onebox 2.2.14 → 2.2.18

data/lib/onebox/engine/google_photos_onebox.rb

@@ -11,9 +11,9 @@ module Onebox
 
       def to_html
         og = get_opengraph
-        return video_html(og) if !og.video_secure_url.nil?
-        return album_html(og) if !og.type.nil? && og.type == "google_photos:photo_album"
-        return image_html(og) if !og.image.nil?
+        return video_html(og) if og.video_secure_url
+        return album_html(og) if og.type == "google_photos:photo_album"
+        return image_html(og) if og.image
         nil
       end
 
@@ -32,7 +32,7 @@ module Onebox
               <h3><a href="#{escaped_url}" target="_blank" rel="nofollow ugc noopener">#{og.title}</a></h3>
               <div class="aspect-image-full-size">
                 <a href="#{escaped_url}" target="_blank" rel="nofollow ugc noopener">
-                  <img src="#{og.get_secure_image}" class="scale-image"/>
+                  <img src="#{og.secure_image_url}" class="scale-image"/>
                   <span class="instagram-video-icon"></span>
                 </a>
               </div>
@@ -53,7 +53,7 @@ module Onebox
                   <span class='album-title'>#{Onebox::Helpers.truncate(album_title, 80)}</span>
                 </span>
               </span>
-              <img src='#{og.get_secure_image}' #{og.title_attr} height='#{og.image_height}' width='#{og.image_width}'>
+              <img src='#{og.secure_image_url}' #{og.title_attr} height='#{og.image_height}' width='#{og.image_width}'>
             </a>
           </div>
         HTML
@@ -64,7 +64,7 @@ module Onebox
 
         <<-HTML
           <a href='#{escaped_url}' target='_blank' rel='noopener' class="onebox">
-            <img src='#{og.get_secure_image}' #{og.title_attr} alt='Google Photos' height='#{og.image_height}' width='#{og.image_width}'>
+            <img src='#{og.secure_image_url}' #{og.title_attr} alt='Google Photos' height='#{og.image_height}' width='#{og.image_width}'>
           </a>
         HTML
       end

data/lib/onebox/engine/imgur_onebox.rb

@@ -40,7 +40,7 @@ module Onebox
                     <span class='album-title'>#{album_title}</span>
                   </span>
                 </span>
-                <img src='#{og.get_secure_image}' #{og.title_attr} height='#{og.image_height}' width='#{og.image_width}'>
+                <img src='#{og.secure_image_url}' #{og.title_attr} height='#{og.image_height}' width='#{og.image_width}'>
               </a>
             </div>
           HTML
@@ -58,7 +58,7 @@ module Onebox
 
         <<-HTML
             <a href='#{escaped_url}' target='_blank' rel='noopener' class="onebox">
-              <img src='#{og.get_secure_image.chomp("?fb")}' #{og.title_attr} alt='Imgur'>
+              <img src='#{og.secure_image_url.chomp("?fb")}' #{og.title_attr} alt='Imgur'>
             </a>
           HTML
       end

data/lib/onebox/engine/instagram_onebox.rb

@@ -18,9 +18,8 @@ module Onebox
         oembed = get_oembed
         raise "No oEmbed data found. Ensure 'facebook_app_access_token' is valid" if oembed.data.empty?
 
-        permalink = clean_url.gsub("/#{oembed.author_name}/", "/")
-
-        { link: permalink,
+        {
+          link: clean_url.gsub("/#{oembed.author_name}/", "/"),
           title: "@#{oembed.author_name}",
           image: oembed.thumbnail_url,
           description: Onebox::Helpers.truncate(oembed.title, 250),

data/lib/onebox/engine/pastebin_onebox.rb

@@ -30,29 +30,25 @@ module Onebox
       end
 
       def lines
-        return @lines if @lines
+        return @lines if defined?(@lines)
         response = Onebox::Helpers.fetch_response("http://pastebin.com/raw/#{paste_key}", redirect_limit: 1) rescue ""
         @lines = response.split("\n")
       end
 
       def paste_key
-        if uri.path =~ /\/raw\//
-          match = uri.path.match(/\/raw\/([^\/]+)/)
-          return match[1] if match && match[1]
-        elsif uri.path =~ /\/download\//
-          match = uri.path.match(/\/download\/([^\/]+)/)
-          return match[1] if match && match[1]
-        elsif uri.path =~ /\/embed\//
-          match = uri.path.match(/\/embed\/([^\/]+)/)
-          return match[1] if match && match[1]
+        regex = case uri
+                when /\/raw\//
+                  /\/raw\/([^\/]+)/
+                when /\/download\//
+                  /\/download\/([^\/]+)/
+                when /\/embed\//
+                  /\/embed\/([^\/]+)/
         else
-          match = uri.path.match(/\/([^\/]+)/)
-          return match[1] if match && match[1]
+                  /\/([^\/]+)/
         end
 
-        nil
-      rescue
-        nil
+        match = uri.path.match(regex)
+        match[1] if match && match[1]
       end
     end
   end

data/lib/onebox/engine/pdf_onebox.rb

@@ -12,25 +12,17 @@ module Onebox
       private
 
       def data
-        pdf_info = get_pdf_info
-        raise "Unable to read pdf file: #{@url}" if pdf_info.nil?
+        begin
+          size = Onebox::Helpers.fetch_content_length(@url)
+        rescue
+          raise "Unable to read pdf file: #{@url}"
+        end
 
-        result = { link: link,
-                   title: pdf_info[:name],
-                   filesize: pdf_info[:filesize]
-                  }
-        result
-      end
-
-      def get_pdf_info
-        uri = URI.parse(@url)
-        size = Onebox::Helpers.fetch_content_length(@url)
         {
+          link: link,
+          title: File.basename(uri.path),
           filesize: size ? Onebox::Helpers.pretty_filesize(size.to_i) : nil,
-          name: File.basename(uri.path)
         }
-      rescue
-        nil
       end
     end
   end

data/lib/onebox/engine/pubmed_onebox.rb

@@ -10,13 +10,14 @@ module Onebox
 
       private
 
-      def get_xml
+      def xml
+        return @xml if defined?(@xml)
         doc = Nokogiri::XML(URI.open(URI.join(@url, "?report=xml&format=text")))
         pre = doc.xpath("//pre")
-        Nokogiri::XML("<root>" + pre.text + "</root>")
+        @xml = Nokogiri::XML("<root>" + pre.text + "</root>")
       end
 
-      def authors_of_xml(xml)
+      def authors
         initials = xml.css("Initials").map { |x| x.content }
         last_names = xml.css("LastName").map { |x| x.content }
         author_list = (initials.zip(last_names)).map { |i, l| i + " " + l }
@@ -27,22 +28,25 @@ module Onebox
         author_list.join(", ")
       end
 
-      def date_of_xml(xml)
-        date_arr = (xml.css("PubDate").children).map { |x| x.content }
-        date_arr = date_arr.select { |s| !s.match(/^\s+$/) }
-        date_arr = (date_arr.map { |s| s.split }).flatten
-        date_arr.sort.reverse.join(" ") # Reverse sort so month before year.
+      def date
+        xml.css("PubDate")
+          .children
+          .map { |x| x.content }
+          .select { |s| !s.match(/^\s+$/) }
+          .map { |s| s.split }
+          .flatten
+          .sort
+          .reverse
+          .join(" ") # Reverse sort so month before year.
       end
 
       def data
-        xml = get_xml
-
         {
           title: xml.css("ArticleTitle").text,
-          authors: authors_of_xml(xml),
+          authors: authors,
           journal: xml.css("Title").text,
           abstract: xml.css("AbstractText").text,
-          date: date_of_xml(xml),
+          date: date,
           link: @url,
           pmid: match[:pmid]
         }

data/lib/onebox/engine/stack_exchange_onebox.rb

@@ -25,7 +25,7 @@ module Onebox
       end
 
       def url
-        domain = URI(@url).host
+        domain = uri.host
         question_id = match[:question_id]
         answer_id = match[:answer_id2] || match[:answer_id1]
 

data/lib/onebox/engine/standard_embed.rb

@@ -7,7 +7,6 @@ require 'onebox/oembed'
 module Onebox
   module Engine
     module StandardEmbed
-
       def self.oembed_providers
         @@oembed_providers ||= {}
       end
@@ -117,8 +116,6 @@ module Onebox
         "{}"
       end
 
-      protected
-
       def get_oembed_url
         oembed_url = nil
 

data/lib/onebox/engine/trello_onebox.rb

@@ -11,12 +11,11 @@ module Onebox
       always_https
 
       def to_html
-        link = "https://trello.com/#{match[:type]}/#{match[:key]}.html"
-
+        src = "https://trello.com/#{match[:type]}/#{match[:key]}.html"
         height = match[:type] == 'b' ? 400 : 200
 
         <<-HTML
-          <iframe src="#{link}" width="100%" height="#{height}" frameborder="0" style="border:0"></iframe>
+          <iframe src="#{src}" width="100%" height="#{height}" frameborder="0" style="border:0"></iframe>
         HTML
       end
 
@@ -25,12 +24,10 @@ module Onebox
       end
 
       private
+
       def match
         return @match if defined?(@match)
-
         @match = @url.match(%{trello\.com/(?<type>[^/]+)/(?<key>[^/]+)/?\W*})
-
-        @match
       end
     end
   end

data/lib/onebox/engine/youku_onebox.rb

@@ -30,12 +30,6 @@ module Onebox
           ></iframe>
         HTML
       end
-
-      private
-
-      def uri
-        @_uri ||= URI(@url)
-      end
     end
   end
 end

data/lib/onebox/helpers.rb

@@ -26,23 +26,7 @@ module Onebox
 
     def self.fetch_html_doc(url, headers = nil, body_cacher = nil)
       response = (fetch_response(url, headers: headers, body_cacher: body_cacher) rescue nil)
-      doc = Nokogiri::HTML(response)
-      uri = Addressable::URI.parse(url)
-
-      ignore_canonical_tag = doc.at('meta[property="og:ignore_canonical"]')
-      should_ignore_canonical = IGNORE_CANONICAL_DOMAINS.map { |hostname| uri.hostname.match?(hostname) }.any?
-
-      unless (ignore_canonical_tag && ignore_canonical_tag['content'].to_s == 'true') || should_ignore_canonical
-        # prefer canonical link
-        canonical_link = doc.at('//link[@rel="canonical"]/@href')
-        canonical_uri = Addressable::URI.parse(canonical_link)
-        if canonical_link && "#{canonical_uri.host}#{canonical_uri.path}" != "#{uri.host}#{uri.path}"
-          response = (fetch_response(canonical_uri.to_s, headers: headers, body_cacher: body_cacher) rescue nil)
-          doc = Nokogiri::HTML(response) if response
-        end
-      end
-
-      doc
+      Nokogiri::HTML(response)
     end
 
     def self.fetch_response(location, redirect_limit: 5, domain: nil, headers: nil, body_cacher: nil)
@@ -63,8 +47,7 @@ module Onebox
       end
 
       result = StringIO.new
-      Net::HTTP.start(uri.host, uri.port, use_ssl: uri.normalized_scheme == 'https') do |http|
-        http.open_timeout = Onebox.options.connect_timeout
+      Net::HTTP.start(uri.host, uri.port, open_timeout: Onebox.options.connect_timeout, use_ssl: uri.normalized_scheme == 'https') do |http|
         http.read_timeout = Onebox.options.timeout
         http.verify_mode = OpenSSL::SSL::VERIFY_NONE  # Work around path building bugs
 
@@ -90,7 +73,8 @@ module Onebox
 
           code = response.code.to_i
           unless code === 200
-            response.error! unless [301, 302].include?(code)
+            response.error! unless [301, 302, 303, 307, 308].include?(code)
+
             return fetch_response(
               response['location'],
               redirect_limit: redirect_limit - 1,
@@ -117,8 +101,7 @@ module Onebox
     def self.fetch_content_length(location)
       uri = URI(location)
 
-      Net::HTTP.start(uri.host, uri.port, use_ssl: uri.is_a?(URI::HTTPS)) do |http|
-        http.open_timeout = Onebox.options.connect_timeout
+      Net::HTTP.start(uri.host, uri.port, open_timeout: Onebox.options.connect_timeout, use_ssl: uri.is_a?(URI::HTTPS)) do |http|
         http.read_timeout = Onebox.options.timeout
         if uri.is_a?(URI::HTTPS)
           http.use_ssl = true

data/lib/onebox/layout.rb

@@ -32,19 +32,7 @@ module Onebox
     private
 
     def uri
-      @uri = URI(link)
-    end
-
-    def checksum
-      @md5.hexdigest("#{VERSION}:#{link}")
-    end
-
-    def link
-      ::Onebox::Helpers.normalize_url_for_output(record[:link])
-    end
-
-    def domain
-      record[:domain] || URI(link || '').host.to_s.sub(/^www\./, '')
+      @uri ||= URI(::Onebox::Helpers.normalize_url_for_output(record[:link]))
     end
 
     def details
@@ -52,7 +40,7 @@ module Onebox
         link: record[:link],
         title: record[:title],
         favicon: record[:favicon],
-        domain: domain,
+        domain: record[:domain] || uri.host.to_s.sub(/^www\./, ''),
         article_published_time: record[:article_published_time],
         article_published_time_title: record[:article_published_time_title],
         metadata_1_label: record[:metadata_1_label],

data/lib/onebox/matcher.rb

@@ -2,8 +2,12 @@
 
 module Onebox
   class Matcher
-    def initialize(link, options = {})
-      @url = link
+    def initialize(url, options = {})
+      begin
+        @uri = URI(url)
+      rescue URI::InvalidURIError
+      end
+
       @options = options
     end
 
@@ -14,12 +18,10 @@ module Onebox
     end
 
     def oneboxed
-      uri = URI(@url)
-      return unless uri.port.nil? || Onebox.options.allowed_ports.include?(uri.port)
-      return unless uri.scheme.nil? || Onebox.options.allowed_schemes.include?(uri.scheme)
-      ordered_engines.find { |engine| engine === uri && has_allowed_iframe_origins?(engine) }
-    rescue URI::InvalidURIError
-      nil
+      return if @uri.nil?
+      return if @uri.port && !Onebox.options.allowed_ports.include?(@uri.port)
+      return if @uri.scheme && !Onebox.options.allowed_schemes.include?(@uri.scheme)
+      ordered_engines.find { |engine| engine === @uri && has_allowed_iframe_origins?(engine) }
     end
 
     def has_allowed_iframe_origins?(engine)

data/lib/onebox/mixins/git_blob_onebox.rb

@@ -25,8 +25,8 @@ module Onebox
       }
 
       module InstanceMethods
-        def initialize(link, timeout = nil)
-          super link, timeout
+        def initialize(url, timeout = nil)
+          super url, timeout
           # merge engine options from global Onebox.options interface
           # self.options = Onebox.options["GithubBlobOnebox"] #  self.class.name.split("::").last.to_s
           # self.options = Onebox.options[self.class.name.split("::").last.to_s] #We can use this a more generic approach. extract the engine class name automatically
@@ -163,11 +163,9 @@ module Onebox
             @file = m[:file]
             @lang = Onebox::FileTypeFinder.from_file_name(m[:file])
 
-            if @lang == "stl" && link.match(/^https?:\/\/(www\.)?github\.com.*\/blob\//)
-
+            if @lang == "stl" && link.match?(/^https?:\/\/(www\.)?github\.com.*\/blob\//)
               @model_file = @lang.dup
               @raw = "https://render.githubusercontent.com/view/solid?url=" + self.raw_template(m)
-
             else
               contents = URI.open(self.raw_template(m), read_timeout: timeout).read
 

data/lib/onebox/open_graph.rb

@@ -17,10 +17,10 @@ module Onebox
       !title.nil? ? "title='#{title}'" : ""
     end
 
-    def get_secure_image
-      secure_link = URI(get(:image))
-      secure_link.scheme = 'https'
-      secure_link.to_s
+    def secure_image_url
+      secure_url = URI(get(:image))
+      secure_url.scheme = 'https'
+      secure_url.to_s
     end
 
     def method_missing(attr, *args, &block)

data/lib/onebox/preview.rb

@@ -6,8 +6,8 @@ module Onebox
     client_exception = defined?(Net::HTTPClientException) ? Net::HTTPClientException : Net::HTTPServerException
     WEB_EXCEPTIONS ||= [client_exception, OpenURI::HTTPError, Timeout::Error, Net::HTTPError, Errno::ECONNREFUSED]
 
-    def initialize(link, options = Onebox.options)
-      @url = link
+    def initialize(url, options = Onebox.options)
+      @url = url
       @options = options.dup
 
       allowed_origins = @options[:allowed_iframe_origins] || Onebox::Engine.all_iframe_origins

data/lib/onebox/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Onebox
-  VERSION = "2.2.14"
+  VERSION = "2.2.18"
 end