From c85f291c72b168cc4714feb911144ddaec371bcc Mon Sep 17 00:00:00 2001
From: dqruid <31595220+dqruid@users.noreply.github.com>
Date: Sat, 1 Aug 2026 01:25:43 +0800
Subject: [PATCH] Support WeChat data-formula math

---
 src/elements/math.base.ts                     |  36 +++++-
 src/elements/math.core.ts                     |   4 +-
 src/elements/math.full.ts                     |   4 +-
 tests/expected/math--wechat-data-formula.md   | 106 ++++++++++++++++++
 tests/fixtures/math--wechat-data-formula.html |  84 ++++++++++++++
 5 files changed, 231 insertions(+), 3 deletions(-)
 create mode 100644 tests/expected/math--wechat-data-formula.md
 create mode 100644 tests/fixtures/math--wechat-data-formula.html

diff --git a/src/elements/math.base.ts b/src/elements/math.base.ts
index 768886c36..1e1b2a301 100644
--- a/src/elements/math.base.ts
+++ b/src/elements/math.base.ts
@@ -7,6 +7,19 @@ export interface MathData {
 	isBlock: boolean;
 }
 
+const DATA_FORMULA_BLOCK_TAGS = new Set([
+	'address', 'article', 'aside', 'blockquote', 'div', 'figure',
+	'footer', 'header', 'main', 'p', 'section'
+]);
+
+/**
+ * WeChat/mdnice renders the same formula in nested wrappers. Only the
+ * outermost data-formula element represents a formula in document flow;
+ * processing descendants as well would emit duplicate Markdown equations.
+ */
+export const isNestedDataFormula = (el: Element): boolean =>
+	el.hasAttribute('data-formula') && Boolean(el.parentElement?.closest('[data-formula]'));
+
 // --- MathJax CHTML (mjx-*) → MathML reconstruction -------------------------
 //
 // When MathJax v3 renders with the CHTML output and the page ships no
@@ -304,6 +317,10 @@ export const getBasicLatexFromElement = (el: Element): string | null => {
 	if (dataMath) {
 		return dataMath;                                                                                                                                             
 	}
+	const dataFormula = el.getAttribute('data-formula');
+	if (dataFormula) {
+		return dataFormula;
+	}
 	const parentDataEntry = el.parentElement?.classList.contains('hurmet-tex')
 		? el.parentElement.getAttribute('data-entry')
 		: null;
@@ -375,6 +392,22 @@ export const isBlockDisplay = (el: Element): boolean => {
 		return true;
 	}
 
+	// WeChat/mdnice uses semantic block wrappers (commonly section/div) for
+	// display equations and span wrappers for inline equations. Explicit
+	// display hints take precedence over the wrapper tag.
+	if (el.hasAttribute('data-formula')) {
+		const dataDisplayAttr = el.getAttribute('data-display');
+		if (displayAttr === 'inline' || displayAttr === 'false' || dataDisplayAttr === 'inline' || dataDisplayAttr === 'false') {
+			return false;
+		}
+		if (displayAttr === 'true' || dataDisplayAttr === 'block' || dataDisplayAttr === 'true') {
+			return true;
+		}
+		if (DATA_FORMULA_BLOCK_TAGS.has(el.tagName.toLowerCase())) {
+			return true;
+		}
+	}
+
 	// Check common class names
 	const classNames = getClassName(el).toLowerCase();
 	if (classNames.includes('display') || classNames.includes('block')) {
@@ -426,7 +459,7 @@ export const isBlockDisplay = (el: Element): boolean => {
 
 // Cheap presence check before running the full mathSelectors scan.
 // Must remain a subset of mathSelectors — every selector here should also appear there.
-export const mathFastCheck = 'math, mjx-container, .MathJax, .katex, img.latex, [data-math], [data-latex], script[type^="math/"]';
+export const mathFastCheck = 'math, mjx-container, .MathJax, .katex, img.latex, [data-math], [data-latex], [data-formula], script[type^="math/"]';
 
 // Shared selector for math elements
 export const mathSelectors = [
@@ -462,6 +495,7 @@ export const mathSelectors = [
 	'math',
 	'[data-math]',
 	'[data-latex]',
+	'[data-formula]',
 	'[data-tex]',
 	'script[type^="math/"]',
 	'annotation[encoding="application/x-tex"]'
diff --git a/src/elements/math.core.ts b/src/elements/math.core.ts
index 9299d64af..7790a2434 100644
--- a/src/elements/math.core.ts
+++ b/src/elements/math.core.ts
@@ -3,6 +3,7 @@ import {
 	getMathMLFromElement,
 	getBasicLatexFromElement as getLatexFromElement,
 	isBlockDisplay,
+	isNestedDataFormula,
 	mathSelectors,
 	mathFastCheck
 } from './math.base';
@@ -43,6 +44,7 @@ export const mathRules = [
 		fastCheck: mathFastCheck,
 		transform: (el: Element, doc: Document): Element => {
 			if (!hasHTMLElementProps(el)) return el;
+			if (isNestedDataFormula(el)) return el;
 
 			const mathData = getMathMLFromElement(el);
 			const latex = getLatexFromElement(el);
@@ -62,4 +64,4 @@ export const mathRules = [
 			return cleanMathEl;
 		}
 	}
-]; 
\ No newline at end of file
+];
diff --git a/src/elements/math.full.ts b/src/elements/math.full.ts
index 8b358ff3f..ad35a40fd 100644
--- a/src/elements/math.full.ts
+++ b/src/elements/math.full.ts
@@ -4,6 +4,7 @@ import {
 	getMathMLFromElement,
 	getBasicLatexFromElement,
 	isBlockDisplay,
+	isNestedDataFormula,
 	mathSelectors,
 	mathFastCheck
 } from './math.base';
@@ -79,6 +80,7 @@ export const mathRules = [
 			if (!('classList' in el) || !('getAttribute' in el) || !('querySelector' in el)) {
 				return el;
 			}
+			if (isNestedDataFormula(el)) return el;
 
 			const mathData = getMathMLFromElement(el);
 			const latex = getLatexFromElement(el);
@@ -98,4 +100,4 @@ export const mathRules = [
 			return cleanMathEl;
 		}
 	}
-]; 
\ No newline at end of file
+];
diff --git a/tests/expected/math--wechat-data-formula.md b/tests/expected/math--wechat-data-formula.md
new file mode 100644
index 000000000..37e8afe04
--- /dev/null
+++ b/tests/expected/math--wechat-data-formula.md
@@ -0,0 +1,106 @@
+```json
+{
+  "title": "WeChat data-formula",
+  "author": "",
+  "site": "",
+  "published": ""
+}
+```
+
+Inline formula $x_1$ should appear once.
+
+Equation 1.
+
+$$
+E_1 = x_1 \tag{1}
+$$
+
+Equation 2.
+
+$$
+E_2 = x_2 \tag{2}
+$$
+
+Equation 3.
+
+$$
+E_3 = x_3 \tag{3}
+$$
+
+Equation 4.
+
+$$
+E_4 = x_4 \tag{4}
+$$
+
+Equation 5.
+
+$$
+E_5 = x_5 \tag{5}
+$$
+
+Equation 6.
+
+$$
+E_6 = x_6 \tag{6}
+$$
+
+Equation 7.
+
+$$
+E_7 = x_7 \tag{7}
+$$
+
+Equation 8.
+
+$$
+E_8 = x_8 \tag{8}
+$$
+
+Equation 9.
+
+$$
+E_9 = x_9 \tag{9}
+$$
+
+Equation 10.
+
+$$
+E_{10} = x_{10} \tag{10}
+$$
+
+Equation 11.
+
+$$
+E_{11} = x_{11} \tag{11}
+$$
+
+Equation 12.
+
+$$
+E_{12} = x_{12} \tag{12}
+$$
+
+Equation 13.
+
+$$
+E_{13} = x_{13} \tag{13}
+$$
+
+Equation 14.
+
+$$
+E_{14} = x_{14} \tag{14}
+$$
+
+Equation 15.
+
+$$
+E_{15} = x_{15} \tag{15}
+$$
+
+Equation 16.
+
+$$
+E_{16} = x_{16} \tag{16}
+$$
diff --git a/tests/fixtures/math--wechat-data-formula.html b/tests/fixtures/math--wechat-data-formula.html
new file mode 100644
index 000000000..8e4ee33e8
--- /dev/null
+++ b/tests/fixtures/math--wechat-data-formula.html
@@ -0,0 +1,84 @@
+<!-- {"url": "https://mp.weixin.qq.com/s/example"} -->
+<!doctype html>
+<html lang="en">
+<head>
+	<meta charset="utf-8">
+	<title>WeChat data-formula</title>
+</head>
+<body>
+	<article>
+		<h1>WeChat data-formula</h1>
+		<p>
+			Inline formula
+			<span data-formula="x_1">
+				<span data-formula="x_1"><svg aria-hidden="true"></svg></span>
+			</span>
+			should appear once.
+		</p>
+		<div>
+			<p>Equation 1.</p>
+			<section data-formula="E_1 = x_1 \tag{1}"><span data-formula="E_1 = x_1 \tag{1}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 2.</p>
+			<section data-formula="E_2 = x_2 \tag{2}"><span data-formula="E_2 = x_2 \tag{2}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 3.</p>
+			<section data-formula="E_3 = x_3 \tag{3}"><span data-formula="E_3 = x_3 \tag{3}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 4.</p>
+			<section data-formula="E_4 = x_4 \tag{4}"><span data-formula="E_4 = x_4 \tag{4}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 5.</p>
+			<section data-formula="E_5 = x_5 \tag{5}"><span data-formula="E_5 = x_5 \tag{5}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 6.</p>
+			<section data-formula="E_6 = x_6 \tag{6}"><span data-formula="E_6 = x_6 \tag{6}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 7.</p>
+			<section data-formula="E_7 = x_7 \tag{7}"><span data-formula="E_7 = x_7 \tag{7}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 8.</p>
+			<section data-formula="E_8 = x_8 \tag{8}"><span data-formula="E_8 = x_8 \tag{8}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 9.</p>
+			<section data-formula="E_9 = x_9 \tag{9}"><span data-formula="E_9 = x_9 \tag{9}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 10.</p>
+			<section data-formula="E_{10} = x_{10} \tag{10}"><span data-formula="E_{10} = x_{10} \tag{10}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 11.</p>
+			<section data-formula="E_{11} = x_{11} \tag{11}"><span data-formula="E_{11} = x_{11} \tag{11}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 12.</p>
+			<section data-formula="E_{12} = x_{12} \tag{12}"><span data-formula="E_{12} = x_{12} \tag{12}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 13.</p>
+			<section data-formula="E_{13} = x_{13} \tag{13}"><span data-formula="E_{13} = x_{13} \tag{13}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 14.</p>
+			<section data-formula="E_{14} = x_{14} \tag{14}"><span data-formula="E_{14} = x_{14} \tag{14}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 15.</p>
+			<section data-formula="E_{15} = x_{15} \tag{15}"><span data-formula="E_{15} = x_{15} \tag{15}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+		<div>
+			<p>Equation 16.</p>
+			<section data-formula="E_{16} = x_{16} \tag{16}"><span data-formula="E_{16} = x_{16} \tag{16}"><svg aria-hidden="true"></svg></span></section>
+		</div>
+	</article>
+</body>
+</html>