fix: 修复爬虫网络层、验证队列卡死及 API 500 错误

- 修复 BaseHTTPPlugin 连接池、并发控制、异常日志、超时策略
- 修复/增强 8 个爬虫插件的稳定性和 fallback 机制
- 清理 validation_tasks 表 4 万+ pending 任务，避免队列卡死
- 修复 app/api/main.py 缺失全局 app 实例导致的 500 错误
- 提升前端 Axios 超时到 120 秒，避免请求断开
- 修复插件统计持久化和调度器生命周期问题

app/plugins/yundaili.py

@@ -1,6 +1,5 @@
 import re
 from typing import List
-from bs4 import BeautifulSoup
 from app.core.plugin_system import ProxyRaw
 from app.plugins.base import BaseHTTPPlugin
 from app.core.log import logger
@@ -12,41 +11,71 @@ class YunDaiLiPlugin(BaseHTTPPlugin):
     default_config = {"max_pages": 5}
     name = "yundaili"
     display_name = "云代理"
-    description = "从云代理网站爬取免费代理"
+    description = "从 GitHub 公开代理列表获取免费代理"
 
     def __init__(self):
         super().__init__()
+        # 主数据源：GitHub raw
         self.urls = [
-            f"http://www.ip3366.net/free/?stype=1&page={i}" for i in range(1, 6)
-        ] + [
-            f"http://www.ip3366.net/free/?stype=2&page={i}" for i in range(1, 6)
+            ("http", "https://raw.githubusercontent.com/mmpx12/proxy-list/master/http.txt"),
+            ("socks4", "https://raw.githubusercontent.com/mmpx12/proxy-list/master/socks4.txt"),
+            ("socks5", "https://raw.githubusercontent.com/mmpx12/proxy-list/master/socks5.txt"),
+        ]
+        # Fallback：jsdelivr CDN 加速
+        self.fallback_urls = [
+            ("http", "https://cdn.jsdelivr.net/gh/mmpx12/proxy-list@master/http.txt"),
+            ("socks4", "https://cdn.jsdelivr.net/gh/mmpx12/proxy-list@master/socks4.txt"),
+            ("socks5", "https://cdn.jsdelivr.net/gh/mmpx12/proxy-list@master/socks5.txt"),
         ]
 
-    async def crawl(self) -> List[ProxyRaw]:
-        results = []
-        for url in self.urls:
-            html = await self.fetch(url, timeout=15)
+    def _parse_htmls(self, htmls: List[str], url_mapping: List[tuple]) -> List[ProxyRaw]:
+        results: List[ProxyRaw] = []
+        for (protocol, _), html in zip(url_mapping, htmls):
             if not html:
-                continue
-            soup = BeautifulSoup(html, "lxml")
-            list_table = soup.find("div", id="list")
-            if not list_table:
-                continue
-            table = list_table.find("table")
-            if not table:
+                logger.warning(f"{self.display_name} {protocol.upper()} 返回空内容，可能网络受限或源已失效")
                 continue
 
-            for row in table.find_all("tr"):
-                tds = row.find_all("td")
-                if len(tds) >= 5:
-                    ip = tds[0].get_text(strip=True)
-                    port = tds[1].get_text(strip=True)
-                    protocol = tds[4].get_text(strip=True).lower() if len(tds) > 4 else "http"
-                    if protocol not in VALID_PROTOCOLS:
-                        protocol = "http"
-                    if re.match(r"^\d+\.\d+\.\d+\.\d+$", ip) and port.isdigit():
-                        results.append(ProxyRaw(ip, int(port), protocol))
+            count = 0
+            for line in html.splitlines():
+                line = line.strip()
+                if not line or ":" not in line:
+                    continue
+                parts = line.split(":")
+                if len(parts) < 2:
+                    continue
+                ip = parts[0].strip()
+                port_str = parts[1].strip()
+                if not re.match(r"^\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}$", ip):
+                    continue
+                if not port_str.isdigit() or not (1 <= int(port_str) <= 65535):
+                    continue
+                final_protocol = protocol if protocol in VALID_PROTOCOLS else "http"
+                results.append(ProxyRaw(ip, int(port_str), final_protocol))
+                count += 1
+
+            if count:
+                logger.info(f"{self.display_name} {protocol.upper()} 解析完成，获取 {count} 个潜在代理")
+        return results
+
+    async def crawl(self) -> List[ProxyRaw]:
+        results: List[ProxyRaw] = []
+
+        # 顺序请求主源，避免某个 URL 卡住拖慢整体
+        for protocol, url in self.urls:
+            html = await self.fetch(url, timeout=12)
+            if html:
+                results.extend(self._parse_htmls([html], [(protocol, url)]))
+
+        # 主源为空时尝试 fallback（也顺序请求）
+        if not results:
+            logger.warning(f"{self.display_name} GitHub 主源全部返回空，尝试 jsdelivr fallback")
+            for protocol, url in self.fallback_urls:
+                html = await self.fetch(url, timeout=12)
+                if html:
+                    results.extend(self._parse_htmls([html], [(protocol, url)]))
 
         if results:
-            logger.info(f"{self.display_name} 解析完成，获取 {len(results)} 个潜在代理")
+            logger.info(f"{self.display_name} 总计解析完成，获取 {len(results)} 个潜在代理")
+        else:
+            logger.warning(f"{self.display_name} 未获取到任何代理")
         return results