Update app.py

app.py

@@ -90,7 +90,8 @@ def do_train(n_samples, n_new_data):
     X_new, y_new = make_blobs(
         n_samples=N_NEW_DATA, centers=[(-7, -1), (-2, 4), (3, 6)], random_state=RANDOM_STATE
     )
-
+    X_all = np.concatenate((X, X_new), axis=0)
+    
     fig2, axes2 = plt.subplots()
     axes2.scatter(X[:, 0], X[:, 1], c=cluster_labels, alpha=0.5, edgecolor="k")
     axes2.scatter(X_new[:, 0], X_new[:, 1], c="black", alpha=1, edgecolor="k")
@@ -105,7 +106,7 @@ def do_train(n_samples, n_new_data):
     probable_clusters = inductive_learner.predict(X_new)
     fig3, axes3 = plt.subplots()
     disp = DecisionBoundaryDisplay.from_estimator(
-        inductive_learner, X, response_method="predict", alpha=0.4, ax=axes3
+        inductive_learner, X_all, response_method="predict", alpha=0.4, ax=axes3
     )
     disp.ax_.set_title("Classify unknown instances with known clusters")
     disp.ax_.scatter(X[:, 0], X[:, 1], c=cluster_labels, alpha=0.5, edgecolor="k")
@@ -114,7 +115,6 @@ def do_train(n_samples, n_new_data):
     
     # recomputing clustering and classify boundary
     t2 = time.time()
-    X_all = np.concatenate((X, X_new), axis=0)
     clusterer = AgglomerativeClustering(n_clusters=3)
     y = clusterer.fit_predict(X_all)
     classifier = RandomForestClassifier(random_state=RANDOM_STATE).fit(X_all, y)