fix: Keep nested h3's in tldr.tech

jocmp · Feb 2, 2025 · 617b694 · 617b694
1 parent d50a54d
commit 617b694
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 6 deletions.
diff --git a/src/extractors/custom/tldr.tech/index.js b/src/extractors/custom/tldr.tech/index.js
@@ -2,7 +2,7 @@ export const TldrTechExtractor = {
   domain: 'tldr.tech',
 
   title: {
-    selectors: [['meta[name="og:title"]', 'value'], 'title'],
+    selectors: ['h1'],
   },
 
   lead_image_url: {
@@ -14,7 +14,7 @@ export const TldrTechExtractor = {
 
     transforms: {
       h2: $node => $node.attr('class', 'mercury-parser-keep'),
-      h3: $node => $node.attr('class', 'mercury-parser-keep'),
+      h3: 'b',
     },
 
     clean: [],

diff --git a/src/extractors/custom/tldr.tech/index.test.js b/src/extractors/custom/tldr.tech/index.test.js
@@ -27,10 +27,7 @@ describe('TldrTechExtractor', () => {
     it('returns the title', async () => {
       const { title } = await result;
 
-      assert.equal(
-        title,
-        `OpenAI eyes $340B valuation 💰, Gemini 2.0 🤖, Javascript Temporal 👨‍💻`
-      );
+      assert.equal(title, `TLDR 2025-01-31`);
     });
     it('returns the lead_image_url', async () => {
       const { lead_image_url } = await result;