MAINT: Adapt sklearn for NumPy default integer change (#27041)

seberg · ogrisel · web-flow · commit 0ab36990c0a7 · 2023-11-16T09:35:16.000-05:00
Co-authored-by: Olivier Grisel &lt;olivier.grisel@ensta.org&gt;
diff --git a/sklearn/cluster/_hdbscan/hdbscan.py b/sklearn/cluster/_hdbscan/hdbscan.py
@@ -124,7 +124,7 @@ def _brute_mst(mutual_reachability, min_samples):
     # Compute the minimum spanning tree for the sparse graph
     sparse_min_spanning_tree = csgraph.minimum_spanning_tree(mutual_reachability)
     rows, cols = sparse_min_spanning_tree.nonzero()
-    mst = np.core.records.fromarrays(
+    mst = np.rec.fromarrays(
         [rows, cols, sparse_min_spanning_tree.data],
         dtype=MST_edge_dtype,
     )
diff --git a/sklearn/utils/_random.pxd b/sklearn/utils/_random.pxd
@@ -16,10 +16,6 @@ cdef enum:
     # 32-bit signed integers (i.e. 2^31 - 1).
     RAND_R_MAX = 2147483647
 
-cpdef sample_without_replacement(cnp.int_t n_population,
-                                 cnp.int_t n_samples,
-                                 method=*,
-                                 random_state=*)
 
 # rand_r replacement using a 32bit XorShift generator
 # See http://www.jstatsoft.org/v08/i14/paper for details
diff --git a/sklearn/utils/_random.pyx b/sklearn/utils/_random.pyx
@@ -19,8 +19,17 @@ from . import check_random_state
 cdef UINT32_t DEFAULT_SEED = 1
 
 
-cpdef _sample_without_replacement_check_input(cnp.int_t n_population,
-                                              cnp.int_t n_samples):
+# Compatibility type to always accept the default int type used by NumPy, both
+# before and after NumPy 2. On Windows, `long` does not always match `cnp.inp_t`.
+# See the comments in the `sample_without_replacement` Python function for more
+# details.
+ctypedef fused default_int:
+    cnp.intp_t
+    long
+
+
+cpdef _sample_without_replacement_check_input(default_int n_population,
+                                              default_int n_samples):
     """ Check that input are consistent for sample_without_replacement"""
     if n_population < 0:
         raise ValueError('n_population should be greater than 0, got %s.'
@@ -33,8 +42,8 @@ cpdef _sample_without_replacement_check_input(cnp.int_t n_population,
 
 
 cpdef _sample_without_replacement_with_tracking_selection(
-        cnp.int_t n_population,
-        cnp.int_t n_samples,
+        default_int n_population,
+        default_int n_samples,
         random_state=None):
     r"""Sample integers without replacement.
 
@@ -76,9 +85,9 @@ cpdef _sample_without_replacement_with_tracking_selection(
     """
     _sample_without_replacement_check_input(n_population, n_samples)
 
-    cdef cnp.int_t i
-    cdef cnp.int_t j
-    cdef cnp.int_t[::1] out = np.empty((n_samples, ), dtype=int)
+    cdef default_int i
+    cdef default_int j
+    cdef default_int[::1] out = np.empty((n_samples, ), dtype=int)
 
     rng = check_random_state(random_state)
     rng_randint = rng.randint
@@ -97,8 +106,8 @@ cpdef _sample_without_replacement_with_tracking_selection(
     return np.asarray(out)
 
 
-cpdef _sample_without_replacement_with_pool(cnp.int_t n_population,
-                                            cnp.int_t n_samples,
+cpdef _sample_without_replacement_with_pool(default_int n_population,
+                                            default_int n_samples,
                                             random_state=None):
     """Sample integers without replacement.
 
@@ -131,10 +140,10 @@ cpdef _sample_without_replacement_with_pool(cnp.int_t n_population,
     """
     _sample_without_replacement_check_input(n_population, n_samples)
 
-    cdef cnp.int_t i
-    cdef cnp.int_t j
-    cdef cnp.int_t[::1] out = np.empty((n_samples,), dtype=int)
-    cdef cnp.int_t[::1] pool = np.empty((n_population,), dtype=int)
+    cdef default_int i
+    cdef default_int j
+    cdef default_int[::1] out = np.empty((n_samples,), dtype=int)
+    cdef default_int[::1] pool = np.empty((n_population,), dtype=int)
 
     rng = check_random_state(random_state)
     rng_randint = rng.randint
@@ -154,8 +163,8 @@ cpdef _sample_without_replacement_with_pool(cnp.int_t n_population,
 
 
 cpdef _sample_without_replacement_with_reservoir_sampling(
-    cnp.int_t n_population,
-    cnp.int_t n_samples,
+    default_int n_population,
+    default_int n_samples,
     random_state=None
 ):
     """Sample integers without replacement.
@@ -191,9 +200,9 @@ cpdef _sample_without_replacement_with_reservoir_sampling(
     """
     _sample_without_replacement_check_input(n_population, n_samples)
 
-    cdef cnp.int_t i
-    cdef cnp.int_t j
-    cdef cnp.int_t[::1] out = np.empty((n_samples, ), dtype=int)
+    cdef default_int i
+    cdef default_int j
+    cdef default_int[::1] out = np.empty((n_samples, ), dtype=int)
 
     rng = check_random_state(random_state)
     rng_randint = rng.randint
@@ -213,8 +222,8 @@ cpdef _sample_without_replacement_with_reservoir_sampling(
     return np.asarray(out)
 
 
-cpdef sample_without_replacement(cnp.int_t n_population,
-                                 cnp.int_t n_samples,
+cdef _sample_without_replacement(default_int n_population,
+                                 default_int n_samples,
                                  method="auto",
                                  random_state=None):
     """Sample integers without replacement.
@@ -303,6 +312,32 @@ cpdef sample_without_replacement(cnp.int_t n_population,
                          % (all_methods, method))
 
 
+def sample_without_replacement(
+        object n_population, object n_samples, method="auto", random_state=None):
+    cdef:
+        cnp.intp_t n_pop_intp, n_samples_intp
+        long n_pop_long, n_samples_long
+
+    # On most platforms `np.int_ is np.intp`.  However, before NumPy 2 the
+    # default integer `np.int_` was a long which is 32bit on 64bit windows
+    # while `intp` is 64bit on 64bit platforms and 32bit on 32bit ones.
+    if np.int_ is np.intp:
+        # Branch always taken on NumPy >=2 (or when not on 64bit windows).
+        # Cython has different rules for conversion of values to integers.
+        # For NumPy <1.26.2 AND Cython 3, this first branch requires `int()`
+        # called explicitly to allow e.g. floats.
+        n_pop_intp = int(n_population)
+        n_samples_intp = int(n_samples)
+        return _sample_without_replacement(
+                n_pop_intp, n_samples_intp, method, random_state)
+    else:
+        # Branch taken on 64bit windows with Numpy<2.0 where `long` is 32bit
+        n_pop_long = n_population
+        n_samples_long = n_samples
+        return _sample_without_replacement(
+                n_pop_long, n_samples_long, method, random_state)
+
+
 def _our_rand_r_py(seed):
     """Python utils to test the our_rand_r function"""
     cdef UINT32_t my_seed = seed
diff --git a/sklearn/utils/estimator_checks.py b/sklearn/utils/estimator_checks.py
@@ -3470,7 +3470,7 @@ def param_filter(p):
                 type,
             }
             # Any numpy numeric such as np.int32.
-            allowed_types.update(np.core.numerictypes.allTypes.values())
+            allowed_types.update(np.sctypeDict.values())
 
             allowed_value = (
                 type(init_param.default) in allowed_types

Original file line number	Diff line number	Diff line change
`@@ -124,7 +124,7 @@ def _brute_mst(mutual_reachability, min_samples):`
`124`	`124`	`# Compute the minimum spanning tree for the sparse graph`
`125`	`125`	`sparse_min_spanning_tree = csgraph.minimum_spanning_tree(mutual_reachability)`
`126`	`126`	`rows, cols = sparse_min_spanning_tree.nonzero()`
`127`		`- mst = np.core.records.fromarrays(`
	`127`	`+ mst = np.rec.fromarrays(`
`128`	`128`	`[rows, cols, sparse_min_spanning_tree.data],`
`129`	`129`	`dtype=MST_edge_dtype,`
`130`	`130`	`)`
Original file line number	Diff line number	Diff line change
`@@ -3470,7 +3470,7 @@ def param_filter(p):`
`3470`	`3470`	`type,`
`3471`	`3471`	`}`
`3472`	`3472`	`# Any numpy numeric such as np.int32.`
`3473`		`- allowed_types.update(np.core.numerictypes.allTypes.values())`
	`3473`	`+ allowed_types.update(np.sctypeDict.values())`
`3474`	`3474`
`3475`	`3475`	`allowed_value = (`
`3476`	`3476`	`type(init_param.default) in allowed_types`