Update app.py

app.py

@@ -44,39 +44,31 @@ selected_features =  [
 ]
 
 def extract_features(sequence):
+    """Extract selected features and normalize them."""
+    
+    # Extract features from different methods
     aa_features = AAComposition.CalculateAADipeptideComposition(sequence)
     auto_features = Autocorrelation.CalculateAutoTotal(sequence)
     ctd_features = CTD.CalculateCTD(sequence)
     pseaac_features = PseudoAAC.GetAPseudoAAC(sequence, lamda=9)
 
+    # Combine all extracted features
     all_features = {**aa_features, **auto_features, **ctd_features, **pseaac_features}
 
-    # Convert to DataFrame
-    feature_df = pd.DataFrame([all_features])
-
-    print("Extracted Features:", feature_df.columns.tolist())  # Debugging line
-
     # Ensure all selected features are present
-    missing_features = [f for f in selected_features if f not in feature_df.columns]
-    extra_features = [f for f in feature_df.columns if f not in selected_features]
-
-    if missing_features:
-        print(f"Missing Features ({len(missing_features)}):", missing_features)
+    feature_dict = {feature: all_features.get(feature, 0) for feature in selected_features}
     
-    if extra_features:
-        print(f"Extra Features ({len(extra_features)}):", extra_features)
+    # Convert to DataFrame
+    feature_df = pd.DataFrame([feature_dict])
 
-    # Fix missing columns by adding them with default values (0)
-    for feature in missing_features:
-        feature_df[feature] = 0
+    # Normalize the features
+    normalized_features = scaler.transform(feature_df)
 
-    # Select only the required features
-    feature_df = feature_df[selected_features]
+    # Convert to a NumPy array in the expected format
+    selected_feature_array = normalized_features.flatten().reshape(1, -1)
 
-    # Normalize
-    normalized_features = scaler.transform(feature_df)
+    return selected_feature_array
 
-    return normalized_features
 
 
 def predict(sequence):