Test usage of Density Estimator in Silhouette visualizer

tests/test_cluster/test_silhouette.py

@@ -25,6 +25,7 @@
 from sklearn.datasets import make_blobs
 from sklearn.cluster import KMeans, MiniBatchKMeans
 from sklearn.cluster import SpectralClustering, AgglomerativeClustering
+from sklearn.mixture import GaussianMixture
 
 from unittest import mock
 from tests.base import VisualTestCase
@@ -84,6 +85,27 @@ def test_integrated_mini_batch_kmeans_silhouette(self):
 
         self.assert_images_similar(visualizer, remove_legend=True)
 
+    @pytest.mark.xfail(sys.platform == "win32", reason="images not close on windows")
+    def test_integrated_gaussian_mixture_silhouette(self):
+        """
+        Test Density Estimator works with silhouette visualizer
+        """
+        # NOTE see #182: cannot use occupancy dataset because of memory usage
+
+        # Generate a blobs data set
+        X, y = make_blobs(
+            n_samples=1000, n_features=12, centers=8, shuffle=False, random_state=0
+        )
+
+        fig = plt.figure()
+        ax = fig.add_subplot()
+
+        visualizer = SilhouetteVisualizer(GaussianMixture(n_components=5, random_state=0), ax=ax)
+        visualizer.fit(X)
+        visualizer.finalize()
+
+        self.assert_images_similar(visualizer, remove_legend=True)
+
     @pytest.mark.skip(reason="no negative silhouette example available yet")
     def test_negative_silhouette_score(self):
         """

yellowbrick/cluster/silhouette.py

@@ -188,9 +188,9 @@ def fit(self, X, y=None, **kwargs):
         # Compute the number of available clusters from the estimator
         if hasattr(self.estimator, "n_clusters"):
             self.n_clusters_ = self.estimator.n_clusters
-	elif hasattr(self.estimator, "n_components"):
-	    self.n_clusters_ = self.estimator.n_components
-	else:
+        elif hasattr(self.estimator, "n_components"):
+	        self.n_clusters_ = self.estimator.n_components
+        else:
             unique_labels = set(labels)
             n_noise_clusters = 1 if -1 in unique_labels else 0
             self.n_clusters_ = len(unique_labels) - n_noise_clusters