Update app.py

app.py

@@ -6,7 +6,14 @@ import joblib  # For loading the saved preprocessing objects
 
 # Load your pre-trained model
 model = load_model('crack_prediction-2.h5')
+with open('poly_features.pkl', 'rb') as f:
+    poly_loaded = pickle.load(f)
 
+# Load the StandardScaler object
+with open('standard_scaler.pkl', 'rb') as f:
+    scaler_loaded = pickle.load(f)
+
+    
 # Load the preprocessing objects
 poly = PolynomialFeatures(degree=5, include_bias=False)
 scaler = StandardScaler()
@@ -16,17 +23,12 @@ def preprocess_input(flange_width, beam_width, geometric_factor):
     """
     Preprocess the input data: apply polynomial features and then scale.
     """
-    # Create input array
-    input_data = np.reshape( [flange_width, beam_width, geometric_factor], (1,-1))
     
     # # Apply Polynomial Features
-    poly = PolynomialFeatures(degree=5, include_bias=False)
-    X_test_poly = poly.fit_transform(input_data)
+    X_test_poly = poly_loaded.transform([[flange_width, beam_width, geometric_factor]])
 
     # Standardize the polynomial features
-    scaler = StandardScaler()
-    X_train_poly_scaled = scaler.fit_transform(X_test_poly)
-    scaled_features = scaler.transform(X_train_poly_scaled)
+    scaled_features = scaler_loaded.transform(X_train_poly_scaled)
     
     return scaled_features