Merge pull request #96 from qutech/numpy-2.0

Support Numpy 2.0

filter_functions/basis.py

@@ -47,7 +47,6 @@
 import numpy as np
 import opt_einsum as oe
 from numpy import linalg as nla
-from numpy.core import ndarray
 from scipy import linalg as sla
 from sparse import COO
 
@@ -56,7 +55,7 @@
 __all__ = ['Basis', 'expand', 'ggm_expand', 'normalize']
 
 
-class Basis(ndarray):
+class Basis(np.ndarray):
     r"""
     Class for operator bases. There are several ways to instantiate a
     Basis object:
@@ -217,22 +216,26 @@ def __eq__(self, other: object) -> bool:
             # Not ndarray
             return np.equal(self, other)
 
-        return np.allclose(self.view(ndarray), other.view(ndarray),
+        return np.allclose(self.view(np.ndarray), other.view(np.ndarray),
                            atol=self._atol, rtol=self._rtol)
 
-    def __contains__(self, item: ndarray) -> bool:
+    def __contains__(self, item: np.ndarray) -> bool:
         """Implement 'in' operator."""
-        return any(np.isclose(item.view(ndarray), self.view(ndarray),
+        return any(np.isclose(item.view(np.ndarray), self.view(np.ndarray),
                               rtol=self._rtol, atol=self._atol).all(axis=(1, 2)))
 
-    def __array_wrap__(self, out_arr, context=None):
+    def __array_wrap__(self, arr, context=None, return_scalar=False):
         """
         Fixes problem that ufuncs return 0-d arrays instead of scalars.
 
         https://github.com/numpy/numpy/issues/5819#issue-72454838
         """
-        if out_arr.ndim:
-            return ndarray.__array_wrap__(self, out_arr, context)
+        try:
+            return super().__array_wrap__(arr, context, return_scalar=True)
+        except TypeError:
+            if arr.ndim:
+                # Numpy < 2
+                return np.ndarray.__array_wrap__(self, arr, context)
 
     def _print_checks(self) -> None:
         """Print checks for debug purposes."""
@@ -265,7 +268,7 @@ def isorthonorm(self) -> bool:
                 actual = U.conj() @ U.T
                 target = np.identity(dim)
                 atol = self._eps*(self.d**2)**3
-                self._isorthonorm = np.allclose(actual.view(ndarray), target,
+                self._isorthonorm = np.allclose(actual.view(np.ndarray), target,
                                                 atol=atol, rtol=self._rtol)
 
         return self._isorthonorm
@@ -278,13 +281,16 @@ def istraceless(self) -> bool:
         if self._istraceless is None:
             trace = np.einsum('...jj', self)
             trace = util.remove_float_errors(trace, self.d**2)
-            nonzero = trace.nonzero()
+            nonzero = np.atleast_1d(trace).nonzero()
             if nonzero[0].size == 0:
                 self._istraceless = True
             elif nonzero[0].size == 1:
                 # Single element has nonzero trace, check if (proportional to)
                 # identity
-                elem = self[nonzero][0].view(ndarray) if self.ndim == 3 else self.view(ndarray)
+                if self.ndim == 3:
+                    elem = self[nonzero][0].view(np.ndarray)
+                else:
+                    elem = self.view(np.ndarray)
                 offdiag_nonzero = elem[~np.eye(self.d, dtype=bool)].nonzero()
                 diag_equal = np.diag(elem) == elem[0, 0]
                 if diag_equal.all() and not offdiag_nonzero[0].any():
@@ -597,7 +603,7 @@ def _full_from_partial(elems: Sequence, traceless: bool, labels: Sequence[str])
             # sort Identity label to the front, default to first if not found
             # (should not happen since traceless checks that it is present)
             id_idx = next((i for i, elem in enumerate(elems)
-                           if np.allclose(Id.view(ndarray), elem.view(ndarray),
+                           if np.allclose(Id.view(np.ndarray), elem.view(np.ndarray),
                                           rtol=elems._rtol, atol=elems._atol)), 0)
             labels.insert(0, labels.pop(id_idx))
 
@@ -606,7 +612,7 @@ def _full_from_partial(elems: Sequence, traceless: bool, labels: Sequence[str])
     return basis, labels
 
 
-def _norm(b: Sequence) -> ndarray:
+def _norm(b: Sequence) -> np.ndarray:
     """Frobenius norm with two singleton dimensions inserted at the end."""
     b = np.asanyarray(b)
     norm = nla.norm(b, axis=(-1, -2))
@@ -633,8 +639,8 @@ def normalize(b: Basis) -> Basis:
     return (b/_norm(b)).squeeze().view(Basis)
 
 
-def expand(M: Union[ndarray, Basis], basis: Union[ndarray, Basis],
-           normalized: bool = True, hermitian: bool = False, tidyup: bool = False) -> ndarray:
+def expand(M: Union[np.ndarray, Basis], basis: Union[np.ndarray, Basis],
+           normalized: bool = True, hermitian: bool = False, tidyup: bool = False) -> np.ndarray:
     r"""
     Expand the array *M* in the basis given by *basis*.
 
@@ -684,8 +690,8 @@ def cast(arr):
     return util.remove_float_errors(coefficients) if tidyup else coefficients
 
 
-def ggm_expand(M: Union[ndarray, Basis], traceless: bool = False,
-               hermitian: bool = False) -> ndarray:
+def ggm_expand(M: Union[np.ndarray, Basis], traceless: bool = False,
+               hermitian: bool = False) -> np.ndarray:
     r"""
     Expand the matrix *M* in a Generalized Gell-Mann basis [Bert08]_.
     This function makes use of the explicit construction prescription of
@@ -767,7 +773,7 @@ def cast(arr):
     return coeffs.squeeze() if square else coeffs
 
 
-def equivalent_pauli_basis_elements(idx: Union[Sequence[int], int], N: int) -> ndarray:
+def equivalent_pauli_basis_elements(idx: Union[Sequence[int], int], N: int) -> np.ndarray:
     """
     Get the indices of the equivalent (up to identities tensored to it)
     basis elements of Pauli bases of qubits at position idx in the total
@@ -780,7 +786,7 @@ def equivalent_pauli_basis_elements(idx: Union[Sequence[int], int], N: int) -> n
     return elem_idx
 
 
-def remap_pauli_basis_elements(order: Sequence[int], N: int) -> ndarray:
+def remap_pauli_basis_elements(order: Sequence[int], N: int) -> np.ndarray:
     """
     For a N-qubit Pauli basis, transpose the order of the subsystems and
     return the indices that permute the old basis to the new.

setup.py

@@ -20,7 +20,7 @@ def extract_version(version_file):
                   'bloch_sphere_visualization': ['qutip', 'matplotlib'],
                   'fancy_progressbar': ['ipynbname', 'jupyter'],
                   'doc': ['jupyter', 'nbsphinx', 'numpydoc', 'sphinx', 'sphinx_rtd_theme',
-                          'ipympl', 'qutip-qip', 'qutip-qtrl'],
+                          'ipympl', 'qutip-qip', 'qutip-qtrl', 'numpy<2'],
                   'tests': ['pytest>=4.6', 'pytest-cov', 'codecov']}
 
 extras_require['all'] = list({dep for deps in extras_require.values() for dep in deps})

tests/test_basis.py

@@ -24,12 +24,11 @@
 from copy import copy
 from itertools import product
 
+import filter_functions as ff
 import numpy as np
 import pytest
-from sparse import COO
-
-import filter_functions as ff
 from filter_functions import util
+from sparse import COO
 from tests import testutil
 from tests.testutil import rng
 
@@ -152,9 +151,21 @@ def test_basis_properties(self):
 
             base._print_checks()
 
-        basis = util.paulis[1].view(ff.Basis)
-        self.assertTrue(basis.isorthonorm)
-        self.assertArrayEqual(basis.T, basis.view(np.ndarray).T)
+        # single element always considered orthonormal
+        orthonorm = rng.normal(size=(d, d))
+        self.assertTrue(orthonorm.view(ff.Basis).isorthonorm)
+
+        herm = testutil.rand_herm(d).squeeze()
+        self.assertTrue(herm.view(ff.Basis).isherm)
+
+        herm[0, 1] += 1
+        self.assertFalse(herm.view(ff.Basis).isherm)
+
+        traceless = testutil.rand_herm_traceless(d).squeeze()
+        self.assertTrue(traceless.view(ff.Basis).istraceless)
+
+        traceless[0, 0] += 1
+        self.assertFalse(traceless.view(ff.Basis).istraceless)
 
     def test_transpose(self):
         arr = rng.normal(size=(2, 3, 3))

tests/test_gradient.py

@@ -134,7 +134,7 @@ def test_gradient_calculation_random_pulse(self):
                 spectral_noise_density=gradient_testutil.one_over_f_noise,
                 c_id=[f'c{i}' for i in range(len(u_ctrl))], n_coeffs_deriv=None
             )
-            self.assertArrayAlmostEqual(ana_grad, fin_diff_grad, rtol=1e-6, atol=1e-8)
+            self.assertArrayAlmostEqual(ana_grad, fin_diff_grad, rtol=1e-5, atol=1e-7)
 
     def test_caching(self):
         """Make sure calculation works with or without cached intermediates."""