fix #356 猫站数据统计问题

2023-09-02 07:57:44 +08:00
parent 3c753686c6
commit ec8c9c996a
3 changed files with 13 additions and 7 deletions
--- a/app/modules/indexer/spider.py
+++ b/app/modules/indexer/spider.py
@ -262,7 +262,12 @@ class TorrentSpider:
                        # 解码为字符串
                        page_source = raw_data.decode(encoding)
                    except Exception as e:
-                        logger.error(f"chardet解码失败：{e}")
+                        logger.debug(f"chardet解码失败：{e}")
+                        # 探测utf-8解码
+                        if re.search(r"charset=\"?utf-8\"?", ret.text, re.IGNORECASE):
+                            ret.encoding = "utf-8"
+                        else:
+                            ret.encoding = ret.apparent_encoding
                        page_source = ret.text
                else:
                    page_source = ret.text