debug average vs max

ga_calibration_by_one_factory.py

@@ -23,12 +23,24 @@ def clip(val: float, bounds: Tuple[float, float]) -> float:
     return max(lo, min(hi, val))
 
 
+def _match_factory_row(df: pd.DataFrame, cols: dict, factory_code: str) -> pd.Series:
+    """Match factory by English name first, fall back to Chinese name."""
+    eng_col = cols["工厂英文名"]
+    cn_col = cols["工厂中文名"]
+    code_norm = factory_code.strip().lower()
+    mask_eng = df[eng_col].astype(str).str.strip().str.lower() == code_norm
+    if mask_eng.any():
+        return df.loc[mask_eng].iloc[0]
+    mask_cn = df[cn_col].astype(str).str.strip() == factory_code.strip()
+    if mask_cn.any():
+        return df.loc[mask_cn].iloc[0]
+    raise ValueError(f"在{eng_col}/{cn_col} 中找不到工厂 {factory_code}")
+
+
 def load_factory_row(csv_path: Path, factory_code: str) -> tuple[pd.Series, pd.DataFrame, dict]:
     df, cols, _ = read_csv_with_encoding(csv_path, required={"工厂中文名", "工厂英文名", "工厂平均磨合系数", "最小误差"})
-    row = df.loc[df[cols["工厂英文名"]].astype(str).str.strip() == factory_code]
-    if row.empty:
-        raise ValueError(f"在 {csv_path} 中找不到工厂英文名: {factory_code}")
-    return row.iloc[0], df, cols
+    row = _match_factory_row(df, cols, factory_code)
+    return row, df, cols
 
 
 def _csv_lock_path(csv_path: Path) -> Path:
@@ -64,16 +76,14 @@ def update_factory_csv(csv_path: Path, factory_code: str, new_factor: float, new
     fd = _acquire_lock(lock_path)
     try:
         df, cols, enc = read_csv_with_encoding(csv_path, required={"工厂中文名", "工厂英文名", "工厂平均磨合系数", "最小误差"})
-        mask = df[cols["工厂英文名"]].astype(str).str.strip() == factory_code
-        if not mask.any():
-            raise ValueError(f"在 {csv_path} 中找不到工厂英文名: {factory_code}")
+        row = _match_factory_row(df, cols, factory_code)
+        mask = df.index == row.name
         df.loc[mask, cols["工厂平均磨合系数"]] = float(new_factor)
         df.loc[mask, cols["最小误差"]] = float(new_error)
         df.to_csv(csv_path, index=False, encoding=enc)
     finally:
         _release_lock(lock_path, fd)
 
-
 def update_production_line_csv(csv_path: Path, factory_name_cn: str, line_ids: List[str], best_genes: List[float]) -> None:
     lock_path = _csv_lock_path(csv_path)
     fd = _acquire_lock(lock_path)
@@ -105,7 +115,7 @@ def evaluate(factory_code: str, line_ids: List[str], genes: List[float]) -> floa
         model.line_factor[lid] = float(val)
     while model.running:
         model.step()
-    return model.mean_abs_error
+    return model.error
 
 
 def mutate(genes: List[float]) -> List[float]: