MAINT Fix build when SKLEARN_OPENMP_PARALLELISM_ENABLED=False (scikit-learn#24682)

Co-authored-by: Guillaume Lemaitre <[email protected]>

Author: lesteve

doc/developers/performance.rst

@@ -341,9 +341,16 @@ memory alignment, direct blas calls...
 Using OpenMP
 ------------
 
-Since scikit-learn can be built without OpenMP, it's necessary to
-protect each direct call to OpenMP. This can be done using the following
-syntax::
+Since scikit-learn can be built without OpenMP, it's necessary to protect each
+direct call to OpenMP.
+
+There are some helpers in
+[sklearn/utils/_openmp_helpers.pyx](https://github.com/scikit-learn/scikit-learn/blob/main/sklearn/utils/_openmp_helpers.pyx)
+that you can reuse for the main useful functionalities and already have the
+necessary protection to be built without OpenMP.
+
+If the helpers are not enough, you need to protect your OpenMP code using the
+following syntax::
 
   # importing OpenMP
   IF SKLEARN_OPENMP_PARALLELISM_ENABLED:

sklearn/metrics/_pairwise_distances_reduction/_base.pyx.tp

@@ -1,5 +1,4 @@
 cimport numpy as cnp
-cimport openmp
 
 from cython cimport final
 from cython.operator cimport dereference as deref
@@ -73,7 +72,8 @@ cpdef DTYPE_t[::1] _sqeuclidean_row_norms32(
         )
 
     with nogil, parallel(num_threads=num_threads):
-        thread_num = openmp.omp_get_thread_num()
+        thread_num = _openmp_thread_num()
+
         for i in prange(n, schedule='static'):
             # Upcasting the i-th row of X from float32 to float64
             for j in range(d):