Update model_predictor.py

model_predictor.py

@@ -159,65 +159,87 @@ class XGBoostModel:
         return input_df
 
     def predict(self, model_name, input_df):
-        """
-        進行股價漲幅預測
-        
-        Returns:
-            dict: 預測結果，包含各時間點的漲幅百分比
-        """
+    """
+    進行股價漲幅預測（已加入自動對齊模型 feature_names 的邏輯）
+    """
+    try:
+        # 載入模型（如果尚未載入）
+        if self.model is None:
+            model_path = f"{model_name}.json"
+            if not self.load_model(model_path):
+                return None
+
+        # 先做基本的特徵預處理（會補缺失欄位為 0，但不重新排序）
+        processed_df = input_df.copy()
+        processed_df = self.preprocess_features(processed_df.copy())
+
+        # 嘗試從已載入的 xgboost 模型中取得訓練時的 feature names
+        expected_features = None
         try:
-            # 載入模型（如果尚未載入）
-            if self.model is None:
-                model_path = f"{model_name}.json"
-                if not self.load_model(model_path):
-                    return None
-            
-            # 預處理特徵
-            processed_df = self.preprocess_features(input_df.copy())
-            
-            print("=== 模型輸入特徵檢查 ===")
-            print(f"輸入形狀: {processed_df.shape}")
-            print("前5個特徵值:")
-            for i, col in enumerate(processed_df.columns[:5]):
-                print(f"  {col}: {processed_df[col].iloc[0]:.6f}")
-            
-            # 進行預測
-            predictions = self.model.predict(processed_df)
-            print(f"原始預測輸出形狀: {predictions.shape}")
-            print(f"原始預測值: {predictions}")
-            
-            # 【修正】處理多輸出預測結果
-            if predictions.ndim == 1:
-                # 單輸出情況 - 只有一個時間點的預測
-                result = {
-                    'Change_pct_t1_pred': float(predictions[0])
-                }
-            else:
-                # 多輸出情況：[t1, t5, t10, t20] - 對應訓練模型的四個輸出
-                result = {}
-                target_keys = ['Change_pct_t1_pred', 'Change_pct_t5_pred', 
-                              'Change_pct_t10_pred', 'Change_pct_t20_pred']
-                
-                for i, key in enumerate(target_keys):
-                    if i < predictions.shape[1]:
-                        result[key] = float(predictions[0][i])
-                    else:
-                        result[key] = 0.0
-            
-            # 輸出預測結果摘要
-            print("=== 漲幅預測結果 ===")
-            for key, value in result.items():
-                days = key.split('_')[2][1:]
-                direction = "上漲" if value > 0 else "下跌"
-                print(f"  {days}日後預測: {value:+.2f}% ({direction})")
-            
-            return result
-            
-        except Exception as e:
-            print(f"預測過程中發生錯誤：{e}")
-            import traceback
-            traceback.print_exc()
-            return None
+            booster = self.model.get_booster()
+            expected_features = getattr(booster, "feature_names", None)
+        except Exception:
+            expected_features = None
+
+        if expected_features:
+            # 檢查缺失或多餘欄位
+            missing = [f for f in expected_features if f not in processed_df.columns]
+            extra = [f for f in processed_df.columns if f not in expected_features]
+
+            if missing:
+                print(f"警告：模型期待以下特徵但輸入缺失，將以 0 補齊: {missing}")
+                for f in missing:
+                    processed_df[f] = 0.0
+
+            if extra:
+                print(f"注意：輸入含有模型未使用的額外特徵，將忽略: {extra}")
+
+            # 依模型期待欄位順序重排（且只保留 expected_features）
+            processed_df = processed_df[expected_features]
+        else:
+            # 如果模型沒有記錄 feature_names，退回到 class 裡預設的 feature_columns（如有）
+            processed_df = processed_df[self.feature_columns]
+
+        # DEBUG 訊息
+        print("=== 模型輸入特徵檢查（對齊後） ===")
+        print(f"輸入形狀: {processed_df.shape}")
+        print("前5個特徵值:")
+        for i, col in enumerate(processed_df.columns[:5]):
+            print(f"  {col}: {processed_df[col].iloc[0]:.6f}")
+
+        # 進行預測
+        predictions = self.model.predict(processed_df)
+        print(f"原始預測輸出形狀: {getattr(predictions, 'shape', str(type(predictions)))}")
+        print(f"原始預測值: {predictions}")
+
+        # 處理多輸出或單輸出
+        if getattr(predictions, 'ndim', 1) == 1:
+            # 單輸出
+            result = {'Change_pct_t1_pred': float(predictions[0])}
+        else:
+            result = {}
+            target_keys = ['Change_pct_t1_pred', 'Change_pct_t5_pred',
+                           'Change_pct_t10_pred', 'Change_pct_t20_pred']
+            for i, key in enumerate(target_keys):
+                if i < predictions.shape[1]:
+                    result[key] = float(predictions[0][i])
+                else:
+                    result[key] = 0.0
+
+        # 列印摘要
+        print("=== 漲幅預測結果 ===")
+        for key, value in result.items():
+            days = key.split('_')[2][1:]
+            direction = "上漲" if value > 0 else "下跌"
+            print(f"  {days}日後預測: {value:+.2f}% ({direction})")
+
+        return result
+
+    except Exception as e:
+        print(f"預測過程中發生錯誤：{e}")
+        import traceback
+        traceback.print_exc()
+        return None
 
     def predict_single_timeframe(self, model_name, input_df, days):
         """預測特定時間框架的漲幅"""