Update app.py

app.py

@@ -13,7 +13,6 @@ mi_features = joblib.load("selected_mi_features.pkl")
 roa_features = joblib.load("selected_roa_features.pkl")
 class_names = joblib.load("class_names.pkl")
 
-# Manually defined feature names (from your dataset)
 FEATURE_COLUMNS = [
     "flow_duration", "header_length", "protocol_type", "duration", "rate", "srate", "drate",
     "fin_flag_number", "syn_flag_number", "rst_flag_number", "psh_flag_number", "ack_flag_number",
@@ -33,28 +32,33 @@ default_row = (
     "9.5,10.39230485,0,0,0,141.55"
 )
 
-user_input = st.text_area("Paste the feature row here (no label column)", default_row, height=150)
+user_input = st.text_area("Paste the feature row here (label column optional)", default_row, height=150)
 
 if st.button("🔍 Predict Attack Type"):
     try:
-        # Auto-detect delimiter (tab or comma)
         delimiter = '\t' if '\t' in user_input else ','
 
-        # Parse input row
-        input_list = [float(x.strip()) for x in user_input.strip().split(delimiter)]
+        input_parts = user_input.strip().split(delimiter)
+        
+        # Handle optional label at the end
+        try:
+            _ = float(input_parts[-1])  # If this fails, it's a string label
+        except ValueError:
+            input_parts = input_parts[:-1]  # Drop the label
+
+        input_list = [float(x.strip()) for x in input_parts]
 
         if len(input_list) != len(FEATURE_COLUMNS):
-            st.error(f"🚫 Expected {len(FEATURE_COLUMNS)} values but got {len(input_list)}.")
+            st.error(f"🚫 Expected {len(FEATURE_COLUMNS)} features but got {len(input_list)}.")
         else:
             input_df = pd.DataFrame([input_list], columns=FEATURE_COLUMNS)
 
-            # Preprocessing
+            # Preprocessing pipeline
             scaled = scaler.transform(input_df)
             var_filtered = var_thresh.transform(scaled)
             mi_selected = pd.DataFrame(var_filtered, columns=np.array(FEATURE_COLUMNS)[var_thresh.get_support()])
             final_features = mi_selected[mi_features].iloc[:, roa_features]
 
-            # Prediction
             prediction = model.predict(final_features)
             predicted_label = label_encoder.inverse_transform(prediction)[0]