Bump min supported NumPy version to 1.26

Author: antonwolfy

conda-recipe/conda_build_config.yaml

@@ -1,5 +1,5 @@
 numpy:
-    - '1.25'
+    - '1.26'
 c_compiler:                   # [linux]
     - gcc                     # [linux]
 cxx_compiler:                 # [linux]

dpnp/tests/test_arraycreation.py

@@ -938,12 +938,9 @@ def test_logspace(dtype, num, endpoint):
     assert_allclose(dpnp_res, np_res, rtol=1e-06)
 
 
+@testing.with_requires("numpy>=1.25.0")
 @pytest.mark.parametrize("axis", [0, 1])
 def test_logspace_axis(axis):
-    if numpy.lib.NumpyVersion(numpy.__version__) < "1.25.0":
-        pytest.skip(
-            "numpy.logspace supports a non-scalar base argument since 1.25.0"
-        )
     func = lambda xp: xp.logspace(
         [2, 3], [20, 15], num=2, base=[[1, 3], [5, 7]], axis=axis
     )

pyproject.toml

@@ -8,7 +8,7 @@ requires = [
   "dpctl>=0.21.0",
   "ninja>=1.11.1; platform_system!='Windows'",
   # NOTE: no DPNP restriction on NumPy version, so follow NumPy's drop schedule
-  "numpy>=1.25.0",
+  "numpy>=1.26.0",
   "scikit-build>=0.18.1",
   "setuptools>=79.0.1",
   "wheel>=0.45.1",
@@ -50,7 +50,7 @@ dependencies = [
   # "intel-cmplr-lib-rt>=0.59.0"
   # WARNING: use the latest dpctl dev version, otherwise stable w/f will fail
   "dpctl>=0.21.0",
-  "numpy>=1.25.0"
+  "numpy>=1.26.0"
 ]
 description = "Data Parallel Extension for NumPy"
 dynamic = ["version"]