fix: androidauthority.com - Clear polls

Remove polls that require JavaScript
jocmp · Jan 15, 2025 · af3a562 · af3a562
1 parent 97a6b1e
commit af3a562
Show file tree

Hide file tree

Showing 3 changed files with 5 additions and 3 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -3,7 +3,7 @@
 ### 2.3.0
 
 - [a45b329e0a] - fix: Update versants.com to parse figures (Josiah Campbell) [#42](https://github.com/jocmp/mercury-parser/pull/42)
-- [68e9b88a8d] - fix: androidauthority.com - Retain h3 tags (Josiah Campbell) [#41](https://github.com/jocmp/mercury-parser/pull/41)
+- [68e9b88a8d] - fix: androidauthority.com - Retain heading tags (Josiah Campbell) [#41](https://github.com/jocmp/mercury-parser/pull/41)
 - [c2bc68449f] - bump version v2.2.9 -> v2.2.10 (jocmp)
 
 ### v2.2.10 (Jan 11, 2025)

diff --git a/src/extractors/custom/www.androidauthority.com/index.js b/src/extractors/custom/www.androidauthority.com/index.js
@@ -20,7 +20,7 @@ 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', '.e_Ac'],
+    selectors: ['.e_Bc', '.d_Dd'],
     transforms: {
       ol: node => {
         node.attr('class', 'mercury-parser-keep');
@@ -29,7 +29,8 @@ export const WwwAndroidauthorityComExtractor = {
       h3: $node => $node.attr('class', 'mercury-parser-keep'),
     },
     clean: [
-      '.d_f .d_nr', // Lead image
+      '.e_Oh', // Polls
+      'picture + div', // Lead image text
     ],
   },
 };
diff --git a/src/utils/dom/constants.js b/src/utils/dom/constants.js
@@ -36,6 +36,7 @@ export const REMOVE_ATTR_LIST = REMOVE_ATTRS.join(',');
 export const WHITELIST_ATTRS = [
   'src',
   'srcset',
+  'start',
   'sizes',
   'type',
   'href',