fix: www.androidauthority.com - Clear polls

Remove polls that require JavaScript
jocmp · Jan 15, 2025 · 22a29c3 · 22a29c3
1 parent 00f01a6
commit 22a29c3
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 2 deletions.
diff --git a/src/extractors/custom/www.androidauthority.com/index.js b/src/extractors/custom/www.androidauthority.com/index.js
@@ -20,16 +20,23 @@ 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: ['main'],
     transforms: {
       ol: node => {
         node.attr('class', 'mercury-parser-keep');
       },
       h2: $node => $node.attr('class', 'mercury-parser-keep'),
       h3: $node => $node.attr('class', 'mercury-parser-keep'),
+      p: node => {
+        if (node.text().startsWith('Affiliate links')) {
+          node.remove();
+        }
+      },
     },
     clean: [
-      '.d_f .d_nr', // Lead image
+      'h1', // Clean title
+      'h1 ~ *', // Clean subtitle
+      '.e_Oh', // Polls
     ],
   },
 };
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',