Test media tag (#4796)

* feat: add html video tag convertion (#4784)

Co-authored-by: Zhenyi Wang <zhenyiwang@intl.zju.edu.cn>

* perf: media tag

---------

Co-authored-by: Zhenyi-Wang <47094597+Zhenyi-Wang@users.noreply.github.com>
Co-authored-by: Zhenyi Wang <zhenyiwang@intl.zju.edu.cn>

docSite/content/zh-cn/docs/development/upgrading/498.md

@@ -23,6 +23,7 @@ weight: 792
 1. Chat log list 优化，避免大数据时超出内存限制。
 2. 预加载 token 计算 worker，避免主任务中并发创建导致线程阻塞。
 3. 工作流节点版本控制交互优化。
+4. 网络获取以及 html2md 优化，支持视频和音频标签的转换。
 
 ## 🐛 修复
 

packages/service/common/string/cheerio.ts

@@ -42,7 +42,7 @@ export const cheerioToHtml = ({
       }
     }
   });
-  selectDom.find('img').each((i, el) => {
+  selectDom.find('img, video, source, audio, iframe').each((i, el) => {
     const src = $(el).attr('src');
     if (src) {
       if (src.startsWith('//')) {

packages/service/worker/htmlStr2Md/utils.ts

@@ -43,6 +43,24 @@ export const html2md = (
     turndownService.remove(['i', 'script', 'iframe', 'style']);
     turndownService.use(turndownPluginGfm.gfm);
 
+    // add custom handling for media tag
+    turndownService.addRule('media', {
+      filter: ['video', 'source', 'audio'],
+      replacement: function (content, node) {
+        const mediaNode = node as HTMLVideoElement | HTMLAudioElement | HTMLSourceElement;
+        const src = mediaNode.getAttribute('src');
+        const sources = mediaNode.getElementsByTagName('source');
+        const firstSourceSrc = sources.length > 0 ? sources[0].getAttribute('src') : null;
+        const mediaSrc = src || firstSourceSrc;
+
+        if (mediaSrc) {
+          return `[${mediaSrc}](${mediaSrc}) `;
+        }
+
+        return content;
+      }
+    });
+
     // Base64 img to id, otherwise it will occupy memory when going to md
     const { processedHtml, images } = processBase64Images(html);
     const md = turndownService.turndown(processedHtml);