Update app.py

app.py

@@ -20,26 +20,15 @@ raw_columns = [
     'ct_dst_sport_ltm', 'ct_dst_src_ltm', 'attack_cat', 'Label'
 ]
 
-# Define the columns that will be added with placeholder values
-missing_columns = ['srcip', 'dstip']
-
-# Function to create a dataframe with missing columns filled with placeholder values
+# Function to preprocess a single input row
 def preprocess_input(row_values):
-    # Ensure the row has 49 values
     if len(row_values) != 49:
         raise ValueError(f"❌ Expected 49 values, but got {len(row_values)}.")
     
-    # Create the DataFrame with the input values
+    # Create DataFrame from input
     input_df = pd.DataFrame([row_values], columns=raw_columns)
 
-    # Add the missing columns with placeholder values (like NaN)
-    for col in missing_columns:
-        input_df[col] = np.nan  # Fill missing columns with NaN (or zero if needed)
-
-    # Ensure that all columns are in the same order as the trained model expects
-    input_df = input_df[raw_columns]
-
-    # Convert columns to numeric where applicable
+    # Convert all columns to numeric
     input_df = input_df.apply(pd.to_numeric, errors='coerce')
 
     # Feature engineering
@@ -47,8 +36,8 @@ def preprocess_input(row_values):
     input_df['byte_ratio'] = input_df['sbytes'] / (input_df['dbytes'] + 1)
     input_df['pkt_ratio'] = input_df['Spkts'] / (input_df['Dpkts'] + 1)
 
-    # Drop correlated features (e.g., ones with high correlation)
-    input_df = input_df.drop(columns=features_to_drop, errors='ignore')
+    # Drop unused or label columns
+    input_df = input_df.drop(columns=features_to_drop + ['attack_cat', 'Label'], errors='ignore')
 
     return input_df
 
@@ -63,7 +52,7 @@ if st.button("Predict"):
         # Parse the input
         values = user_input.strip().split("\t")
 
-        # Preprocess the input (adding missing columns and performing feature engineering)
+        # Preprocess the input row
         processed_df = preprocess_input(values)
 
         # Predict using the preprocessed data