fix(tests): handle error case in clustering tests

Author: codie3611

test/unit_tests/dimred/svd/test_clustering_betas.py

@@ -47,11 +47,87 @@ def test_group_betas(self):
 
         for cluster_type in ClusterType.get_cluster_type_name():
             for detector_type in DetectorType.get_detector_type_name():
-                beta_clusters, name_clusters = group_betas(
-                    fake_names, fake_betas, cluster=cluster_type, detector=detector_type
+                
+                # As some clustering algorithms require parameters, we need to provide them
+                # to avoid errors
+                if cluster_type == "KMeans" or cluster_type == "SpectralClustering":
+                    cluster_params = {"n_clusters": 2}
+                else:
+                    cluster_params = {}
+                
+                result = group_betas(
+                    fake_names,
+                    fake_betas,
+                    cluster=cluster_type,
+                    detector=detector_type,
+                    cluster_params=cluster_params,
                 )
 
+                if isinstance(result, str):
+                    self.fail(f"group_betas returned an error: {result}")
+                
+                beta_clusters, name_clusters = result
+
                 # verify correct output
                 self.assertIsInstance(beta_clusters, list)
                 self.assertIsInstance(name_clusters, list)
                 self.assertEqual(len(beta_clusters), len(name_clusters))
+
+
+    def test_group_betas_no_detector(self):
+        """tests correct function of the group_betas function
+        in clustering_betas.py without an outlier detector"""
+
+        fake_names = np.array([f"betas_{i}" for i in range(25)])
+        fake_cluster_0 = np.random.rand(12, 3) + 5
+        fake_cluster_1 = np.random.rand(12, 3) - 5
+        fake_betas = np.stack([*fake_cluster_0, *fake_cluster_1, np.array([0, 0, 0])])
+        expected_clusters = 2
+
+        # test with recommended settings
+        beta_clusters, name_clusters = group_betas(
+            fake_names,
+            fake_betas,
+            cluster=ClusterType.KMeans,
+            cluster_params={"n_clusters": expected_clusters},
+        )
+
+        # verify correct type of output
+        self.assertIsInstance(beta_clusters, list)
+        self.assertIsInstance(name_clusters, list)
+
+        # verify that beta_clusters and name_clusters correspond to each other
+        self.assertEqual(len(beta_clusters), len(name_clusters))
+        # verify that beta_clusters contains as many clusters as searched for
+        self.assertEqual(len(beta_clusters), expected_clusters)
+
+        # verify that entries correspond to each other
+        for c, cluster in enumerate(name_clusters):
+            for e, entry in enumerate(cluster):
+                index = np.where(fake_names == entry)[0]
+                self.assertTrue((fake_betas[index] - beta_clusters[c][e]).max() == 0)
+
+        # verify different keyword combinations
+
+        for cluster_type in ClusterType.get_cluster_type_name():
+
+            # As some clustering algorithms require parameters, we need to provide them
+            # to avoid errors
+            if cluster_type == "KMeans" or cluster_type == "SpectralClustering":
+                cluster_params = {"n_clusters": 2}
+            else:
+                cluster_params = {}
+
+            result = group_betas(
+                fake_names, fake_betas, cluster=cluster_type, cluster_params=cluster_params
+            )
+
+            if isinstance(result, str):
+                self.fail(f"group_betas returned an error: {result}")
+
+            beta_clusters, name_clusters = result
+
+            # verify correct output
+            self.assertIsInstance(beta_clusters, list)
+            self.assertIsInstance(name_clusters, list)
+            self.assertEqual(len(beta_clusters), len(name_clusters))