Anthony22-dev
diff --git a/‎sklearn/cluster/tests/test_hierarchical.py
Lines changed: 18 additions & 19 deletions b/‎sklearn/cluster/tests/test_hierarchical.py
Lines changed: 18 additions & 19 deletions
@@ -176,10 +176,10 @@ def test_agglomerative_clustering_distances(
         assert not hasattr(clustering, "distances_")
 
 
-def test_agglomerative_clustering():
+def test_agglomerative_clustering(global_random_seed):
     # Check that we obtain the correct number of clusters with
     # agglomerative clustering.
-    rng = np.random.RandomState(0)
+    rng = np.random.RandomState(global_random_seed)
     mask = np.ones([10, 10], dtype=bool)
     n_samples = 100
     X = rng.randn(n_samples, 50)
@@ -280,9 +280,9 @@ def test_agglomerative_clustering_memory_mapped():
     AgglomerativeClustering(metric="euclidean", linkage="single").fit(Xmm)
 
 
-def test_ward_agglomeration():
+def test_ward_agglomeration(global_random_seed):
     # Check that we obtain the correct solution in a simplistic case
-    rng = np.random.RandomState(0)
+    rng = np.random.RandomState(global_random_seed)
     mask = np.ones([10, 10], dtype=bool)
     X = rng.randn(50, 100)
     connectivity = grid_to_graph(*mask.shape)
@@ -330,10 +330,10 @@ def assess_same_labelling(cut1, cut2):
     assert (co_clust[0] == co_clust[1]).all()
 
 
-def test_sparse_scikit_vs_scipy():
+def test_sparse_scikit_vs_scipy(global_random_seed):
     # Test scikit linkage with full connectivity (i.e. unstructured) vs scipy
     n, p, k = 10, 5, 3
-    rng = np.random.RandomState(0)
+    rng = np.random.RandomState(global_random_seed)
 
     # Not using a lil_matrix here, just to check that non sparse
     # matrices are well handled
@@ -370,10 +370,9 @@ def test_sparse_scikit_vs_scipy():
 
 # Make sure our custom mst_linkage_core gives
 # the same results as scipy's builtin
-@pytest.mark.parametrize("seed", range(5))
-def test_vector_scikit_single_vs_scipy_single(seed):
+def test_vector_scikit_single_vs_scipy_single(global_random_seed):
     n_samples, n_features, n_clusters = 10, 5, 3
-    rng = np.random.RandomState(seed)
+    rng = np.random.RandomState(global_random_seed)
     X = 0.1 * rng.normal(size=(n_samples, n_features))
     X -= 4.0 * np.arange(n_samples)[:, np.newaxis]
     X -= X.mean(axis=1)[:, np.newaxis]
@@ -468,13 +467,13 @@ def test_connectivity_propagation():
     ward.fit(X)
 
 
-def test_ward_tree_children_order():
+def test_ward_tree_children_order(global_random_seed):
     # Check that children are ordered in the same way for both structured and
     # unstructured versions of ward_tree.
 
     # test on five random datasets
     n, p = 10, 5
-    rng = np.random.RandomState(0)
+    rng = np.random.RandomState(global_random_seed)
 
     connectivity = np.ones((n, n))
     for i in range(5):
@@ -488,13 +487,13 @@ def test_ward_tree_children_order():
         assert_array_equal(out_unstructured[0], out_structured[0])
 
 
-def test_ward_linkage_tree_return_distance():
+def test_ward_linkage_tree_return_distance(global_random_seed):
     # Test return_distance option on linkage and ward trees
 
     # test that return_distance when set true, gives same
     # output on both structured and unstructured clustering.
     n, p = 10, 5
-    rng = np.random.RandomState(0)
+    rng = np.random.RandomState(global_random_seed)
 
     connectivity = np.ones((n, n))
     for i in range(5):
@@ -726,10 +725,10 @@ def increment(self, *args, **kwargs):
 
 
 @pytest.mark.parametrize("linkage", ["ward", "complete", "average"])
-def test_agglomerative_clustering_with_distance_threshold(linkage):
+def test_agglomerative_clustering_with_distance_threshold(linkage, global_random_seed):
     # Check that we obtain the correct number of clusters with
     # agglomerative clustering with distance_threshold.
-    rng = np.random.RandomState(0)
+    rng = np.random.RandomState(global_random_seed)
     mask = np.ones([10, 10], dtype=bool)
     n_samples = 100
     X = rng.randn(n_samples, 50)
@@ -764,8 +763,8 @@ def test_agglomerative_clustering_with_distance_threshold(linkage):
         assert np.array_equiv(clusters_produced, clusters_at_threshold)
 
 
-def test_small_distance_threshold():
-    rng = np.random.RandomState(0)
+def test_small_distance_threshold(global_random_seed):
+    rng = np.random.RandomState(global_random_seed)
     n_samples = 10
     X = rng.randint(-300, 300, size=(n_samples, 3))
     # this should result in all data in their own clusters, given that
@@ -781,8 +780,8 @@ def test_small_distance_threshold():
     assert clustering.n_clusters_ == n_samples
 
 
-def test_cluster_distances_with_distance_threshold():
-    rng = np.random.RandomState(0)
+def test_cluster_distances_with_distance_threshold(global_random_seed):
+    rng = np.random.RandomState(global_random_seed)
     n_samples = 100
     X = rng.randint(-10, 10, size=(n_samples, 3))
     # check the distances within the clusters and with other clusters