全局修正

src/core/prediction/inference_batch.py

@@ -12,7 +12,7 @@ warnings.filterwarnings('ignore')
 import sys
 import os
 
-from src.preprocessing.spectral_Preprocessing import Preprocessing
+from src.preprocessing.spectral_Preprocessing import Preprocessing, get_preprocessing_transformer
 from src.core.utils.split_methods import spxy, ks
 
 # try:
@@ -22,6 +22,7 @@ from src.core.utils.split_methods import spxy, ks
 
 # 机器学习相关导入
 from sklearn.model_selection import train_test_split
+from sklearn.pipeline import Pipeline
 
 
 class WaterQualityInference:
@@ -423,33 +424,20 @@ class WaterQualityInference:
                 from src.utils.water_index import WaterQualityIndexCalculator
                 calc = WaterQualityIndexCalculator()
 
-                # 提取纯计算方法（排除 find_closest_wavelength 和 calculate_all_indices，
-                # 以及不返回 Series 的辅助方法）
-                algorithm_methods = []
-                for m in dir(calc):
-                    if m.startswith('_'):
-                        continue
-                    if m in ['find_closest_wavelength', 'calculate_all_indices']:
-                        continue
-                    attr = getattr(calc, m)
-                    if callable(attr):
-                        algorithm_methods.append(m)
-
-                original_col_count = spectra.shape[1]
-                for algo_name in algorithm_methods:
-                    try:
-                        algo_func = getattr(calc, algo_name)
-                        result = algo_func(spectra)
-                        # 只追加返回 Series 且长度为样本数的合法结果
-                        if isinstance(result, pd.Series) and len(result) == len(spectra):
-                            spectra[algo_name] = result.values
-                        else:
-                            spectra[algo_name] = np.nan
-                    except Exception:
-                        spectra[algo_name] = np.nan
-
-                print(f"[特征补全] 完成！光谱列已扩充至 {spectra.shape[1]} 列"
-                      f"（追加了 {spectra.shape[1] - original_col_count} 个 WQI 指数）")
+                # CSV 驱动的 WaterQualityIndexCalculator：所有公式名通过 list_available() 拿；
+                # 一次性 calculate_many() 批量计算。彻底摆脱 dir(calc) 反射扫描 + 单 algo_func
+                # 调用这种碎片化写法（Calculator 早已重构为公式驱动，不再有独立公式方法）。
+                formulas = calc.list_available()
+                if not formulas:
+                    print("[特征补全] Calculator 未持有任何公式，跳过补全")
+                else:
+                    results_df = calc.calculate_many(formulas, spectra)
+                    # results_df 是列对齐的 WQI 计算结果（每列一个公式，行数=样本数）
+                    if isinstance(results_df, pd.DataFrame) and not results_df.empty:
+                        original_col_count = spectra.shape[1]
+                        spectra = pd.concat([spectra, results_df], axis=1)
+                        print(f"[特征补全] 完成！光谱列已扩充至 {spectra.shape[1]} 列"
+                              f"（追加了 {spectra.shape[1] - original_col_count} 个 WQI 指数）")
             except Exception as e:
                 print(f"[特征补全] 失败，将使用原始光谱特征: {e}")
 
@@ -471,6 +459,14 @@ class WaterQualityInference:
 
         print(f"[特征对齐] 最终输入维度: {spectra.shape}")
 
+        # ---- Pipeline 化分支：模型内置 scaler/MSC.mean_spectrum_ 等状态时，跳过手动 Preprocessing ----
+        if isinstance(model, Pipeline):
+            print(f"[Pipeline] 检测到模型是 sklearn Pipeline，"
+                  f"其内置预处理步骤（{list(model.named_steps.keys())[0]}）将处理原始光谱，"
+                  f"无需外部 Preprocessing")
+            return spectra.values
+
+        # ---- 兼容路径：旧 .joblib（裸模型 + preprocess_method 字符串）回退手动 Preprocessing ----
         try:
             # 应用预处理
             spectra_processed = Preprocessing(actual_preprocess_method, spectra)
@@ -479,7 +475,8 @@ class WaterQualityInference:
             if isinstance(spectra_processed, pd.DataFrame):
                 spectra_processed = spectra_processed.values
 
-            print(f"预处理后数据形状: {spectra_processed.shape}")
+            print(f"  [Legacy] 旧裸模型 + 手动 Preprocessing({actual_preprocess_method}) 完成，"
+                  f"数据形状: {spectra_processed.shape}")
 
             return spectra_processed