Update app.py

app.py

@@ -152,27 +152,27 @@ def get_lagged_stock_data(date_str, tickers=("DHI", "LEN", "^SP500-60"), lag_day
 # 4. PREDICTION FUNCTION (DIPERBAIKI)
 # ==============================================================================
 def predict_sale_amount(assessed_value, town, property_residential, list_year, prediction_date):
+    
     # 0. Safety Check
     if 'model' not in globals() or model is None:
-        print("ERROR: Model belum dimuat. Cek path file model.")
         return 0.0
 
-    print("\n--- START PREDICTION ---")
+    print("\n--- 🔍 START PREDICTION PIPELINE ---")
     
-    # 1. Validasi Input
+    # 1. Validasi Input & Log Transform
     if assessed_value <= 0: assessed_value = 1.0
     assessed_value_log = np.log1p(assessed_value)
     
-    # 2. Ambil Data Saham
+    # 2. Ambil Data Saham (Cached)
     dhi, len_stock, sp500 = get_lagged_stock_data(prediction_date)
     
-    # 3. Mapping Input
+    # 3. Mapping Input User ke Format Training
     clean_town = get_exact_category(town, town_map, fallback='***unknown***')
     clean_prop = get_exact_category(property_residential, prop_map, fallback='Single family')
     clean_year = str(list_year)
 
-    # 4. Buat DataFrame
-    # PASTIKAN NAMA KOLOM INI SAMA PERSIS DENGAN X_TRAIN
+    # 4. Buat DataFrame Input
+    # Pastikan urutan kolom ini SAMA PERSIS dengan X_train
     input_data = pd.DataFrame([[
         assessed_value_log,
         clean_town,
@@ -183,44 +183,50 @@ def predict_sale_amount(assessed_value, town, property_residential, list_year, p
         sp500
     ]], columns=['Assessed Value log', 'Town', 'Property_Residential', 'List Year', 'DHI_lag7', 'LEN_lag7', '^SP500-60_lag7'])
 
-    # 5. Set Kategori
+    # 5. Set Tipe Data Categorical
     input_data['Town'] = pd.Categorical(input_data['Town'], categories=TRAIN_TOWNS)
     input_data['Property_Residential'] = pd.Categorical(input_data['Property_Residential'], categories=TRAIN_PROP_TYPES)
     input_data['List Year'] = pd.Categorical(input_data['List Year'], categories=TRAIN_YEARS)
 
-    print(f"DEBUG: Town Code: {input_data['Town'].cat.codes.iloc[0]}")
+    # =========================================================================
+    # ✨ FITUR BARU: PRINT SEMUA FITUR DENGAN CANTIK
+    # =========================================================================
+    print(f"\n📊 FINAL INPUT FEATURES (Apa yang dilihat Model):")
+    print("="*60)
+    print(f"{'FEATURE NAME':<25} | {'RAW VALUE':<20} | {'FINAL VALUE/CODE'}")
+    print("-" * 60)
+    
+    for col in input_data.columns:
+        raw_val = input_data[col].iloc[0]
+        final_display = raw_val
+        
+        # Jika kolom Kategori, kita tampilkan juga Kode Angkanya
+        if hasattr(input_data[col], 'cat'):
+            code = input_data[col].cat.codes.iloc[0]
+            final_display = f"Code: {code}"
+            # Warning jika code -1 (Unknown)
+            if code == -1: final_display += " ⚠️ (UNKNOWN)"
+            
+        print(f"{col:<25} | {str(raw_val):<20} | {final_display}")
+    
+    print("="*60 + "\n")
+    # =========================================================================
 
-    # --- DEBUGGING TAHAP LANJUT (CEK KOLOM) ---
-    print("DEBUG CHECK: Memeriksa kesesuaian kolom...")
+    # 6. Prediksi
     try:
-        # Cek kolom yang diharapkan model (jika atribut tersedia)
-        if hasattr(model, 'feature_name'):
-            expected_cols = model.feature_name()
-            current_cols = input_data.columns.tolist()
-            
-            print(f"Model expects: {expected_cols}")
-            print(f"We provides:   {current_cols}")
-            
-            if expected_cols != current_cols:
-                print("!!! CRITICAL WARNING: Kolom tidak cocok! Prediksi pasti gagal.")
-        
-        # 6. Lakukan Prediksi
-        print("DEBUG: Executing model.predict()...")
+        print("🤖 Executing Model Prediction...")
         pred_log = model.predict(input_data)[0]
-        print(f"DEBUG: Raw Log Prediction result: {pred_log}")
-        
         pred_val = np.expm1(pred_log)
-        print(f"DEBUG: Final Value: {pred_val}")
+        
+        print(f"✅ Prediction Success!")
+        print(f"   > Log Result: {pred_log:.4f}")
+        print(f"   > Real Price: ${pred_val:,.2f}")
+        print("------------------------------------")
         
         return float(pred_val)
 
     except Exception as e:
-        # Tampilkan Error LENGKAP di Terminal
-        import traceback
-        traceback.print_exc()
-        print(f"ERROR FATAL: {str(e)}")
-        
-        # Return 0.0 agar Gradio tidak blank (tapi kita tahu ada error dari log)
+        print(f"❌ PREDICTION ERROR: {str(e)}")
         return 0.0
 
 # ==============================================================================