From 093424c6e0c74d271198903c8366e580a73787ed Mon Sep 17 00:00:00 2001
From: filesite <machete@filesite.io>
Date: Mon, 11 Sep 2023 11:01:56 +0800
Subject: [PATCH] bot for xigua done

---
 bot/Xigua.mjs       | 49 +++++++++++++++++++++++++++++++++++++++++++++
 test/scrap_test.mjs | 22 +++++++++++++++-----
 2 files changed, 66 insertions(+), 5 deletions(-)
 create mode 100644 bot/Xigua.mjs

diff --git a/bot/Xigua.mjs b/bot/Xigua.mjs
new file mode 100644
index 0000000..8f08e4e
--- /dev/null
+++ b/bot/Xigua.mjs
@@ -0,0 +1,49 @@
+import Hero from '@ulixee/hero';
+import configs from '../config.mjs';
+
+class Xigua {
+  constructor(heroCloudServer) {
+    this.heroServer = heroCloudServer ? heroCloudServer : '';
+  }
+
+  async scrap(url) {
+    let data = {};
+
+    try {
+        let options = {};
+        if (this.heroServer) {
+            options.connectionToCore = this.heroServer;
+        }
+
+        const hero = new Hero(options);
+        await hero.goto(url, configs.heroBotOptions);
+
+        //等待所有内容加载完成
+        const tab = await hero.activeTab;
+        await tab.waitForLoad('AllContentLoaded', {timeoutMs: configs.heroTabOptions.timeoutMs});
+
+        //解析网页HTML数据
+        const elems = await hero.detach( hero.document.querySelectorAll('meta') );
+        let meta_name = '';
+        for (const elem of elems) {
+          meta_name = elem.getAttribute('name');
+          if (!meta_name) {continue;}
+          meta_name = meta_name.toLowerCase();
+          if (meta_name.indexOf('og:image') > -1) {
+            data.cover = elem.getAttribute('content');
+          }else if (meta_name.indexOf('og:title') > -1) {
+            data.title = elem.getAttribute('content');
+          }
+        }
+
+        await hero.close();
+    }catch(error) {
+        console.error("Error got when request %s via hero: %s", url, error);
+    };
+
+    return data;
+  }
+
+}
+
+export default Xigua;
diff --git a/test/scrap_test.mjs b/test/scrap_test.mjs
index 478cc3e..b617751 100644
--- a/test/scrap_test.mjs
+++ b/test/scrap_test.mjs
@@ -1,10 +1,9 @@
 import Douyin from '../bot/Douyin.mjs';
 import Kuaishou from '../bot/Kuaishou.mjs';
+import Xigua from '../bot/Xigua.mjs';
 import configs from '../config.mjs';
 
 (async () => {
-    configs.heroTabOptions.timeoutMs = 10000;   //所有内容加载完成超时
-    console.log('Hero配置', configs);
     const heroCloudServer = 'ws://192.168.3.13:1818';
 
     //抖音测试
@@ -15,10 +14,23 @@ import configs from '../config.mjs';
     //console.log("解析结果:\n%s", JSON.stringify(data));
 
     //快手测试
-    const kuaishou = new Kuaishou(heroCloudServer);
-    const url = 'https://www.kuaishou.com/f/X8FTguiIjZQVwE7';
+    //configs.heroTabOptions.timeoutMs = 10000;   //所有内容加载完成超时
+    //console.log('Hero配置', configs);
+    //const kuaishou = new Kuaishou(heroCloudServer);
+    //const url = 'https://www.kuaishou.com/f/X8FTguiIjZQVwE7';
+    //console.log('请求中: %s ...', url);
+    //const data = await kuaishou.scrap(url);
+    //console.log("解析结果:\n%s", JSON.stringify(data));
+
+    //西瓜测试
+    const url = 'https://www.ixigua.com/7092326495246516749';
+    configs.heroBotOptions.referrer = url;
+    configs.heroBotOptions.userAgent = 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/116.0.0.0 Safari/537.36';
+    console.log('Hero配置', configs);
+
+    const xigua = new Xigua(heroCloudServer);
     console.log('请求中: %s ...', url);
-    const data = await kuaishou.scrap(url);
+    const data = await xigua.scrap(`${url}?wid_try=1`);
     console.log("解析结果:\n%s", JSON.stringify(data));
 
 })();