fix: 支持CSV自动识别UTF-8和GBK编码

app/warehouse.py

@@ -182,16 +182,31 @@ class MetrixWarehouse:
             raise RuntimeError(str(result))
         return int(result.get("affected_rows", 0)) if isinstance(result, dict) else 0
 
-    def import_csv(self, file_path: str, table_name: str) -> int:
+    def import_csv(self, file_path: str, table_name: str, encoding: str | None = None) -> int:
         """走 Metrix 平台导入 API 追加导入到已存在的表（字段映射由平台按列名处理）。"""
-        job_id = self.client.import_csv(
-            self.conn_id, table_name, Path(file_path), mode="append",
-            database=self.database, create_table=False,
-        )
-        job = self.client.wait_job(job_id)
-        if job.get("status") != "success":
-            raise RuntimeError(job.get("error_code") or job.get("status") or "导入失败")
-        return int(job.get("row_count") or 0)
+        source_path = Path(file_path)
+        upload_path = source_path
+        converted_path: Path | None = None
+        if encoding and encoding.lower() not in {"utf-8", "utf-8-sig"}:
+            converted_path = source_path.with_name(f"{source_path.stem}_utf8{source_path.suffix}")
+            with source_path.open("r", encoding=encoding, newline="") as source:
+                with converted_path.open("w", encoding="utf-8-sig", newline="") as target:
+                    while chunk := source.read(1024 * 1024):
+                        target.write(chunk)
+            upload_path = converted_path
+
+        try:
+            job_id = self.client.import_csv(
+                self.conn_id, table_name, upload_path, mode="append",
+                database=self.database, create_table=False,
+            )
+            job = self.client.wait_job(job_id)
+            if job.get("status") != "success":
+                raise RuntimeError(job.get("error_code") or job.get("status") or "导入失败")
+            return int(job.get("row_count") or 0)
+        finally:
+            if converted_path is not None:
+                converted_path.unlink(missing_ok=True)
 
     def execute_sql(self, sql: str) -> Tuple[bool, Any]:
         try: