Merge pull request #182 from yakimka/replace_lazy_images

Replace lazy pictures with original ones
extractus · Aug 16, 2021 · 9fc5e7c · 9fc5e7c
2 parents 3dee9b7 + 066b5ae
commit 9fc5e7c
Show file tree

Hide file tree

Showing 3 changed files with 52 additions and 1 deletion.
diff --git a/src/utils/parseFromHtml.test.js b/src/utils/parseFromHtml.test.js
@@ -31,3 +31,11 @@ test(`test parseFromHtml a webpage with article but no source`, async () => {
   const result = await parseFromHtml(html, ['abcd']);
   expect(result).toBeInstanceOf(Object);
 });
+
+test(`test parseFromHtml replace src with data-src in img tag`, async () => {
+  const html = readFileSync('./test-data/html-article-with-data-src.html', 'utf8');
+  const result = await parseFromHtml(html, ['abcd']);
+
+  expect(result.content).toContain('<img src="https://somewhere.any/image1.jpg" />');
+  expect(result.content).toContain('<img src="https://somewhere.any/image2.jpg" />');
+});
diff --git a/src/utils/standalizeArticle.js b/src/utils/standalizeArticle.js
@@ -23,7 +23,7 @@ module.exports = (htmlArticle, url) => {
   });
 
   $('img').each((i, elem) => {
-    const src = $(elem).attr('src');
+    const src = $(elem).attr('data-src') || $(elem).attr('src');
     if (src) {
       $(elem).attr('src', absolutifyUrl(url, src));
     }

diff --git a/test-data/html-article-with-data-src.html b/test-data/html-article-with-data-src.html
@@ -0,0 +1,43 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <meta charset="utf-8">
+    <meta content="width=device-width,initial-scale=1" name="viewport">
+    <meta content="https://somewhere.any/pages/no-article" property="og:url"/>
+    <title>TechNews</title>
+</head>
+<body>
+<article>
+    <img data-src="https://somewhere.any/image1.jpg" src="https://somewhere.any/lazy-image.jpg">
+    <img src="https://somewhere.any/image2.jpg">
+
+    Few can name an easygoing deer that isn't a frank hamster. One cannot separate tangerines from
+    excellent oranges. The literature would have us believe that a dazzling seal is not but a bee?
+    Some posit the nice strawberry to be less than fair-minded. Washing and polishing the car,a
+    monkey is the monkey of a turtle;
+
+    What we don't know for sure is whether or not the spiders could be said to resemble ambitious
+    cherries. Of course, a funny zebra's rabbit comes with it the thought that the receptive orange
+    is a fox. If this was somewhat unclear, a sheep is a strawberry's chimpanzee. The zeitgeist
+    contends that few can name a kind-hearted grapes that isn't an inventive squirrel! Few can name
+    a friendly deer that isn't a decorous grapes? What we don't know for sure is whether or not
+    those flies are nothing more than camels.
+
+    It's very tricky, if not impossible, a courageous hippopotamus is a turtle of the mind. Few can
+    name an affable nectarine that isn't a courteous bird? An alligator is an imaginative fish? Few
+    can name a quiet blackberry that isn't an independent currant. They were lost without the sedate
+    fox that composed their puppy; An ant sees a kumquat as a romantic lion. A lemon is a squirrel's
+    currant.
+
+    To be more specific, those turtles are nothing more than fishes. A grape can hardly be
+    considered a shrewd goldfish without also being an owl. Some unbiased goats are thought of
+    simply as tangerines.
+
+    Shouting with happiness, a courageous elephant is a duck of the mind? Some posit the upbeat
+    hippopotamus to be less than enchanting. It's an undeniable fact, really; authors often
+    misinterpret the grape as an endurable rabbit, when in actuality it feels more like a tough
+    dolphin. We know that a cherry can hardly be considered a responsible apricot without also being
+    a nectarine.
+</article>
+</body>
+</html>