update test to pass with numpy-2.3

Author: vtavana

.github/workflows/conda-package.yml

@@ -22,7 +22,7 @@ env:
   test-env-name: 'test'
   rerun-tests-on-failure: 'true'
   rerun-tests-max-attempts: 2
-  rerun-tests-timeout: 35
+  rerun-tests-timeout: 40
 
 jobs:
   build:

dpnp/tests/test_linalg.py

@@ -23,7 +23,6 @@
     get_integer_float_dtypes,
     has_support_aspect64,
     is_cpu_device,
-    is_cuda_device,
     numpy_version,
     requires_intel_mkl_version,
 )
@@ -2104,11 +2103,12 @@ def test_empty(self, shape, ord, axis, keepdims):
                 assert_raises(ValueError, dpnp.linalg.norm, ia, **kwarg)
                 assert_raises(ValueError, numpy.linalg.norm, a, **kwarg)
             else:
-                # TODO: when similar changes in numpy are available, instead
-                # of assert_equal with zero, we should compare with numpy
-                # ord in [None, 1, 2]
-                assert_equal(dpnp.linalg.norm(ia, **kwarg), 0.0)
-                assert_raises(ValueError, numpy.linalg.norm, a, **kwarg)
+                if numpy_version() >= "2.3.0":
+                    result = dpnp.linalg.norm(ia, **kwarg)
+                    expected = numpy.linalg.norm(a, **kwarg)
+                    assert_dtype_allclose(result, expected)
+                else:
+                    assert_equal(dpnp.linalg.norm(ia, **kwarg), 0.0)
         else:
             result = dpnp.linalg.norm(ia, **kwarg)
             expected = numpy.linalg.norm(a, **kwarg)

dpnp/tests/test_product.py

@@ -12,8 +12,6 @@
     assert_dtype_allclose,
     generate_random_numpy_array,
     get_all_dtypes,
-    get_complex_dtypes,
-    is_win_platform,
     numpy_version,
 )
 from .third_party.cupy import testing
@@ -845,6 +843,8 @@ def test_dtype_matrix(self, dt_in1, dt_in2, dt_out, shape1, shape2):
             assert_raises(TypeError, dpnp.matmul, ia, ib, out=iout)
             assert_raises(TypeError, numpy.matmul, a, b, out=out)
 
+    # TODO: include numpy-2.3 when numpy-issue-29164 is resolved
+    @testing.with_requires("numpy<2.3")
     @pytest.mark.parametrize("dtype", _selected_dtypes)
     @pytest.mark.parametrize("order1", ["C", "F", "A"])
     @pytest.mark.parametrize("order2", ["C", "F", "A"])
@@ -882,6 +882,8 @@ def test_order(self, dtype, order1, order2, order, shape1, shape2):
         assert result.flags.f_contiguous == expected.flags.f_contiguous
         assert_dtype_allclose(result, expected)
 
+    # TODO: include numpy-2.3 when numpy-issue-29164 is resolved
+    @testing.with_requires("numpy<2.3")
     @pytest.mark.parametrize("dtype", _selected_dtypes)
     @pytest.mark.parametrize(
         "stride",

dpnp/tests/testing/array.py

@@ -49,7 +49,10 @@ def _assert(assert_func, result, expected, *args, **kwargs):
     ]
     # For numpy < 2.0, some tests will fail for dtype mismatch
     dev = dpctl.select_default_device()
-    if numpy.__version__ >= "2.0.0" and dev.has_aspect_fp64:
+    if (
+        numpy.lib.NumpyVersion(numpy.__version__) >= "2.0.0"
+        and dev.has_aspect_fp64
+    ):
         strict = kwargs.setdefault("strict", True)
         if flag:
             if strict:

dpnp/tests/third_party/cupy/manipulation_tests/test_add_remove.py

@@ -340,7 +340,8 @@ def test_unique_inverse(self, xp, dtype, attr):
         a = testing.shaped_random((100, 100), xp, dtype)
         return getattr(xp.unique_inverse(a), attr)
 
-    @testing.with_requires("numpy>=2.0")
+    # TODO: include numpy-2.3 when dpnp-issue-2476 is addressed
+    @testing.with_requires("numpy>=2.0", "numpy<2.3")
     @testing.for_all_dtypes(no_float16=True, no_bool=True, no_complex=True)
     @testing.numpy_cupy_array_equal()
     def test_unique_values(self, xp, dtype):

dpnp/tests/third_party/cupy/math_tests/test_matmul.py

@@ -83,6 +83,8 @@ def test_cupy_matmul(self, xp, dtype1, dtype2):
         return xp.matmul(x1, x2)
 
 
+# TODO: include numpy-2.3 when numpy-issue-29164 is resolved
+@testing.with_requires("numpy<2.3")
 @testing.parameterize(
     *testing.product(
         {