Update app.py

app.py

@@ -9,12 +9,8 @@ from pyod.models.combination import aom, moa, average
 from pyod.utils.utility import standardizer
 from sklearn.decomposition import PCA
 from sklearn.metrics import precision_score, recall_score
-import base64
 from datetime import datetime
 
-# Configuration
-st.set_option('deprecation.showPyplotGlobalUse', False)
-
 def generate_report(data, predictions, model_names, metrics):
     report = f"""
     Network Anomaly Detection Report
@@ -37,11 +33,10 @@ def generate_report(data, predictions, model_names, metrics):
     return report
 
 def plot_3d_projections(data, predictions):
-    pca = PCA(n_components=3)
-    projections = pca.fit_transform(data)
-    
     fig = plt.figure(figsize=(10, 7))
     ax = fig.add_subplot(111, projection='3d')
+    pca = PCA(n_components=3)
+    projections = pca.fit_transform(data)
     
     normal = projections[predictions == 0]
     anomalies = projections[predictions == 1]
@@ -121,18 +116,22 @@ def main():
         progress_bar.progress((i+1)/len(selected_models))
 
     # Ensemble prediction
-    if ensemble_method == "Average":
-        combined_scores = average(train_scores)
-    elif ensemble_method == "MOA":
-        combined_scores = moa(train_scores)
-    else:
-        combined_scores = aom(train_scores)
+    try:
+        if ensemble_method == "Average":
+            combined_scores = average(train_scores)
+        elif ensemble_method == "MOA":
+            combined_scores = moa(train_scores)
+        else:
+            combined_scores = aom(train_scores)
+    except Exception as e:
+        st.error(f"Ensemble method failed: {str(e)}")
+        return
 
     threshold = np.percentile(combined_scores, 100*(1-contamination))
     predictions = (combined_scores > threshold).astype(int)
 
     # Performance metrics
-    if uploaded_file is None:  # Use synthetic ground truth
+    if uploaded_file is None:
         y_true = np.zeros(n_samples)
         y_true[anomaly_idx] = 1
         precision = precision_score(y_true, predictions)
@@ -161,18 +160,18 @@ def main():
     
     tab1, tab2 = st.tabs(["2D Projection", "3D Projection"])
     with tab1:
+        fig, ax = plt.subplots(figsize=(10, 6))
         pca = PCA(n_components=2)
         viz_data = pca.fit_transform(X_norm)
-        plt.figure(figsize=(10, 6))
-        plt.scatter(viz_data[predictions==0, 0], viz_data[predictions==0, 1],
+        ax.scatter(viz_data[predictions==0, 0], viz_data[predictions==0, 1],
                     c='blue', label='Normal', alpha=0.6)
-        plt.scatter(viz_data[predictions==1, 0], viz_data[predictions==1, 1],
+        ax.scatter(viz_data[predictions==1, 0], viz_data[predictions==1, 1],
                     c='red', marker='x', label='Anomaly')
-        plt.xlabel("Principal Component 1")
-        plt.ylabel("Principal Component 2")
-        plt.title("PCA Projection of Network Data")
-        plt.legend()
-        st.pyplot()
+        ax.set_xlabel("Principal Component 1")
+        ax.set_ylabel("Principal Component 2")
+        ax.set_title("PCA Projection of Network Data")
+        ax.legend()
+        st.pyplot(fig)
     
     with tab2:
         st.pyplot(plot_3d_projections(X_norm, predictions))