onebox 2.2.8 → 2.2.13

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 5899faff4b0c2e96d766892a810e866e28a188dbd7b9e8b2c7bdd6bc5a7d7490
-  data.tar.gz: 7d739d48349f32dc3f0675d85b29446d6f758f276b845543cc9f9d359c9eb664
+  metadata.gz: ee0229502d76efea853fff81a0e756e33867b7034219bb56805dcf891ab42207
+  data.tar.gz: 9e6f581f0dd784f0edb841c05f4faa785bae4cfaf4ac2c482050034ead36b912
 SHA512:
-  metadata.gz: 790964df42c83ea05ff052727a7ad4b567f6064be1a25bf974eeac661a6f7daa3696c64686f21bb83df30ad280caffc36b2039e3c0b799a3171713d61c88eff2
-  data.tar.gz: de30b5e5e6110fd72c0aa792aebffa7571c50f04604efedac648e754a527f4945a784328eb43ce0352e9b06eeac908ea23ff0dcdf3ce4d84ecb5e8c3dd64dfd6
+  metadata.gz: 7139dd42d40ebf30b4228c41018db4408fd53c196838c0b8d94afae27ba30d1ba258fc57fbea7ff432619f473a2e22c86b62251c094591f9a9ec90693b768655
+  data.tar.gz: 69d506c95f5fbf5526967195ce5f9c1224c1af71f72f0477063e4a8a0e59b8ab4300270cc28aae2cfc74f5e41187b3e3aed11b1656e83c0e766293eb932ca84a

data/.rspec

@@ -1,5 +1 @@
 --color
---format documentation
---fail-fast
---backtrace
-

data/lib/onebox/engine.rb

@@ -136,7 +136,7 @@ module Onebox
       end
 
       def always_https?
-        @https
+        defined?(@https) ? @https : false
       end
     end
   end

data/lib/onebox/engine/allowlisted_generic_onebox.rb

@@ -374,15 +374,17 @@ module Onebox
         escaped_image_src = ::Onebox::Helpers.normalize_url_for_output(data[:image])
 
         <<-HTML
-            <video title='#{data[:title]}'
-                   width='#{data[:video_width]}'
-                   height='#{data[:video_height]}'
-                   style='max-width:100%'
-                   poster='#{escaped_image_src}'
-                   controls=''>
-              <source src='#{escaped_video_src}'>
-            </video>
-          HTML
+          <video
+            title='#{data[:title]}'
+            width='#{data[:video_width]}'
+            height='#{data[:video_height]}'
+            style='max-width:100%'
+            poster='#{escaped_image_src}'
+            controls=''
+          >
+            <source src='#{escaped_video_src}'>
+          </video>
+        HTML
       end
 
       def embedded_html

data/lib/onebox/engine/amazon_onebox.rb

@@ -11,31 +11,46 @@ module Onebox
       include HTML
 
       always_https
-      matches_regexp(/^https?:\/\/(?:www\.)?(?:smile\.)?(amazon|amzn)\.(?<tld>com|ca|de|it|es|fr|co\.jp|co\.uk|cn|in|com\.br|com\.mx)\//)
+      matches_regexp(/^https?:\/\/(?:www\.)?(?:smile\.)?(amazon|amzn)\.(?<tld>com|ca|de|it|es|fr|co\.jp|co\.uk|cn|in|com\.br|com\.mx|nl|pl|sa|sg|se|com\.tr|ae)\//)
 
       def url
+        @raw ||= nil
+
+        # If possible, fetch the cached HTML body immediately so we can
+        # try to grab the canonical URL from that document,
+        # rather than guess at the best URL structure to use
+        if body_cacher&.respond_to?('cache_response_body?')
+          if body_cacher.cache_response_body?(uri.to_s) && body_cacher.cached_response_body_exists?(uri.to_s)
+            @raw ||= Onebox::Helpers.fetch_html_doc(@url, http_params, body_cacher)
+          end
+        end
+
+        if @raw
+          canonical_link = @raw.at('//link[@rel="canonical"]/@href')
+          return canonical_link.to_s if canonical_link
+        end
+
         if match && match[:id]
-          return "https://www.amazon.#{tld}/gp/aw/d/#{Onebox::Helpers.uri_encode(match[:id])}"
+          return "https://www.amazon.#{tld}/dp/#{Onebox::Helpers.uri_encode(match[:id])}"
         end
 
         @url
       end
 
       def tld
-        @tld || @@matcher.match(@url)["tld"]
+        @tld ||= @@matcher.match(@url)["tld"]
       end
 
       def http_params
-        {
-          'User-Agent' =>
-          'Mozilla/5.0 (iPhone; CPU iPhone OS 12_2 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Mobile/15E148'
-        }
+        if @options && @options[:user_agent]
+          { 'User-Agent' => @options[:user_agent] }
+        end
       end
 
       private
 
       def match
-        @match ||= @url.match(/(?:d|g)p\/(?:product\/|video\/detail\/)?(?<id>[^\/]+)(?:\/|$)/mi)
+        @match ||= @url.match(/(?:d|g)p\/(?:product\/|video\/detail\/)?(?<id>[A-Z0-9]+)(?:\/|\?|$)/mi)
       end
 
       def image
@@ -50,6 +65,10 @@ module Onebox
         end
 
         if (landing_image = raw.css("#landingImage")) && landing_image.any?
+          attributes = landing_image.first.attributes
+
+          return attributes["data-old-hires"].to_s if attributes["data-old-hires"]
+
           landing_image.first["src"].to_s
         end
 
@@ -100,7 +119,7 @@ module Onebox
           end
 
           result = {
-            link: link,
+            link: url,
             title: title,
             by_info: authors,
             image: og.image || image,
@@ -131,7 +150,7 @@ module Onebox
           end
 
           result = {
-            link: link,
+            link: url,
             title: title,
             by_info: authors,
             image: og.image || image,
@@ -147,7 +166,7 @@ module Onebox
         else
           title = og.title || CGI.unescapeHTML(raw.css("title").inner_text)
           result = {
-            link: link,
+            link: url,
             title: title,
             image: og.image || image,
             price: price
@@ -157,7 +176,10 @@ module Onebox
           result[:by_info] = Onebox::Helpers.clean(result[:by_info].inner_html) if result[:by_info]
 
           summary = raw.at("#productDescription")
-          result[:description] = og.description || (summary && summary.inner_text) || CGI.unescapeHTML(Onebox::Helpers.truncate(raw.css("meta[name=description]").first["content"], 250))
+
+          description = og.description || summary&.inner_text
+          description ||= raw.css("meta[name=description]").first&.[]("content")
+          result[:description] = CGI.unescapeHTML(Onebox::Helpers.truncate(description, 250)) if description
         end
 
         result[:price] = nil if result[:price].start_with?("$0") || result[:price] == 0

data/lib/onebox/engine/asciinema_onebox.rb

@@ -7,7 +7,6 @@ module Onebox
       include StandardEmbed
 
       always_https
-
       matches_regexp(/^https?:\/\/asciinema\.org\/a\/[\p{Alnum}_\-]+$/)
 
       def to_html

data/lib/onebox/engine/audioboom_onebox.rb

@@ -13,14 +13,17 @@ module Onebox
         oembed = get_oembed
 
         <<-HTML
-          <img src="#{oembed.thumbnail_url}" style="max-width: #{oembed.width}px; max-height: #{oembed.height}px;" #{oembed.title_attr}>
+          <img
+            src="#{oembed.thumbnail_url}"
+            style="max-width: #{oembed.width}px; max-height: #{oembed.height}px;"
+            #{oembed.title_attr}
+          >
         HTML
       end
 
       def to_html
         get_oembed.html
       end
-
     end
   end
 end

data/lib/onebox/engine/bandcamp_onebox.rb

@@ -20,16 +20,16 @@ module Onebox
         escaped_src = og.video_secure_url || og.video
 
         <<-HTML
-          <iframe src="#{escaped_src}"
-                  width="#{og.video_width}"
-                  height="#{og.video_height}"
-                  scrolling="no"
-                  frameborder="0"
-                  allowfullscreen>
-          </iframe>
+          <iframe
+            src="#{escaped_src}"
+            width="#{og.video_width}"
+            height="#{og.video_height}"
+            scrolling="no"
+            frameborder="0"
+            allowfullscreen
+          ></iframe>
         HTML
       end
-
     end
   end
 end

data/lib/onebox/engine/cloudapp_onebox.rb

@@ -25,28 +25,28 @@ module Onebox
 
       def link_html(og)
         <<-HTML
-            <a href='#{og.url}' target='_blank' rel='noopener'>
-              #{og.title}
-            </a>
-          HTML
+          <a href='#{og.url}' target='_blank' rel='noopener'>
+            #{og.title}
+          </a>
+        HTML
       end
 
       def video_html(og)
         direct_src = ::Onebox::Helpers.normalize_url_for_output("#{og.get(:url)}/#{og.title}")
 
         <<-HTML
-            <video width='480' height='360' #{og.title_attr} controls loop>
-              <source src='#{direct_src}' type='video/mp4'>
-            </video>
-          HTML
+          <video width='480' height='360' #{og.title_attr} controls loop>
+            <source src='#{direct_src}' type='video/mp4'>
+          </video>
+        HTML
       end
 
       def image_html(og)
         <<-HTML
-            <a href='#{og.url}' target='_blank' class='onebox' rel='noopener'>
-              <img src='#{og.image}' #{og.title_attr} alt='CloudApp' width='480'>
-            </a>
-          HTML
+          <a href='#{og.url}' target='_blank' class='onebox' rel='noopener'>
+            <img src='#{og.image}' #{og.title_attr} alt='CloudApp' width='480'>
+          </a>
+        HTML
       end
     end
   end

data/lib/onebox/engine/coub_onebox.rb

@@ -17,7 +17,6 @@ module Onebox
       def to_html
         get_oembed.html
       end
-
     end
   end
 end

data/lib/onebox/engine/facebook_media_onebox.rb

@@ -14,13 +14,14 @@ module Onebox
         metadata = get_twitter
         if metadata.present? && metadata[:card] == "player" && metadata[:player].present?
           <<-HTML
-            <iframe src="#{metadata[:player]}"
-                    width="#{metadata[:player_width]}"
-                    height="#{metadata[:player_height]}"
-                    scrolling="no"
-                    frameborder="0"
-                    allowfullscreen>
-            </iframe>
+            <iframe
+              src="#{metadata[:player]}"
+              width="#{metadata[:player_width]}"
+              height="#{metadata[:player_height]}"
+              scrolling="no"
+              frameborder="0"
+              allowfullscreen
+            ></iframe>
           HTML
         else
           html = Onebox::Engine::AllowlistedGenericOnebox.new(@url, @timeout).to_html

data/lib/onebox/engine/five_hundred_px_onebox.rb

@@ -13,7 +13,6 @@ module Onebox
         og = get_opengraph
         "<img src='#{og.image}' width='#{og.image_width}' height='#{og.image_height}' class='onebox' #{og.title_attr}>"
       end
-
     end
   end
 end

data/lib/onebox/engine/flickr_onebox.rb

@@ -25,29 +25,28 @@ module Onebox
         album_title = "[Album] #{og.title}"
 
         <<-HTML
-            <div class='onebox flickr-album'>
-              <a href='#{escaped_url}' target='_blank' rel='noopener'>
-                <span class='outer-box' style='max-width:#{og.image_width}px'>
-                  <span class='inner-box'>
-                    <span class='album-title'>#{album_title}</span>
-                  </span>
+          <div class='onebox flickr-album'>
+            <a href='#{escaped_url}' target='_blank' rel='noopener'>
+              <span class='outer-box' style='max-width:#{og.image_width}px'>
+                <span class='inner-box'>
+                  <span class='album-title'>#{album_title}</span>
                 </span>
-                <img src='#{og.get_secure_image}' #{og.title_attr} height='#{og.image_height}' width='#{og.image_width}'>
-              </a>
-            </div>
-          HTML
+              </span>
+              <img src='#{og.get_secure_image}' #{og.title_attr} height='#{og.image_height}' width='#{og.image_width}'>
+            </a>
+          </div>
+        HTML
       end
 
       def image_html(og)
         escaped_url = ::Onebox::Helpers.normalize_url_for_output(url)
 
         <<-HTML
-            <a href='#{escaped_url}' target='_blank' rel='noopener' class="onebox">
-              <img src='#{og.get_secure_image}' #{og.title_attr} alt='Imgur' height='#{og.image_height}' width='#{og.image_width}'>
-            </a>
-          HTML
+          <a href='#{escaped_url}' target='_blank' rel='noopener' class="onebox">
+            <img src='#{og.get_secure_image}' #{og.title_attr} alt='Imgur' height='#{og.image_height}' width='#{og.image_width}'>
+          </a>
+        HTML
       end
-
     end
   end
 end

data/lib/onebox/engine/gfycat_onebox.rb

@@ -63,7 +63,7 @@ module Onebox
 
       def nokogiri_page
         @nokogiri_page ||= begin
-          response = Onebox::Helpers.fetch_response(url, 10) rescue nil
+          response = Onebox::Helpers.fetch_response(url, redirect_limit: 10) rescue nil
           Nokogiri::HTML(response)
         end
       end

data/lib/onebox/engine/giphy_onebox.rb

@@ -18,7 +18,6 @@ module Onebox
           </a>
         HTML
       end
-
     end
   end
 end

data/lib/onebox/engine/github_blob_onebox.rb

@@ -8,17 +8,21 @@ module Onebox
       def self.git_regexp
         /^https?:\/\/(www\.)?github\.com.*\/blob\//
       end
+
       def self.onebox_name
         "githubblob"
       end
 
       include Onebox::Mixins::GitBlobOnebox
+
       def raw_regexp
         /github\.com\/(?<user>[^\/]+)\/(?<repo>[^\/]+)\/blob\/(?<sha1>[^\/]+)\/(?<file>[^#]+)(#(L(?<from>[^-]*)(-L(?<to>.*))?))?/mi
       end
+
       def raw_template(m)
         "https://raw.githubusercontent.com/#{m[:user]}/#{m[:repo]}/#{m[:sha1]}/#{m[:file]}"
       end
+
       def title
         Sanitize.fragment(Onebox::Helpers.uri_unencode(link).sub(/^https?\:\/\/github\.com\//, ''))
       end

data/lib/onebox/engine/github_commit_onebox.rb

@@ -17,11 +17,10 @@ module Onebox
       private
 
       def match
-        return @match if @match
+        return @match if defined?(@match)
 
         @match = @url.match(%{github\.com/(?<owner>[^/]+)/(?<repository>[^/]+)/commit/(?<sha>[^/]+)})
-
-        @match = @url.match(%{github\.com/(?<owner>[^/]+)/(?<repository>[^/]+)/pull/(?<pr>[^/]+)/commit/(?<sha>[^/]+)}) if @match.nil?
+        @match ||= @url.match(%{github\.com/(?<owner>[^/]+)/(?<repository>[^/]+)/pull/(?<pr>[^/]+)/commit/(?<sha>[^/]+)})
 
         @match
       end

data/lib/onebox/engine/github_gist_onebox.rb

@@ -9,7 +9,7 @@ module Onebox
 
       MAX_FILES = 3
 
-      matches_regexp Regexp.new("^http(?:s)?://gist\\.(?:(?:\\w)+\\.)?(github)\\.com(?:/)?")
+      matches_regexp(/^http(?:s)?:\/\/gist\.(?:(?:\w)+\.)?(github)\.com(?:\/)?/)
       always_https
 
       def url
@@ -76,7 +76,6 @@ module Onebox
           @lines ||= @json["content"].split("\n")
         end
       end
-
     end
   end
 end

data/lib/onebox/engine/github_issue_onebox.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require_relative '../mixins/github_body'
+
 module Onebox
   module Engine
     class GithubIssueOnebox
@@ -7,7 +9,9 @@ module Onebox
       include Engine
       include LayoutSupport
       include JSON
-      matches_regexp Regexp.new("^https?:\/\/(?:www\.)?(?:(?:\w)+\.)?github\.com\/(?<org>.+)\/(?<repo>.+)\/issues\/([[:digit:]]+)")
+      include Onebox::Mixins::GithubBody
+
+      matches_regexp(/^https?:\/\/(?:www\.)?(?:(?:\w)+\.)?github\.com\/(?<org>.+)\/(?<repo>.+)\/issues\/([[:digit:]]+)/)
       always_https
 
       def url
@@ -22,32 +26,26 @@ module Onebox
       end
 
       def data
-        @raw ||= ::MultiJson.load(URI.open(url, "Accept" => "application/vnd.github.v3.text+json", read_timeout: timeout)) #custom Accept header so we can get body as text.
-        body_text = @raw["body_text"]
-
-        content_words = body_text.gsub("\n\n", "\n").gsub("\n", "<br>").split(" ") #one pass of removing double newline, then we change \n to <br> and later on we revert it back to \n this is a workaround to avoid losing newlines after we join it back.
-        max_words = 20
-        short_content =  content_words[0..max_words].join(" ")
-        short_content += "..." if content_words.length > max_words
-
-        created_at = Time.parse(@raw['created_at'])
-        closed_at = Time.parse(@raw['closed_at']) if @raw['closed_at']
-
+        created_at = Time.parse(raw['created_at'])
+        closed_at = Time.parse(raw['closed_at']) if raw['closed_at']
+        body, excerpt = compute_body(raw['body'])
         ulink = URI(link)
+
         {
           link: @url,
-          title: @raw["title"],
-          content: short_content.gsub("<br>", "\n"),
-          labels: @raw["labels"],
-          user: @raw['user'],
+          title: raw["title"],
+          body: body,
+          excerpt: excerpt,
+          labels: raw["labels"],
+          user: raw['user'],
           created_at: created_at.strftime("%I:%M%p - %d %b %y %Z"),
           created_at_date: created_at.strftime("%F"),
           created_at_time: created_at.strftime("%T"),
           closed_at: closed_at&.strftime("%I:%M%p - %d %b %y %Z"),
           closed_at_date: closed_at&.strftime("%F"),
           closed_at_time: closed_at&.strftime("%T"),
-          closed_by: @raw['closed_by'],
-          avatar: "https://avatars1.githubusercontent.com/u/#{@raw['user']['id']}?v=2&s=96",
+          closed_by: raw['closed_by'],
+          avatar: "https://avatars1.githubusercontent.com/u/#{raw['user']['id']}?v=2&s=96",
           domain: "#{ulink.host}/#{ulink.path.split('/')[1]}/#{ulink.path.split('/')[2]}",
         }
       end