feat: Add custom parser - channelnewsasia.com (#44)

CHANGELOG.md

@@ -1,5 +1,10 @@
 # Mercury Parser Changelog
 
+### 2.3.1
+
+- [ba949c0c87] - feat: Add custom parser - channelnewsasia.com (Josiah Campbell) [#44](https://github.com/jocmp/mercury-parser/pull/44)
+- [6abdf5a862] - chore: Fix references to Postlight Parser in README (Josiah Campbell)
+
 ### 2.3.0 (Jan 15, 2025)
 
 - [a45b329e0a] - fix: Update versants.com to parse figures (Josiah Campbell) [#42](https://github.com/jocmp/mercury-parser/pull/42)

src/extractors/custom/index.js

@@ -148,3 +148,4 @@ export * from './techcrunch.com';
 export * from './www.hardwarezone.com.sg';
 export * from './www.spiegel.de';
 export * from './mobilesyrup.com';
+export * from './www.channelnewsasia.com';

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

@@ -0,0 +1,36 @@
+export const WwwChannelnewsasiaComExtractor = {
+  domain: 'www.channelnewsasia.com',
+
+  title: {
+    selectors: [['meta[name="og:title"]', 'value']],
+  },
+
+  author: {
+    selectors: [
+      '.link--author-profile',
+      ['meta[name="cXenseParse:author"]', 'value'],
+    ],
+  },
+
+  date_published: {
+    selectors: ['.article-publish:not(span)'],
+    format: 'DD MMM YYYY HH:mma',
+    timezone: 'Asia/Singapore',
+  },
+
+  dek: {
+    selectors: ['.content-detail__description'],
+  },
+
+  lead_image_url: {
+    selectors: [['meta[name="og:image"]', 'value']],
+  },
+
+  content: {
+    selectors: ['section[data-title="Content"]'],
+
+    transforms: {},
+
+    clean: [],
+  },
+};

src/extractors/custom/www.channelnewsasia.com/index.test.js

@@ -0,0 +1,86 @@
+import assert from 'assert';
+import URL from 'url';
+import cheerio from 'cheerio';
+
+import Parser from 'mercury';
+import getExtractor from 'extractors/get-extractor';
+import { excerptContent } from 'utils/text';
+
+const fs = require('fs');
+
+describe('WwwChannelnewsasiaComExtractor', () => {
+  describe('initial test case', () => {
+    let result;
+    let url;
+    beforeAll(() => {
+      url =
+        'https://www.channelnewsasia.com/singapore/police-arrest-suspects-ica-change-addresses-unauthorised-attempts-4869916';
+      const html = fs.readFileSync(
+        './fixtures/www.channelnewsasia.com/1737324313613.html'
+      );
+      result = Parser.parse(url, { html, fallback: false });
+    });
+
+    it('is selected properly', () => {
+      const extractor = getExtractor(url);
+      assert.equal(extractor.domain, URL.parse(url).hostname);
+    });
+
+    it('returns the title', async () => {
+      const { title } = await result;
+
+      assert.equal(
+        title,
+        `Another six people arrested for unauthorised address changes using ICA online service`
+      );
+    });
+
+    it('returns the author', async () => {
+      const { author } = await result;
+
+      assert.equal(author, 'Daphne Yow');
+    });
+
+    it('returns the date_published', async () => {
+      const { date_published } = await result;
+
+      assert.equal(date_published, '2025-01-17T13:06:00.000Z');
+    });
+
+    it('returns the dek', async () => {
+      const { dek } = await result;
+
+      assert.equal(
+        dek,
+        'This brings the total number of people arrested so far to 13.'
+      );
+    });
+
+    it('returns the lead_image_url', async () => {
+      const { lead_image_url } = await result;
+
+      assert.equal(
+        lead_image_url,
+        `https://dam.mediacorp.sg/image/upload/s--WYyV_F7F--/c_crop,h_574,w_1021,x_4,y_1/f_auto,q_auto/c_fill,g_auto,h_676,w_1200/v1/mediacorp/cna/image/2022/11/17/ica3.jpg?itok=smLA3pam`
+      );
+    });
+
+    it('returns the content', async () => {
+      const { content } = await result;
+
+      const $ = cheerio.load(content || '');
+
+      const first13 = excerptContent(
+        $('*')
+          .first()
+          .text(),
+        13
+      );
+
+      assert.equal(
+        first13,
+        'SINGAPORE: The police have arrested another six people in relation to a series'
+      );
+    });
+  });
+});