feat: arstechnica.com extractor

src/extractors/collect-all-pages.js

@@ -30,7 +30,6 @@ export default async function collectAllPages({
       html,
       $,
       metaCache,
-      contentOnly: true,
       extractedTitle: title,
       previousUrls,
     };

src/extractors/custom/arstechnica.com/index.js

@@ -0,0 +1,63 @@
+export const ArstechnicaComExtractor = {
+  domain: 'arstechnica.com',
+
+  // Articles from this site are often paginated, but I was unable to write a CSS
+  // selector to find the next page. On the last page, there will be a link with a CSS
+  // selector indicating that the previous page is next. But the parser appears to find
+  // the next page without this extractor finding it, as long as the fallback option is
+  // left at its default value of true.
+
+  title: {
+    selectors: ['title'],
+  },
+
+  author: {
+    selectors: ['*[rel="author"] *[itemprop="name"]'],
+  },
+
+  date_published: {
+    selectors: [['.byline time', 'datetime']],
+  },
+
+  dek: {
+    selectors: ['h2[itemprop="description"]'],
+  },
+
+  lead_image_url: {
+    selectors: [['meta[name="og:image"]', 'value']],
+  },
+
+  content: {
+    selectors: ['div[itemprop="articleBody"]'],
+
+    // Is there anything in the content you selected that needs transformed
+    // before it's consumable content? E.g., unusual lazy loaded images
+    transforms: {
+      h2: $node => {
+        // Some pages have an element h2 that is significant, and that the parser will
+        // remove if not following a paragraph. Adding this empty paragraph fixes it, and
+        // the empty paragraph will be removed anyway.
+        $node.before('<p></p>');
+      },
+    },
+
+    // Is there anything that is in the result that shouldn't be?
+    // The clean selectors will remove anything that matches from
+    // the result.
+    clean: [
+      // Remove enlarge links and separators inside image captions.
+      'figcaption .enlarge-link',
+      'figcaption .sep',
+
+      // I could not transform the video into usable elements, so I
+      // removed them.
+      'figure.video',
+
+      // Image galleries that do not work.
+      '.gallery',
+
+      'aside',
+      '.sidebar',
+    ],
+  },
+};

src/extractors/custom/arstechnica.com/index.test.js

@@ -0,0 +1,159 @@
+import assert from 'assert';
+import URL from 'url';
+import cheerio from 'cheerio';
+
+import Mercury from 'mercury';
+import getExtractor from 'extractors/get-extractor';
+import { excerptContent } from 'utils/text';
+
+const fs = require('fs');
+
+describe('ArstechnicaComExtractor', () => {
+  describe('initial test case', () => {
+    let result;
+    let url;
+    beforeAll(() => {
+      url =
+        'https://arstechnica.com/gadgets/2016/08/the-connected-renter-how-to-make-your-apartment-smarter/';
+      const html = fs.readFileSync(
+        './fixtures/arstechnica.com/1587927767738.html'
+      );
+      result = Mercury.parse(url, { html, fallback: false });
+    });
+
+    it('is selected properly', () => {
+      // This test should be passing by default.
+      // It sanity checks that the correct parser
+      // is being selected for URLs from this domain
+      const extractor = getExtractor(url);
+      assert.equal(extractor.domain, URL.parse(url).hostname);
+    });
+
+    it('returns the title', async () => {
+      // To pass this test, fill out the title selector
+      // in ./src/extractors/custom/arstechnica.com/index.js.
+      const { title } = await result;
+
+      // Update these values with the expected values from
+      // the article.
+      assert.equal(
+        title,
+        `The connected renter: How to make your apartment smarter`
+      );
+    });
+
+    it('returns the author', async () => {
+      // To pass this test, fill out the author selector
+      // in ./src/extractors/custom/arstechnica.com/index.js.
+      const { author } = await result;
+
+      // Update these values with the expected values from
+      // the article.
+      assert.equal(author, `Valentina Palladino`);
+    });
+
+    it('returns the date_published', async () => {
+      // To pass this test, fill out the date_published selector
+      // in ./src/extractors/custom/arstechnica.com/index.js.
+      const { date_published } = await result;
+
+      // Update these values with the expected values from
+      // the article.
+      assert.equal(date_published, `2016-08-10T11:15:53.000Z`);
+    });
+
+    it('returns the dek', async () => {
+      // To pass this test, fill out the dek selector
+      // in ./src/extractors/custom/arstechnica.com/index.js.
+      const { dek } = await result;
+
+      // Update these values with the expected values from
+      // the article.
+      assert.equal(
+        dek,
+        'Turning your rented space into a smart home can be tricky; we have some advice.'
+      );
+    });
+
+    it('returns the lead_image_url', async () => {
+      // To pass this test, fill out the lead_image_url selector
+      // in ./src/extractors/custom/arstechnica.com/index.js.
+      const { lead_image_url } = await result;
+
+      // Update these values with the expected values from
+      // the article.
+      assert.equal(
+        lead_image_url,
+        `https://cdn.arstechnica.net/wp-content/uploads/2016/07/smartapartment_hero5-640x215.jpg`
+      );
+    });
+
+    // it('returns the pages_rendered', async () => {
+    //     // To pass this test, fill out the pages_rendered selector
+    //     // in ./src/extractors/custom/arstechnica.com/index.js.
+    //     const { pages_rendered } = await result
+    //
+    //     // Update these values with the expected values from
+    //     // the article.
+    //     assert.equal(pages_rendered, `3`)
+    //   });
+
+    it('returns the content', async () => {
+      // To pass this test, fill out the content selector
+      // in ./src/extractors/custom/arstechnica.com/index.js.
+      // You may also want to make use of the clean and transform
+      // options.
+      const { content } = await result;
+
+      const $ = cheerio.load(content || '');
+
+      const first13 = excerptContent(
+        $('*')
+          .first()
+          .text(),
+        13
+      );
+
+      // Update these values with the expected values from
+      // the article.
+      assert.equal(
+        first13,
+        "Name a home appliance or product, and there's probably a smart version of"
+      );
+    });
+  });
+
+  describe('Keep the first h2 on subsequent pages test', () => {
+    let result;
+    let url;
+    beforeAll(() => {
+      url =
+        'https://arstechnica.com/science/2020/04/should-you-wear-a-face-mask-heres-all-the-data-we-have/2/';
+      const html = fs.readFileSync(
+        './fixtures/arstechnica.com/1587927767738.html'
+      );
+      result = Mercury.parse(url, { html, fallback: false });
+    });
+
+    it('is selected properly', () => {
+      // This test should be passing by default.
+      // It sanity checks that the correct parser
+      // is being selected for URLs from this domain
+      const extractor = getExtractor(url);
+      assert.equal(extractor.domain, URL.parse(url).hostname);
+    });
+
+    it('returns the content', async () => {
+      // To pass this test, fill out the content selector
+      // in ./src/extractors/custom/arstechnica.com/index.js.
+      // You may also want to make use of the clean and transform
+      // options.
+      const { content } = await result;
+
+      const $ = cheerio.load(content || '');
+
+      const h2 = $('h2');
+      assert.equal(h2.length, 1);
+    });
+  });
+});

src/extractors/custom/index.js

@@ -139,3 +139,4 @@ export * from './pastebin.com';
 export * from './www.abendblatt.de';
 export * from './www.gruene.de';
 export * from './www.engadget.com';
+export * from './arstechnica.com';