postrank-uri.rb in postrank-uri-1.0.9

- old
+ new

@@ -4,10 +4,25 @@
 require 'domainatrix'
 require 'digest/md5'
 require 'nokogiri'
 require 'yaml'
 
+module Addressable
+  class URI
+    def domain
+      begin
+        dp = Domainatrix.parse(self)
+      rescue
+        return nil
+      end
+
+      dom = dp.public_suffix
+      dom = dp.domain.downcase + "." + dom unless dp.domain.empty?
+    end
+  end
+end
+
 module PostRank
   module URI
 
     c18ndb = YAML.load_file(File.dirname(__FILE__) + '/postrank-uri/c18n.yml')
 
@@ -85,11 +100,11 @@
 
     def extract_href(text, host = nil)
       urls = []
       Nokogiri.HTML(text).search('a').each do |a|
         begin
-          url = clean(a.attr('href'), false)
+          url = clean(a.attr('href'), :raw => true)
           if url.host.empty?
             next if host.nil?
             url.host = host
           end
 
@@ -111,17 +126,17 @@
       uri.tr('+', ' ').gsub(URIREGEX[:unescape]) do
         [$1.delete('%')].pack('H*')
       end
     end
 
-    def clean(uri, string = true)
+    def clean(uri, opts = {})
       uri = normalize(c18n(unescape(uri)))
-      string ? uri.to_s : uri
+      opts[:raw] ? uri : uri.to_s
     end
 
-    def hash(uri)
-      Digest::MD5.hexdigest(clean(uri))
+    def hash(uri, opts = {})
+      Digest::MD5.hexdigest(opts[:skip_clean] ? uri : clean(uri))
     end
 
     def normalize(uri)
       u = parse(uri)
       u.path = u.path.squeeze('/')
@@ -156,10 +171,10 @@
 
       elsif uri.host =~ /myspace\.com/ && uri.path =~ /PostTo/
         embedded = uri.query_values['u']
       end
 
-      uri = clean(embedded, false) if embedded
+      uri = clean(embedded, :raw => true) if embedded
       uri
     end
 
     def parse(uri)
       return uri if uri.is_a? Addressable::URI
\ No newline at end of file