feat: ML模型波长记忆 + 跨传感器光谱重采样 + 数据加载规范化

训练端 (modeling_batch.py):
- _extract_train_wavelengths() 从列名提取训练波长写入 metadata
- load_data_batch/load_data_single 改为基于列名语义智能提取特征
- 废除 feature_start_column 硬编码位置索引

推理端 (inference_batch.py):
- preprocess_spectra() 分支A: train_wavelengths存在→np.interp精确重采样
- preprocess_spectra() 分支B: 无波长元数据→解析列名→np.interp到400-800nm标准网格
- 废除暴力截断和零值填充

src/gui/panels/step8_ml_train_panel.py

@@ -528,16 +528,14 @@ class Step8MlTrainPanel(QWidget):
 
     def get_training_params(self):
         """获取模型训练参数"""
-        # ★ 2026-07-01 安全保护：float() 前验证 currentText() 是否为有效数值
+        # ★ v2：currentText() 为列名字符串，传 None 走智能识别
         feature_text = self.feature_start.currentText()
-        try:
-            feature_start = float(feature_text)
-        except (ValueError, TypeError):
-            feature_start = 374.285004  # 默认光谱起始列
+        if not feature_text or feature_text.startswith("（"):
+            feature_text = None
 
         return {
             'pipeline_type': 'machine_learning',
-            'feature_start': feature_start,
+            'feature_start': feature_text,
             'cv_folds': self.cv_folds.value(),
             'preprocess_methods': [method for method, cb in self.preproc_checkboxes.items() if cb.isChecked()],
             'model_types': [model for model, cb in self.model_checkboxes.items() if cb.isChecked()],