Adding scrape files for testing
This commit is contained in:
Max committed
1 parent
7d255fe687
commit
8d60aae65c
40 files changed
+4423
No files matched your search
Vendored
+7
@@ -0,0 +1,7 @@
|
||||
# HTML fixtures
|
||||
|
||||
Frozen, hand-trimmed snapshots used by `server/tests/helpers/SiteScrape.test.js`.
|
||||
|
||||
These are **not** fetched live. Each file isolates one parsing situation so a
|
||||
failure points at a specific branch. Add a new file rather than growing an
|
||||
existing one when covering a new case.
|
||||
@@ -0,0 +1,9 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head>
|
||||
<title>Icons By Size</title>
|
||||
<link rel="icon" href="https://example.com/small.png" sizes="16x16">
|
||||
<link rel="apple-touch-icon" href="https://example.com/touch-180.png" sizes="180x180">
|
||||
</head>
|
||||
<body><p>x</p></body>
|
||||
</html>
|
||||
@@ -0,0 +1,8 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head>
|
||||
<title>Generated OG Image</title>
|
||||
<meta property="og:image" content="https://example.com/api/og?title=Hello&theme=dark">
|
||||
</head>
|
||||
<body><p>x</p></body>
|
||||
</html>
|
||||
+5
@@ -0,0 +1,5 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head></head>
|
||||
<body><h1>Heading Is All We Have</h1><p>No title element, no meta tags.</p></body>
|
||||
</html>
|
||||
@@ -0,0 +1,10 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head><title>Broken Images</title></head>
|
||||
<body>
|
||||
<img>
|
||||
<img alt="no src here">
|
||||
<img src="">
|
||||
<img data-src="https://cdn.example.com/lazy.jpg">
|
||||
</body>
|
||||
</html>
|
||||
+15
@@ -0,0 +1,15 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head>
|
||||
<script type="application/ld+json">
|
||||
{
|
||||
"@context": "https://schema.org",
|
||||
"@type": "NewsArticle",
|
||||
"headline": "JSON-LD Headline",
|
||||
"description": "Described in structured data.",
|
||||
"image": ["https://cdn.example.com/ld-primary.jpg", "https://cdn.example.com/ld-second.jpg"]
|
||||
}
|
||||
</script>
|
||||
</head>
|
||||
<body><p>x</p></body>
|
||||
</html>
|
||||
+15
@@ -0,0 +1,15 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head>
|
||||
<script type="application/ld+json">
|
||||
{
|
||||
"@context": "https://schema.org",
|
||||
"@graph": [
|
||||
{ "@type": "Organization", "name": "Not The Article" },
|
||||
{ "@type": "BlogPosting", "headline": "Graph Headline", "image": { "url": "https://cdn.example.com/graph.png" } }
|
||||
]
|
||||
}
|
||||
</script>
|
||||
</head>
|
||||
<body><p>x</p></body>
|
||||
</html>
|
||||
@@ -0,0 +1,11 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head>
|
||||
<title>Survived Bad JSON</title>
|
||||
<script type="application/ld+json">
|
||||
{ "@type": "Article", "headline": "unterminated,,, }
|
||||
</script>
|
||||
<script type="application/ld+json"></script>
|
||||
</head>
|
||||
<body><p>x</p></body>
|
||||
</html>
|
||||
+16
@@ -0,0 +1,16 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head><title>Beekeeping Guide</title></head>
|
||||
<body>
|
||||
<nav>Home About Contact Login Signup Navigation Navigation</nav>
|
||||
<script>var tracking = "tracking tracking tracking tracking tracking";</script>
|
||||
<style>.beehive { color: red }</style>
|
||||
<article class="content">
|
||||
<p>Beekeeping is rewarding. Beekeeping requires patience. A beehive needs space.</p>
|
||||
<p>The beehive must stay dry. Honey harvest comes later. Honey harvest rewards patience.</p>
|
||||
<p>Beekeeping beginners should buy a beehive and study honey production carefully.</p>
|
||||
<p>Honey production depends on the beehive and on steady beekeeping habits.</p>
|
||||
</article>
|
||||
<footer>Footer footer footer links copyright</footer>
|
||||
</body>
|
||||
</html>
|
||||
@@ -0,0 +1,7 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head><title>Unclosed Tags
|
||||
<body>
|
||||
<div><p>Paragraph never closed
|
||||
<ul><li>item one<li>item two
|
||||
</html>
|
||||
+2
@@ -0,0 +1,2 @@
|
||||
<!doctype html>
|
||||
<html><head></head><body></body></html>
|
||||
+14
@@ -0,0 +1,14 @@
|
||||
<!doctype html>
|
||||
<html lang="en">
|
||||
<head>
|
||||
<title>Fallback Title Nobody Should See</title>
|
||||
<meta property="og:title" content="The Real Title">
|
||||
<meta property="og:description" content="A proper social description of the page.">
|
||||
<meta property="og:site_name" content="Example Journal">
|
||||
<meta property="og:image" content="https://cdn.example.com/hero.jpg">
|
||||
<meta property="og:url" content="https://example.com/canonical-post">
|
||||
<meta name="description" content="A lesser meta description.">
|
||||
<link rel="canonical" href="https://example.com/canonical-post">
|
||||
</head>
|
||||
<body><article><p>Body text.</p></article></body>
|
||||
</html>
|
||||
@@ -0,0 +1,8 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head>
|
||||
<title>Protocol Relative</title>
|
||||
<meta property="og:image" content="//cdn.example.com/proto.jpg">
|
||||
</head>
|
||||
<body><p>x</p></body>
|
||||
</html>
|
||||
@@ -0,0 +1,9 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head>
|
||||
<title>Relative Assets</title>
|
||||
<link rel="canonical" href="https://example.com/blog/2026/post">
|
||||
<meta property="og:image" content="../images/hero.png">
|
||||
</head>
|
||||
<body><p>x</p></body>
|
||||
</html>
|
||||
@@ -0,0 +1,8 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head>
|
||||
<title>Icon Only</title>
|
||||
<link rel="shortcut icon" href="/favicon.ico">
|
||||
</head>
|
||||
<body><p>x</p></body>
|
||||
</html>
|
||||
@@ -0,0 +1,9 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head><title>Sized</title></head>
|
||||
<body>
|
||||
<img src="https://cdn.example.com/thumb.jpg" width="80" height="60">
|
||||
<img src="https://cdn.example.com/big.jpg" width="1200" height="800">
|
||||
<img src="https://cdn.example.com/mid.jpg" width="400" height="300">
|
||||
</body>
|
||||
</html>
|
||||
+7
@@ -0,0 +1,7 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head><title>Srcset</title></head>
|
||||
<body>
|
||||
<img srcset="https://cdn.example.com/small.jpg 320w, https://cdn.example.com/large.jpg 1600w, https://cdn.example.com/medium.jpg 800w">
|
||||
</body>
|
||||
</html>
|
||||
@@ -0,0 +1,8 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head>
|
||||
<title>SVG Only</title>
|
||||
<meta property="og:image" content="https://cdn.example.com/logo.svg">
|
||||
</head>
|
||||
<body><p>x</p></body>
|
||||
</html>
|
||||
+5
@@ -0,0 +1,5 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head><title> Just A Title </title></head>
|
||||
<body><h1>An H1 Heading</h1><p>No metadata here at all.</p></body>
|
||||
</html>
|
||||
@@ -0,0 +1,9 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head><title>Pixels</title></head>
|
||||
<body>
|
||||
<img src="https://analytics.example.net/1x1.gif" width="1" height="1">
|
||||
<img src="https://cdn.example.com/spacer.gif">
|
||||
<img src="https://cdn.example.com/real-photo.jpg" width="800" height="600">
|
||||
</body>
|
||||
</html>
|
||||
+10
@@ -0,0 +1,10 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head>
|
||||
<title>Document Title</title>
|
||||
<meta name="twitter:title" content="Twitter Card Title">
|
||||
<meta name="twitter:description" content="Twitter card description.">
|
||||
<meta name="twitter:image" content="https://cdn.example.com/twitter.png">
|
||||
</head>
|
||||
<body><p>x</p></body>
|
||||
</html>
|
||||
@@ -0,0 +1,8 @@
|
||||
<!doctype html>
|
||||
<html>
|
||||
<head>
|
||||
<title>Modern Formats</title>
|
||||
<meta property="og:image" content="https://cdn.example.com/hero.webp">
|
||||
</head>
|
||||
<body><p>x</p></body>
|
||||
</html>
|
||||
Reference in new issue
Block a user