feat: Keep headings if they use the KEEP_CLASS

CHANGELOG.md

@@ -1,9 +1,6 @@
 # Mercury Parser Changelog
 
-### v2.2.11 (Jan 13, 2025)
-
-- [68e9b88a8d] - fix: androidauthority.com - Retain h3 tags (Josiah Campbell) [#41](https://github.com/jocmp/mercury-parser/pull/41)
-- [c2bc68449f] - bump version v2.2.9 -> v2.2.10 (jocmp)
+### v2.3.0
 
 ### v2.2.10 (Jan 11, 2025)
 

src/extractors/custom/www.androidauthority.com/index.js

@@ -20,13 +20,13 @@ export const WwwAndroidauthorityComExtractor = {
   // remove if not following a paragraph. Adding this empty paragraph fixes it, and
   // the empty paragraph will be removed anyway.
   content: {
-    selectors: ['.d_Dd'],
+    selectors: ['.d_Dd', '.e_Ac'],
     transforms: {
       ol: node => {
         node.attr('class', 'mercury-parser-keep');
       },
-      h2: $node => $node.before('<p></p>'),
-      h3: $node => $node.before('<p></p>'),
+      h2: $node => $node.attr('class', 'mercury-parser-keep'),
+      h3: $node => $node.attr('class', 'mercury-parser-keep'),
     },
     clean: [
       '.d_f .d_nr', // Lead image

src/utils/dom/clean-headers.js

@@ -1,11 +1,16 @@
 import { getWeight } from 'extractors/generic/content/scoring';
 
-import { HEADER_TAG_LIST } from './constants';
+import { HEADER_TAG_LIST, KEEP_CLASS } from './constants';
 import { normalizeSpaces } from '../text';
 
 export default function cleanHeaders($article, $, title = '') {
   $(HEADER_TAG_LIST, $article).each((index, header) => {
     const $header = $(header);
+
+    if ($(header).hasClass(KEEP_CLASS)) {
+      return $header;
+    }
+
     // Remove any headers that appear before all other p tags in the
     // document. This probably means that it was part of the title, a
     // subtitle or something else extraneous like a datestamp or byline,

src/utils/dom/clean-headers.test.js

@@ -69,4 +69,24 @@ describe('cleanHeaders(article, $)', () => {
     `
     );
   });
+
+  it('keeps headers with keep class', () => {
+    const $ = cheerio.load(`
+      <div>
+        <h3 class="mercury-parser-keep">Keep me</h3>
+        <p>What do you think?</p>
+      </div>
+    `);
+
+    const result = cleanHeaders($('*').first(), $);
+    assertClean(
+      result.html(),
+      `
+      <div>
+        <h3 class="mercury-parser-keep">Keep me</h3>
+        <p>What do you think?</p>
+      </div>
+    `
+    );
+  });
 });