParamResolver, recursive: bool) ->
     def _has_unitary_(self) -> bool:
         return not self._is_parameterized_()
 
-    def _apply_unitary_(self, args: cirq.ApplyUnitaryArgs) -> npt.NDArray[np.complex_]:
+    def _apply_unitary_(self, args: cirq.ApplyUnitaryArgs) -> npt.NDArray[np.complex128]:
         zo = args.subspace_index(0b01)
         oz = args.subspace_index(0b10)
         args.available_buffer[zo] = args.target_tensor[zo]
@@ -184,7 +184,7 @@ class ZXPowGate(cirq.EigenGate):
 
     """
 
-    def _eigen_components(self) -> list[tuple[float, npt.NDArray[np.float_]]]:
+    def _eigen_components(self) -> list[tuple[float, npt.NDArray[np.float64]]]:
         return [
             (
                 0.0,
@@ -914,7 +914,7 @@ def rz_rads(self) -> cirq.TParamVal:
     def _num_qubits_(self) -> int:
         return 2
 
-    def _unitary_(self) -> npt.NDArray[np.complex_] | None:
+    def _unitary_(self) -> npt.NDArray[np.complex128] | None:
         if self._is_parameterized_():
             return None
         return np.diag(
@@ -990,7 +990,7 @@ def _json_dict_(self) -> dict[str, Any]:
 class DDPowGate(cirq.EigenGate):
     r"""The Dipole-Dipole gate for EeroQ hardware"""
 
-    def _eigen_components(self) -> list[tuple[float, npt.NDArray[np.float_]]]:
+    def _eigen_components(self) -> list[tuple[float, npt.NDArray[np.float64]]]:
         return [
             (
                 -0.5,

cirq-superstaq/cirq_superstaq/ops/qudit_gates.py

@@ -48,7 +48,7 @@ def _equal_up_to_global_phase_(self, other: Any, atol: float) -> bool | None:
     def _has_unitary_(self) -> bool:
         return True
 
-    def _apply_unitary_(self, args: cirq.ApplyUnitaryArgs) -> npt.NDArray[np.complex_] | None:
+    def _apply_unitary_(self, args: cirq.ApplyUnitaryArgs) -> npt.NDArray[np.complex128] | None:
         for i in range(self._dimension):
             for j in range(i):
                 idx0 = args.subspace_index(i * self._dimension + j)
@@ -118,7 +118,7 @@ def _swapped_state_indices(self) -> tuple[int, int]:
     def _eigen_shifts(self) -> list[float]:
         return [0.0, 0.5, -0.5]
 
-    def _eigen_components(self) -> list[tuple[float, npt.NDArray[np.float_]]]:
+    def _eigen_components(self) -> list[tuple[float, npt.NDArray[np.float64]]]:
         idx0, idx1 = self._swapped_state_indices()
 
         d = self.dimension**2
@@ -191,7 +191,7 @@ def dimension(self) -> int:
     def _qid_shape_(self) -> tuple[int, int]:
         return self.dimension, self.dimension
 
-    def _eigen_components(self) -> list[tuple[float, npt.NDArray[np.float_]]]:
+    def _eigen_components(self) -> list[tuple[float, npt.NDArray[np.float64]]]:
         eigen_components = []
         exponents = np.kron(range(self.dimension), range(self.dimension)) % self.dimension
         for x in sorted(set(exponents)):
@@ -304,7 +304,7 @@ def _with_exponent(self, exponent: cirq.TParamVal) -> VirtualZPowGate:
             global_shift=self._global_shift,
         )
 
-    def _eigen_components(self) -> list[tuple[float, npt.NDArray[np.float_]]]:
+    def _eigen_components(self) -> list[tuple[float, npt.NDArray[np.float64]]]:
         projector_phase = np.zeros(self._dimension)
         projector_phase[self._level :] = 1
         projector_rest = 1 - projector_phase
@@ -370,7 +370,7 @@ def _target_state(self) -> int:
     def _qid_shape_(self) -> tuple[int]:
         return (3,)
 
-    def _eigen_components(self) -> list[tuple[float, npt.NDArray[np.float_]]]:
+    def _eigen_components(self) -> list[tuple[float, npt.NDArray[np.float64]]]:
         d = self._qid_shape_()[0]
         projector_phase = np.zeros((d, d))
         projector_phase[self._target_state, self._target_state] = 1
@@ -542,7 +542,7 @@ def _resolve_parameters_(
     def _has_unitary_(self) -> bool:
         return cirq.has_unitary(self._sub_gate)
 
-    def _apply_unitary_(self, args: cirq.ApplyUnitaryArgs) -> npt.NDArray[np.complex_] | None:
+    def _apply_unitary_(self, args: cirq.ApplyUnitaryArgs) -> npt.NDArray[np.complex128] | None:
         if not cirq.has_unitary(self._sub_gate):
             return NotImplemented
 

cirq-superstaq/cirq_superstaq/service.py

@@ -419,7 +419,7 @@ def aqt_compile_eca(
         target: str = "aqt_keysight_qpu",
         atol: float | None = None,
         gate_defs: None | (
-            Mapping[str, npt.NDArray[np.complex_] | cirq.Gate | cirq.Operation | None]
+            Mapping[str, npt.NDArray[np.number[Any]] | cirq.Gate | cirq.Operation | None]
         ) = None,
         **kwargs: Any,
     ) -> css.compiler_output.CompilerOutput:
@@ -483,7 +483,7 @@ def aqt_compile(
         random_seed: int | None = None,
         atol: float | None = None,
         gate_defs: None | (
-            Mapping[str, npt.NDArray[np.complex_] | cirq.Gate | cirq.Operation | None]
+            Mapping[str, npt.NDArray[np.number[Any]] | cirq.Gate | cirq.Operation | None]
         ) = None,
         gateset: Mapping[str, Sequence[Sequence[int]]] | None = None,
         pulses: object = None,
@@ -816,7 +816,7 @@ def _get_compile_request_json(
 
     def supercheq(
         self, files: list[list[int]], num_qubits: int, depth: int
-    ) -> tuple[list[cirq.Circuit], npt.NDArray[np.float_]]:
+    ) -> tuple[list[cirq.Circuit], npt.NDArray[np.float64]]:
         """Returns the randomly generated circuits and the fidelity matrix for inputted files.
 
         Args:
@@ -1123,7 +1123,7 @@ def process_cb(self, job_id: str, counts: list[dict[str, int]] | None = None) ->
 
             def _objective(
                 x: np.typing.NDArray[np.int_], A: float, p: float
-            ) -> np.typing.NDArray[np.float_]:
+            ) -> np.typing.NDArray[np.float64]:
                 return A * p**x
 
             fit_data: defaultdict[str, float] = defaultdict(float)
@@ -1167,7 +1167,7 @@ def plot(self, circuits_and_metadata: dict[str, Any]) -> None:
 
         def _objective(
             x: np.typing.NDArray[np.int_], A: float, p: float
-        ) -> np.typing.NDArray[np.float_]:
+        ) -> np.typing.NDArray[np.float64]:
             return A * p**x
 
         e_f = 0.0

qiskit-superstaq/qiskit_superstaq/serialization.py

@@ -74,7 +74,7 @@ def json_encoder(val: object) -> dict[str, object]:
     raise TypeError(f"Object of type {type(val)} is not JSON serializable.")
 
 
-def json_resolver(val: T) -> T | npt.NDArray[np.complex_]:
+def json_resolver(val: T) -> T | npt.NDArray[np.complex128]:
     """Hook to deserialize objects that were serialized via `json_encoder()`.
 
     Args:

qiskit-superstaq/qiskit_superstaq/superstaq_backend.py

@@ -244,7 +244,7 @@ def aqt_compile(
         num_eca_circuits: int | None = None,
         random_seed: int | None = None,
         atol: float | None = None,
-        gate_defs: Mapping[str, str | npt.NDArray[np.complex_] | None] | None = None,
+        gate_defs: Mapping[str, str | npt.NDArray[np.number[Any]] | None] | None = None,
         gateset: Mapping[str, Sequence[Sequence[int]]] | None = None,
         pulses: object = None,
         variables: object = None,

qiskit-superstaq/qiskit_superstaq/superstaq_provider.py

@@ -187,7 +187,7 @@ def aqt_compile(
         num_eca_circuits: int | None = None,
         random_seed: int | None = None,
         atol: float | None = None,
-        gate_defs: Mapping[str, str | npt.NDArray[np.complex_] | None] | None = None,
+        gate_defs: Mapping[str, str | npt.NDArray[np.number[Any]] | None] | None = None,
         gateset: Mapping[str, Sequence[Sequence[int]]] | None = None,
         pulses: object = None,
         variables: object = None,
@@ -252,7 +252,7 @@ def aqt_compile_eca(
         random_seed: int | None = None,
         target: str = "aqt_keysight_qpu",
         atol: float | None = None,
-        gate_defs: Mapping[str, str | npt.NDArray[np.complex_] | None] | None = None,
+        gate_defs: Mapping[str, str | npt.NDArray[np.number[Any]] | None] | None = None,
         **kwargs: Any,
     ) -> qss.compiler_output.CompilerOutput:
         """Compiles and optimizes the given circuit(s) for the Advanced Quantum Testbed (AQT) at
@@ -458,7 +458,7 @@ def cq_compile(
 
     def supercheq(
         self, files: list[list[int]], num_qubits: int, depth: int
-    ) -> tuple[list[qiskit.QuantumCircuit], npt.NDArray[np.float_]]:
+    ) -> tuple[list[qiskit.QuantumCircuit], npt.NDArray[np.float64]]:
         """Returns Supercheq randomly generated circuits and the corresponding fidelity matrices.
 
         References:

supermarq-benchmarks/supermarq/benchmarks/qaoa_vanilla_proxy.py

@@ -96,8 +96,8 @@ def _get_expectation_value_from_probs(self, probabilities: Mapping[str, float])
             expectation_value += probability * self._get_energy_for_bitstring(bitstring)
         return expectation_value
 
-    def _get_opt_angles(self) -> tuple[npt.NDArray[np.float_], float]:
-        def f(params: npt.NDArray[np.float_]) -> float:
+    def _get_opt_angles(self) -> tuple[npt.NDArray[np.float64], float]:
+        def f(params: npt.NDArray[np.float64]) -> float:
             """The objective function to minimize.
 
             Args:
@@ -118,7 +118,7 @@ def f(params: npt.NDArray[np.float_]) -> float:
 
         return out["x"], out["fun"]
 
-    def _gen_angles(self) -> npt.NDArray[np.float_]:
+    def _gen_angles(self) -> npt.NDArray[np.float64]:
         # Classically simulate the variational optimization 5 times,
         # return the parameters from the best performing simulation
         best_params, best_cost = np.zeros(2), 0.0

supermarq-benchmarks/supermarq/benchmarks/vqe_proxy.py

@@ -46,7 +46,7 @@ def _gen_tfim_hamiltonian(self) -> list[tuple[str, int | tuple[int, int], int]]:
         hamiltonian.append(("ZZ", (self.num_qubits - 1, 0), 1))
         return hamiltonian
 
-    def _gen_ansatz(self, params: npt.NDArray[np.float_]) -> list[cirq.Circuit]:
+    def _gen_ansatz(self, params: npt.NDArray[np.float64]) -> list[cirq.Circuit]:
         qubits = cirq.LineQubit.range(self.num_qubits)
         z_circuit = cirq.Circuit()
 
@@ -117,8 +117,8 @@ def _get_expectation_value_from_probs(
 
         return avg_energy
 
-    def _get_opt_angles(self) -> tuple[npt.NDArray[np.float_], float]:
-        def f(params: npt.NDArray[np.float_]) -> float:
+    def _get_opt_angles(self) -> tuple[npt.NDArray[np.float64], float]:
+        def f(params: npt.NDArray[np.float64]) -> float:
             """The objective function to minimize.
 
             Args:
@@ -141,7 +141,7 @@ def f(params: npt.NDArray[np.float_]) -> float:
 
         return out["x"], out["fun"]
 
-    def _gen_angles(self) -> npt.NDArray[np.float_]:
+    def _gen_angles(self) -> npt.NDArray[np.float64]:
         """Classically simulate the variational optimization and return
         the final parameters.
         """

supermarq-benchmarks/supermarq/plotting.py

@@ -239,7 +239,7 @@ def plot_benchmark(
 
 
 def heatmap(
-    data: npt.NDArray[np.float_],
+    data: npt.NDArray[np.floating[Any]],
     ax: matplotlib.axes.Axes,
     row_labels: list[str],
     col_labels: list[str],
@@ -306,7 +306,7 @@ def heatmap(
 
 def annotate_heatmap(
     im: matplotlib.image.AxesImage,
-    data: npt.NDArray[np.float_] | None = None,
+    data: npt.NDArray[np.floating[Any]] | None = None,
     valfmt: Any = "{x:.2f}",
     textcolors: tuple[str, str] = ("black", "white"),
     threshold: float | None = None,
@@ -360,7 +360,7 @@ def annotate_heatmap(
     return texts
 
 
-def radar_factory(num_vars: int) -> npt.NDArray[np.float_]:
+def radar_factory(num_vars: int) -> npt.NDArray[np.float64]:
     """Create a radar chart with `num_vars` axes.
 
     (https://matplotlib.org/stable/gallery/specialty_plots/radar_chart.html) This function
@@ -398,7 +398,7 @@ class RadarAxesMeta(PolarAxes):
 
     def __init__(self, *args: Any, **kwargs: Any) -> None:
         """Initializes the `RadarAxesMeta` class."""
-        self.theta: npt.NDArray[np.float_]
+        self.theta: npt.NDArray[np.float64]
         self.num_vars: int
         self.frame: str
 

supermarq-benchmarks/supermarq/qcvv/irb.py

@@ -23,6 +23,7 @@
 import cirq
 import cirq.circuits
 import numpy as np
+import numpy.typing as npt
 import pandas as pd
 import scipy
 import seaborn as sns
@@ -635,7 +636,7 @@ def analyze_results(self, plot_results: bool = True) -> IRBResults | RBResults:
 
     def _fit_decay(
         self, experiment: str = "RB"
-    ) -> tuple[np.typing.NDArray[np.float_], np.typing.NDArray[np.float_]]:
+    ) -> tuple[np.typing.NDArray[np.float64], np.typing.NDArray[np.float64]]:
         """Fits the exponential decay function to either the RB or IRB results.
 
         Args:
@@ -660,9 +661,7 @@ def _fit_decay(
         return popt, np.sqrt(np.diag(pcov))
 
     @staticmethod
-    def exp_decay(
-        x: np.typing.NDArray[np.float_], A: float, alpha: float, B: float
-    ) -> np.typing.NDArray[np.float_]:
+    def exp_decay(x: npt.ArrayLike, A: float, alpha: float, B: float) -> npt.ArrayLike:
         r"""Exponential decay of the form
 
         .. math::
@@ -678,4 +677,4 @@ def exp_decay(
         Returns:
             Exponential decay applied to x.
         """
-        return A * (alpha**x) + B
+        return A * (np.asarray(alpha) ** x) + B