优化表情异步补处理的参数提取与下载重试

- 对表情 XML 先做 HTML 实体还原，避免下载地址中的 & 影响请求
- 优先使用 XML 节点提取 aeskey、md5、len 和多种下载地址，正则作为补充兜底
- 异步补处理时按 encrypturl、cdnurl、thumburl 顺序重试下载
- 所有地址失败时记录业务告警，避免无意义异常栈刷屏

utils/wechat/message_to_db.py

@@ -3,6 +3,7 @@ import time
 from datetime import datetime, timedelta
 import xml.etree.ElementTree as ET
 import concurrent.futures  # 添加线程池支持
+import html
 import os
 import base64
 import re
@@ -126,26 +127,61 @@ class MessageStorage:
         if not xml_content:
             return {}
 
-        aeskey_match = self._aeskey_re.search(xml_content)
-        if not aeskey_match:
-            return {}
+        normalized_xml = html.unescape(xml_content)
+        aeskey = ""
+        md5 = ""
+        length = 0
+        urls = []
 
-        url_match = (
-            self._emoji_cdn_re.search(xml_content)
-            or self._emoji_encrypt_re.search(xml_content)
-            or self._emoji_thumb_re.search(xml_content)
-        )
-        if not url_match:
-            return {}
+        try:
+            root = ET.fromstring(normalized_xml)
+            emoji_node = root.find(".//emoji")
+            if emoji_node is not None:
+                aeskey = (emoji_node.attrib.get("aeskey") or "").strip()
+                md5 = (emoji_node.attrib.get("md5") or "").strip()
+                try:
+                    length = int((emoji_node.attrib.get("len") or "0").strip() or 0)
+                except Exception:
+                    length = 0
+                urls.extend([
+                    (emoji_node.attrib.get("encrypturl") or "").strip(),
+                    (emoji_node.attrib.get("cdnurl") or "").strip(),
+                    (emoji_node.attrib.get("thumburl") or "").strip(),
+                ])
+        except Exception:
+            pass
 
-        md5_match = re.search(r'md5="(.*?)"', xml_content)
-        length_match = re.search(r'len="(\d+)"', xml_content)
+        if not aeskey:
+            aeskey_match = self._aeskey_re.search(normalized_xml)
+            aeskey = aeskey_match.group(1).strip() if aeskey_match else ""
+
+        if not md5:
+            md5_match = re.search(r'md5="(.*?)"', normalized_xml)
+            md5 = md5_match.group(1).strip() if md5_match else ""
+
+        if not length:
+            length_match = re.search(r'len="(\d+)"', normalized_xml)
+            length = int(length_match.group(1)) if length_match else 0
+
+        encrypt_match = self._emoji_encrypt_re.search(normalized_xml)
+        cdn_match = self._emoji_cdn_re.search(normalized_xml)
+        thumb_match = self._emoji_thumb_re.search(normalized_xml)
+        urls.extend([
+            encrypt_match.group(1).strip() if encrypt_match else "",
+            cdn_match.group(1).strip() if cdn_match else "",
+            thumb_match.group(1).strip() if thumb_match else "",
+        ])
+        urls = [url for index, url in enumerate(urls) if url and url not in urls[:index]]
+
+        if not aeskey or not urls:
+            return {}
 
         return {
-            "aeskey": aeskey_match.group(1),
-            "url": url_match.group(1),
-            "md5": md5_match.group(1) if md5_match else "",
-            "length": int(length_match.group(1)) if length_match else 0,
+            "aeskey": aeskey,
+            "url": urls[0],
+            "urls": urls,
+            "md5": md5,
+            "length": length,
         }
 
     async def _process_emoji_record(self, msg_record: Dict) -> bool:
@@ -163,8 +199,24 @@ class MessageStorage:
             return False
 
         try:
-            base64_str = await self.client.download_cdn_file(emoji_info["aeskey"], emoji_info["url"])
+            base64_str = None
+            last_error = None
+            for file_url in emoji_info.get("urls", []) or [emoji_info.get("url", "")]:
+                if not file_url:
+                    continue
+                try:
+                    base64_str = await self.client.download_cdn_file(emoji_info["aeskey"], file_url)
+                    if base64_str:
+                        break
+                except Exception as e:
+                    last_error = e
+                    logger.warning(
+                        f"表情下载地址尝试失败: msg_id={message_id}, url={file_url[:120]}, error={e}"
+                    )
             if not base64_str:
+                if last_error:
+                    logger.warning(f"表情下载全部地址均失败: msg_id={message_id}, error={last_error}")
+                    return False
                 logger.warning(f"表情下载返回为空: msg_id={message_id}")
                 return False