From 0cdbfe8e9a1347d0e21515aad1f6333cba124d0c Mon Sep 17 00:00:00 2001
From: Mimi <1119186082@qq.com>
Date: Fri, 7 Aug 2026 09:47:20 +0800
Subject: [PATCH] fix(tocObj): extract semantic math text

---
 lib/toc_obj.ts       | 35 ++++++++++++++++++++++++++++++++---
 test/toc_obj.spec.ts | 33 +++++++++++++++++++++++++++++++++
 2 files changed, 65 insertions(+), 3 deletions(-)

diff --git a/lib/toc_obj.ts b/lib/toc_obj.ts
index 478669b6..05dfa089 100644
--- a/lib/toc_obj.ts
+++ b/lib/toc_obj.ts
@@ -1,5 +1,5 @@
 import { DomHandler, DomUtils, Parser } from 'htmlparser2';
-import type { Element } from 'domhandler';
+import type { AnyNode, Element } from 'domhandler';
 import escapeHTML from './escape_html';
 const nonWord = /^\s*[^a-zA-Z0-9]\s*$/;
 
@@ -20,6 +20,35 @@ const isUnnumbered = ({ attribs = {} }) => {
   return attribs['data-toc-unnumbered'] === 'true';
 };
 
+const isElement = (node: AnyNode): node is Element => 'tagName' in node;
+
+/**
+ * Extract text from the semantic representation of a DOM subtree.
+ */
+const semanticText = (node: AnyNode): string => {
+  if (node.type === 'text') return node.data;
+  if (!('children' in node)) return '';
+
+  if (isElement(node)) {
+    const { attribs, name } = node;
+    if (attribs['aria-hidden']?.toLowerCase() === 'true' || 'hidden' in attribs) return '';
+
+    if (name === 'semantics') {
+      const tex = node.children.find(child => isElement(child)
+        && child.name === 'annotation'
+        && child.attribs.encoding?.toLowerCase() === 'application/x-tex');
+      if (tex) return semanticText(tex);
+
+      const presentation = node.children.find(child => isElement(child)
+        && child.name !== 'annotation'
+        && child.name !== 'annotation-xml');
+      return presentation ? semanticText(presentation) : '';
+    }
+  }
+
+  return node.children.map(semanticText).join('');
+};
+
 interface Result {
   text: string;
   id: string;
@@ -48,7 +77,7 @@ function tocObj(str: string, options = {}) {
     const unnumbered = isUnnumbered(el);
     let text = '';
     for (const element of el.children) {
-      const elText = DomUtils.textContent(element);
+      const elText = semanticText(element);
       // Skip permalink symbol wrapped in <a>
       // permalink is a single non-word character, word = [a-Z0-9]
       // permalink may be wrapped in whitespace(s)
@@ -56,7 +85,7 @@ function tocObj(str: string, options = {}) {
         text += escapeHTML(elText);
       }
     }
-    if (!text) text = escapeHTML(DomUtils.textContent(el));
+    if (!text) text = escapeHTML(semanticText(el));
 
     const res: Result = { text, id, level };
     if (unnumbered) res.unnumbered = true;
diff --git a/test/toc_obj.spec.ts b/test/toc_obj.spec.ts
index 40ad65c1..2a9b886a 100644
--- a/test/toc_obj.spec.ts
+++ b/test/toc_obj.spec.ts
@@ -164,6 +164,39 @@ describe('tocObj', () => {
       result.forEach(str => str[0].text.should.eql('foobarbaz'));
     });
 
+    it('hidden elements', () => {
+      const input = '<h1>Visible<span aria-hidden="true">Aria hidden</span><span hidden>Hidden</span></h1>';
+      const result = tocObj(input);
+
+      result[0].text.should.eql('Visible');
+    });
+
+    it('KaTeX MathML', () => {
+      const input = [
+        '<h2 id="energy">Energy <span class="katex">',
+        '<span class="katex-mathml"><math><semantics>',
+        '<mrow><mi>E</mi><mo>=</mo><mi>m</mi><msup><mi>c</mi><mn>2</mn></msup></mrow>',
+        '<annotation encoding="application/x-tex">E=mc^2</annotation>',
+        '</semantics></math></span>',
+        '<span class="katex-html" aria-hidden="true">E=mc2</span>',
+        '</span></h2>'
+      ].join('');
+      const result = tocObj(input);
+
+      result[0].text.should.eql('Energy E&#x3D;mc^2');
+    });
+
+    it('MathML without TeX annotation', () => {
+      const input = [
+        '<h2><math><semantics><mrow><mi>x</mi><mo>+</mo><mn>1</mn></mrow>',
+        '<annotation encoding="application/mathml-content+xml">duplicate</annotation>',
+        '</semantics></math></h2>'
+      ].join('');
+      const result = tocObj(input);
+
+      result[0].text.should.eql('x+1');
+    });
+
     it('unnumbered headings', () => {
       const input = [
         '<h1 id="title_1" data-toc-unnumbered="true">Title 1</h1>',