chore: add toMarkdown test case

ycjcl868 · ycjcl868 · commit 2e04b7cdd883 · 2025-06-02T17:20:23.000+08:00
diff --git a/packages/agent-infra/shared/src/browser/to-markdown.ts b/packages/agent-infra/shared/src/browser/to-markdown.ts
@@ -3,11 +3,12 @@
  * SPDX-License-Identifier: Apache-2.0
  */
 
-import Turndown from 'turndown';
+import Turndown, { TagName } from 'turndown';
 import { gfm } from 'turndown-plugin-gfm';
 
 export interface ToMarkdownOptions extends Turndown.Options {
   gfmExtension?: boolean;
+  removeTags?: TagName[];
 }
 
 /**
@@ -29,6 +30,7 @@ export function toMarkdown(
       emDelimiter = '*',
       strongDelimiter = '**',
       gfmExtension = true,
+      removeTags = ['script', 'style', 'link'],
     } = options;
 
     const turndown = new Turndown({
@@ -38,6 +40,9 @@ export function toMarkdown(
       strongDelimiter,
     });
 
+    // issue: https://github.com/mixmark-io/turndown/issues/210#issuecomment-353666857
+    turndown.remove(removeTags);
+
     if (gfmExtension) {
       turndown.use(gfm);
     }
diff --git a/packages/agent-infra/shared/tests/browser/to-markdown.test.ts b/packages/agent-infra/shared/tests/browser/to-markdown.test.ts
@@ -0,0 +1,44 @@
+import { describe, it, expect } from 'vitest';
+import { toMarkdown } from '../../src/browser/to-markdown';
+
+describe('toMarkdown', () => {
+  it('should convert HTML to Markdown', () => {
+    const html = '<h1>Hello, world!</h1>';
+    const markdown = toMarkdown(html);
+    expect(markdown).toBe('# Hello, world!');
+  });
+
+  it('should remove tags', () => {
+    const html = `<html>
+  <head>
+    <script formula-runtime >function e(e){for(var r=1;r<arguments.length;r++){}}</script>
+    <body>
+      <div id="app"><!--[--><!--[--><!--[--><!--[--><!--[--><!----><!---->
+        <div id="global" data-logged="0" class="layout limit" style="--40bdee49:1728px;" data-v-34b87540>
+          <div class="header-container" data-v-34b87540 style="--67f219a2:1728px;" data-v-5c1b2170>
+            <header class="mask-paper" data-v-5c1b2170>
+              <a aria-current="page" href="/explore" class="active router-link-exact-active" id="link-guide" style="display:flex;" data-v-5c1b2170>
+                <img crossorigin="anonymous" class="header-logo" style="pointer-events:none;" src="data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAM0AAABgCAYAAAC" data-v-5c1b2170>
+              </a><!--[-->
+              <div class="input-box" data-v-721de8bd>
+                <p>Hello World</p>
+                <input id="search-input" value="" type="text" spellcheck="false" class="search-input" placeholder="登录" autocomplete="off" data-v-721de8bd><!---->
+                <div class="input-button" data-v-721de8bd><!---->
+                  <div class="search-icon" data-v-721de8bd>
+                    <svg class="reds-icon" width="20" height="20" data-v-721de8bd data-v-55b36ac6><use xlink:href="#search" data-v-55b36ac6></use></svg>
+                  </div>
+                </div>
+              </div>
+            </header>
+            </div>
+          </div>
+        </div>
+    </body>
+</html>`;
+    const markdown = toMarkdown(html);
+    expect(markdown)
+      .toEqual(`[![](data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAM0AAABgCAYAAAC)](/explore)
+
+Hello World`);
+  });
+});
diff --git a/packages/agent-infra/shared/vitest.config.mts b/packages/agent-infra/shared/vitest.config.mts
@@ -0,0 +1,12 @@
+/**
+ * Copyright (c) 2025 Bytedance, Inc. and its affiliates.
+ * SPDX-License-Identifier: Apache-2.0
+ */
+import { defineConfig } from 'vitest/config';
+
+export default defineConfig({
+  test: {
+    environment: 'node',
+    include: ['**/*.{test,spec}.{js,mjs,cjs,ts,mts,cts,jsx,tsx}'],
+  },
+});
