fix: medium bug (#129)

* fix: improved medium parser for images and multi-section content

* fix: duplicate video

Author: adampash

fixtures/medium.com/1485902752952.html

@@ -19,6 +19,7 @@ export const MediumExtractor = {
 
   content: {
     selectors: [
+      ['.section-content'],
       '.section-content',
       'article > div > section',
     ],
@@ -36,10 +37,20 @@ export const MediumExtractor = {
           const [_, youtubeId] = thumb.match(ytRe) // eslint-disable-line
           $node.attr('src', `https://www.youtube.com/embed/${youtubeId}`);
           const $parent = $node.parents('figure');
-          $parent.prepend($node.clone());
-          $node.remove();
+          const $caption = $parent.find('figcaption');
+          $parent.empty().append([$node, $caption]);
         }
       },
+
+      // rewrite figures to pull out image and caption, remove rest
+      figure: ($node) => {
+        // ignore if figure has an iframe
+        if ($node.find('iframe').length > 0) return;
+
+        const $img = $node.find('img').slice(-1)[0];
+        const $caption = $node.find('figcaption');
+        $node.empty().append([$img, $caption]);
+      },
     },
 
     // Is there anything that is in the result that shouldn't be?

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

@@ -77,4 +77,26 @@ describe('MediumExtractor', () => {
       assert.equal(first13, 'Video of WTF? My talk at the White House Frontiers ConferenceLast Thursday, I');
     });
   });
+
+  describe('works with another url', () => {
+    let result;
+    let url;
+    beforeAll(() => {
+      url =
+        'https://medium.com/@JakobUlbrich/flag-attributes-in-android-how-to-use-them-ac4ec8aee7d1#.h949wjmyw';
+      const html = fs.readFileSync('./fixtures/medium.com/1485902752952.html');
+      result =
+        Mercury.parse(url, html, { fallback: false });
+    });
+
+    it('returns the content', async () => {
+      const { content } = await result;
+
+      const $ = cheerio.load(content || '');
+
+      const first13 = excerptContent($.text(), 13);
+
+      assert.equal(first13, 'I’m sure you have seen something like the following line very often while');
+    });
+  });
 });