LinkDetailsExtractor adjustments (#31357)

This commit is contained in:
Christian Schmidt 2024-11-22 09:00:14 +01:00 committed by GitHub
parent 68c7782940
commit 0518613dd7
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
3 changed files with 21 additions and 9 deletions

View file

@ -157,7 +157,7 @@ class LinkDetailsExtractor
end
def title
html_entities.decode(structured_data&.headline || opengraph_tag('og:title') || document.xpath('//title').map(&:content).first)&.strip
html_entities.decode(structured_data&.headline || opengraph_tag('og:title') || head.at_xpath('title')&.content)&.strip
end
def description
@ -205,11 +205,11 @@ class LinkDetailsExtractor
end
def language
valid_locale_or_nil(structured_data&.language || opengraph_tag('og:locale') || document.xpath('//html').pick('lang'))
valid_locale_or_nil(structured_data&.language || opengraph_tag('og:locale') || document.root.attr('lang'))
end
def icon
valid_url_or_nil(structured_data&.publisher_icon || link_tag('apple-touch-icon') || link_tag('shortcut icon'))
valid_url_or_nil(structured_data&.publisher_icon || link_tag('apple-touch-icon') || link_tag('icon'))
end
private
@ -237,18 +237,20 @@ class LinkDetailsExtractor
end
def link_tag(name)
document.xpath("//link[nokogiri:link_rel_include(@rel, '#{name}')]", NokogiriHandler).pick('href')
head.at_xpath("//link[nokogiri:link_rel_include(@rel, '#{name}')]", NokogiriHandler)&.attr('href')
end
def opengraph_tag(name)
document.xpath("//meta[@property=\"#{name}\" or @name=\"#{name}\"]").pick('content')
head.at_xpath("//meta[nokogiri:casecmp(@property, '#{name}') or nokogiri:casecmp(@name, '#{name}')]", NokogiriHandler)&.attr('content')
end
def meta_tag(name)
document.xpath("//meta[@name=\"#{name}\"]").pick('content')
head.at_xpath("//meta[nokogiri:casecmp(@name, '#{name}')]", NokogiriHandler)&.attr('content')
end
def structured_data
return @structured_data if defined?(@structured_data)
# Some publications have more than one JSON-LD definition on the page,
# and some of those definitions aren't valid JSON either, so we have
# to loop through here until we find something that is the right type
@ -273,6 +275,10 @@ class LinkDetailsExtractor
@document ||= detect_encoding_and_parse_document
end
def head
@head ||= document.at_xpath('/html/head')
end
def detect_encoding_and_parse_document
html = nil
encoding = nil

View file

@ -8,5 +8,9 @@ class NokogiriHandler
def link_rel_include(token_list, token)
token_list.to_s.downcase.split(WHITE_SPACE).include?(token.downcase)
end
def casecmp(str1, str2)
str1.to_s.casecmp?(str2.to_s)
end
end
end

View file

@ -49,7 +49,8 @@ RSpec.describe LinkDetailsExtractor do
<html lang="en">
<head>
<title>Man bites dog</title>
<meta name="description" content="A dog&#39;s tale">
<meta name="descripTION" content="A dog&#39;s tale">
<link rel="pretty IcoN" href="/favicon.ico">
</head>
</html>
HTML
@ -59,7 +60,8 @@ RSpec.describe LinkDetailsExtractor do
.to have_attributes(
title: eq('Man bites dog'),
description: eq("A dog's tale"),
language: eq('en')
language: eq('en'),
icon: eq('https://example.com/favicon.ico')
)
end
end
@ -256,7 +258,7 @@ RSpec.describe LinkDetailsExtractor do
<head>
<meta property="og:url" content="https://example.com/dog.html">
<meta property="og:title" content="Man bites dog">
<meta property="og:description" content="A dog's tale">
<meta property="OG:description" content="A dog's tale">
<meta property="article:published_time" content="2022-01-31T19:53:00+00:00">
<meta property="og:author" content="Charlie Brown">
<meta property="og:locale" content="en">