PageRankr 1.2.0 → 1.3.0

data/CHANGELOG.md

@@ -1,5 +1,9 @@
 # Change Log
 
+## Version 1.3.0
+
+* Lots of refactoring. Should be much easier to extend and temporarily fix if needed.
+
 ## Version 1.2.0
 
 * Changed backlinks method with no search engines specified to use all of them

data/PageRankr.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = %q{PageRankr}
-  s.version = "1.2.0"
+  s.version = "1.3.0"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Allen Madsen"]
-  s.date = %q{2010-07-04}
+  s.date = %q{2010-07-05}
   s.description = %q{Easy way to retrieve Google Page Rank, Alexa Rank, and backlink counts}
   s.email = %q{blatyo@gmail.com}
   s.extra_rdoc_files = [
@@ -26,10 +26,18 @@ Gem::Specification.new do |s|
      "Rakefile",
      "VERSION",
      "lib/page_rankr.rb",
-     "lib/page_rankr/alexa.rb",
      "lib/page_rankr/backlinks.rb",
-     "lib/page_rankr/google.rb",
-     "lib/page_rankr/google/checksum.rb"
+     "lib/page_rankr/backlinks/alexa.rb",
+     "lib/page_rankr/backlinks/alltheweb.rb",
+     "lib/page_rankr/backlinks/altavista.rb",
+     "lib/page_rankr/backlinks/backlink.rb",
+     "lib/page_rankr/backlinks/bing.rb",
+     "lib/page_rankr/backlinks/google.rb",
+     "lib/page_rankr/backlinks/yahoo.rb",
+     "lib/page_rankr/ranks.rb",
+     "lib/page_rankr/ranks/alexa.rb",
+     "lib/page_rankr/ranks/google.rb",
+     "lib/page_rankr/ranks/google/checksum.rb"
   ]
   s.homepage = %q{http://github.com/blatyo/page_rankr}
   s.rdoc_options = ["--charset=UTF-8"]

data/README.md

@@ -42,6 +42,44 @@ There are two valid rank trackers supported: `:alexa, :google`.
 Alexa ranks are descending where 1 is the most popular. If a site has an alexa rank of 0 then the site is unranked.
 Google page ranks are in the range 0-10 where 10 is the most popular. If a site is unindexed then the rank will be -1.
 
+## Fix it!
+
+If you ever find something is broken it should now be much easier to fix it with version >= 1.3.0. For example, if the xpath used to lookup a backlink is broken, just override the method for that class to provide the correct xpath.
+
+    module PageRankr
+      class Backlinks
+        class Google < Backlink
+          def xpath
+            "my new awesome xpath"
+          end
+        end
+      end
+    end
+
+## Extend it!
+
+If you ever come across a site that provides a rank or backlinks you can hook that class up to automatically be use with PageRankr.
+
+    module PageRankr
+      class Backlinks
+        class Foo < Backlink
+          def url(site)
+            "http://example.com/?q=#{site}"
+          end
+          
+          def xpath
+            "//backlinks/text()"
+          end
+          
+          def clean(backlink_count)
+            #do some of my own cleaning
+            super(backlink_count) # strips letters, commas, and a few other nasty things and converts it to an integer
+          end
+        end
+      end
+    end
+    
+Then, just make sure you require the class and PageRankr and whenver you call PageRankr.backlinks it'll be able to use your class.
 
 ## Note on Patches/Pull Requests
  

data/VERSION

@@ -1 +1 @@
-1.2.0
+1.3.0

data/lib/page_rankr.rb

@@ -1,28 +1,15 @@
-path = File.expand_path(File.dirname(__FILE__)) + '/page_rankr/'
-require path + 'backlinks'
-require path + 'alexa'
-require path + 'google'
+require File.join("page_rankr", "backlinks")
+require File.join("page_rankr", "ranks")
 
 module PageRankr
   class << self
     def backlinks(site, *search_engines)
-      Backlinks.lookup site, *search_engines
+      Backlinks.new.lookup site, *search_engines
     end
     alias_method :backlink, :backlinks
 
     def ranks(site, *rank_trackers)
-      rank_trackers = [:google, :alexa] if rank_trackers.empty?
-      
-      ranks = {}
-      rank_trackers.each do |tracker|
-        case tracker
-          when :google
-            ranks[tracker] = Google.lookup(site)
-          when :alexa
-            ranks[tracker] = Alexa.lookup(site)
-        end
-      end
-      ranks
+      Ranks.new.lookup site, *rank_trackers
     end
     alias_method :rank, :ranks
   end

data/lib/page_rankr/backlinks.rb

@@ -1,39 +1,30 @@
-require 'open-uri'
-require 'cgi'
-require 'nokogiri'
+require File.join("page_rankr", "backlinks", "backlink")
+require File.join("page_rankr", "backlinks", "alexa")
+require File.join("page_rankr", "backlinks", "alltheweb")
+require File.join("page_rankr", "backlinks", "altavista")
+require File.join("page_rankr", "backlinks", "bing")
+require File.join("page_rankr", "backlinks", "google")
+require File.join("page_rankr", "backlinks", "yahoo")
 
 module PageRankr
   class Backlinks
-    SEARCH_ENGINES = [:google, :bing, :yahoo, :altavista, :alltheweb, :alexa]
+    attr_accessor :search_engines
     
-    SEARCH_ENGINE_URLS = {
-      :google    => "http://www.google.com/search?q=link%3A",
-      :bing      => "http://www.bing.com/search?q=link%3A",
-      :yahoo     => "http://siteexplorer.search.yahoo.com/search?p=",
-      :altavista => "http://www.altavista.com/web/results?q=link%3A",
-      :alltheweb => "http://www.alltheweb.com/search?q=link%3A",
-      :alexa     => "http://data.alexa.com/data?cli=10&dat=snbamz&url="
-    }
-
-    SEARCH_EGNINE_PATHS = {
-      :google    => "//div[@id='resultStats']/text()",
-      :bing      => "//span[@class='sb_count']/text()",
-      :yahoo     => "//ol[@id='results-tab']/li[2]/a/text()",
-      :altavista => "//a[@class='lbl']/text()",
-      :alltheweb => "//span[@class='ofSoMany']/text()",
-      :alexa     => "//linksin/@num"
-    }
+    def initialize
+      @search_engines = self.class.constants
+      @search_engines.delete(:Backlink)
+    end
 
-    def self.lookup(site, *search_engines)
-      search_engines = SEARCH_ENGINES if search_engines.empty?
+    def lookup(site, *engines)
+      engines = search_engines if engines.empty?
       
       backlinks = {}
-      search_engines.each do |engine|
-        next unless SEARCH_ENGINE_URLS[engine]
-        doc = Nokogiri::HTML(open(SEARCH_ENGINE_URLS[engine] + CGI.escape(site)))
-        count = doc.at(SEARCH_EGNINE_PATHS[engine]).to_s
-        count = count.gsub('1-10', '').gsub(/[a-zA-Z,\s\(\)]/, '')
-        backlinks[engine] = count.to_i
+      engines.each do |engine|
+        name, klass = engine.to_s.capitalize, self.class
+        
+        next unless klass.const_defined? name
+        
+        backlinks[engine.to_s.downcase.to_sym] = klass.const_get(name).new(site).backlinks
       end
       backlinks
     end

data/lib/page_rankr/backlinks/alexa.rb

@@ -0,0 +1,15 @@
+require 'cgi'
+
+module PageRankr
+  class Backlinks
+    class Alexa < Backlink
+      def url(site)
+        "http://data.alexa.com/data?cli=10&dat=snbamz&url=#{CGI.escape(site)}"
+      end
+      
+      def xpath
+        "//linksin/@num"
+      end
+    end
+  end
+end

data/lib/page_rankr/backlinks/alltheweb.rb

@@ -0,0 +1,15 @@
+require 'cgi'
+
+module PageRankr
+  class Backlinks
+    class Alltheweb < Backlink
+      def url(site)
+        "http://www.alltheweb.com/search?q=link%3A#{CGI.escape(site)}"
+      end
+      
+      def xpath
+        "//span[@class='ofSoMany']/text()"
+      end
+    end
+  end
+end

data/lib/page_rankr/backlinks/altavista.rb

@@ -0,0 +1,15 @@
+require 'cgi'
+
+module PageRankr
+  class Backlinks
+    class Altavista < Backlink
+      def url(site)
+        "http://www.altavista.com/web/results?q=link%3A#{CGI.escape(site)}"
+      end
+      
+      def xpath
+        "//a[@class='lbl']/text()"
+      end
+    end
+  end
+end

data/lib/page_rankr/backlinks/backlink.rb

@@ -0,0 +1,18 @@
+require 'nokogiri'
+require 'open-uri'
+
+module PageRankr
+  class Backlinks
+    class Backlink
+      attr_reader :backlinks
+      
+      def initialize(site)
+        @backlinks = clean Nokogiri::HTML(open url(site)).at(xpath).to_s
+      end
+      
+      def clean(backlink_count)
+        backlink_count.gsub(/[a-zA-Z,\s\(\)]/, '').to_i
+      end
+    end
+  end
+end

data/lib/page_rankr/backlinks/bing.rb

@@ -0,0 +1,19 @@
+require 'cgi'
+
+module PageRankr
+  class Backlinks
+    class Bing < Backlink
+      def url(site)
+        "http://www.bing.com/search?q=link%3A#{CGI.escape(site)}"
+      end
+      
+      def xpath
+        "//span[@class='sb_count']/text()"
+      end
+      
+      def clean(backlink_count)
+        super(backlink_count.gsub('1-10', ''))
+      end
+    end
+  end
+end

data/lib/page_rankr/backlinks/google.rb

@@ -0,0 +1,15 @@
+require 'cgi'
+
+module PageRankr
+  class Backlinks
+    class Google < Backlink
+      def url(site)
+        "http://www.google.com/search?q=link%3A#{CGI.escape(site)}"
+      end
+      
+      def xpath
+        "//div[@id='resultStats']/text()"
+      end
+    end
+  end
+end

data/lib/page_rankr/backlinks/yahoo.rb

@@ -0,0 +1,15 @@
+require 'cgi'
+
+module PageRankr
+  class Backlinks
+    class Yahoo < Backlink
+      def url(site)
+        "http://siteexplorer.search.yahoo.com/search?p=#{CGI.escape(site)}"
+      end
+      
+      def xpath
+        "//ol[@id='results-tab']/li[2]/a/text()"
+      end
+    end
+  end
+end

data/lib/page_rankr/ranks.rb

@@ -0,0 +1,26 @@
+require File.join("page_rankr", "ranks", "alexa")
+require File.join("page_rankr", "ranks", "google")
+
+module PageRankr
+  class Ranks
+    attr_reader :rank_trackers
+    
+    def initialize
+      @rank_trackers = self.class.constants
+    end
+    
+    def lookup(site, *trackers)
+      trackers = rank_trackers if trackers.empty?
+      
+      ranks = {}
+      trackers.each do |tracker|
+        name, klass = tracker.to_s.capitalize, self.class
+        
+        next unless klass.const_defined? name
+        
+        ranks[tracker.to_s.downcase.to_sym] = klass.const_get(name).new(site).rank
+      end
+      ranks
+    end
+  end
+end

data/lib/page_rankr/ranks/alexa.rb

@@ -0,0 +1,23 @@
+require 'open-uri'
+require 'cgi'
+require 'nokogiri'
+
+module PageRankr
+  class Ranks
+    class Alexa
+      attr_reader :rank
+    
+      def initialize(site)
+        @rank = Nokogiri::HTML(open(url(site))).search(xpath).to_s.to_i
+      end
+      
+      def xpath
+        "//reach/@rank"
+      end
+      
+      def url(site)
+        "http://data.alexa.com/data?cli=10&dat=snbamz&url=#{CGI.escape(site)}"
+      end
+    end
+  end
+end

data/lib/page_rankr/{google.rb → ranks/google.rb}

@@ -1,20 +1,23 @@
-path = File.expand_path(File.dirname(__FILE__)) + '/google/'
 require "open-uri"
-require path + 'checksum'
+require File.join("page_rankr", "ranks", "google", "checksum")
 
 module PageRankr
-  module Google
-    class << self
-      def lookup(site)
+  class Ranks
+    class Google
+      attr_reader :rank
+    
+      def initialize(site)
         checksum = Checksum.generate(site)
         begin
-          open(url(site, checksum)) {|io| io.read.scan(/Rank_\d+:\d+:(\d+)/)[0][0].to_i}
+          @rank = open(url(site, checksum)) {|io| io.read.scan(regex)[0][0].to_i}
         rescue
           -1
         end
       end
-
-      private
+      
+      def regex
+        /Rank_\d+:\d+:(\d+)/
+      end
 
       def url(site, checksum)
         "http://toolbarqueries.google.com/search?client=navclient-auto&ch=#{checksum}&features=Rank&q=info:#{site}"

data/lib/page_rankr/ranks/google/checksum.rb

@@ -0,0 +1,76 @@
+module PageRankr
+  class Ranks
+    class Google
+      class Checksum
+        class << self
+          def generate(site)
+            bytes  = byte_array('info:' + site)
+            length = bytes.length
+            a = b = 0x9E3779B9
+            c = 0xE6359A60
+            k, len = 0, length
+            while(len >= 12)
+              a = m(a + bytes[k + 0] + (bytes[k + 1] << 8) + (bytes[k +  2] << 16) + (bytes[k +  3] << 24))
+              b = m(b + bytes[k + 4] + (bytes[k + 5] << 8) + (bytes[k +  6] << 16) + (bytes[k +  7] << 24))
+              c = m(c + bytes[k + 8] + (bytes[k + 9] << 8) + (bytes[k + 10] << 16) + (bytes[k + 11] << 24))
+
+              a, b, c = mix(a, b, c)
+              k += 12
+              len -= 12
+            end
+
+            c = c + length
+
+            c = mix(*toss(a, b, c, bytes, len, k))[2]
+            "6" + c.to_s
+          end
+
+          private
+
+          def byte_array(site)
+            bytes = []
+            site.each_byte {|b| bytes << b}
+            bytes
+          end
+
+          # Need to keep numbers in the unsigned int 32 range
+          def m(v)
+            v % 0x100000000
+          end
+
+          def mix(a, b, c)
+            a, b, c = m(a), m(b), m(c)
+
+            a = m(a-b-c) ^ m(c >> 13)
+            b = m(b-c-a) ^ m(a << 8)
+            c = m(c-a-b) ^ m(b >> 13)
+
+            a = m(a-b-c) ^ m(c >> 12)
+            b = m(b-c-a) ^ m(a << 16)
+            c = m(c-a-b) ^ m(b >> 5)
+
+            a = m(a-b-c) ^ m(c >> 3)
+            b = m(b-c-a) ^ m(a << 10)
+            c = m(c-a-b) ^ m(b >> 15)
+
+            [a, b, c]
+          end
+
+          def toss(a, b, c, bytes, len, k)
+            case len
+              when 9..11
+                c = c + (bytes[k+len-1] << ((len % 8) * 8))
+              when 5..8
+                b = b + (bytes[k+len-1] << ((len % 5) * 8))
+              when 1..4
+                a = a + (bytes[k+len-1] << ((len - 1) * 8))
+              else
+                return [a, b, c]
+            end
+            toss(a, b, c, bytes, len-1, k)
+          end
+        end
+      end
+    end
+  end
+end

metadata

@@ -4,9 +4,9 @@ version: !ruby/object:Gem::Version
   prerelease: false
   segments: 
   - 1
-  - 2
+  - 3
   - 0
-  version: 1.2.0
+  version: 1.3.0
 platform: ruby
 authors: 
 - Allen Madsen
@@ -14,7 +14,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2010-07-04 00:00:00 -04:00
+date: 2010-07-05 00:00:00 -04:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -64,10 +64,18 @@ files:
 - Rakefile
 - VERSION
 - lib/page_rankr.rb
-- lib/page_rankr/alexa.rb
 - lib/page_rankr/backlinks.rb
-- lib/page_rankr/google.rb
-- lib/page_rankr/google/checksum.rb
+- lib/page_rankr/backlinks/alexa.rb
+- lib/page_rankr/backlinks/alltheweb.rb
+- lib/page_rankr/backlinks/altavista.rb
+- lib/page_rankr/backlinks/backlink.rb
+- lib/page_rankr/backlinks/bing.rb
+- lib/page_rankr/backlinks/google.rb
+- lib/page_rankr/backlinks/yahoo.rb
+- lib/page_rankr/ranks.rb
+- lib/page_rankr/ranks/alexa.rb
+- lib/page_rankr/ranks/google.rb
+- lib/page_rankr/ranks/google/checksum.rb
 has_rdoc: true
 homepage: http://github.com/blatyo/page_rankr
 licenses: []

data/lib/page_rankr/alexa.rb

@@ -1,14 +0,0 @@
-require 'open-uri'
-require 'cgi'
-require 'nokogiri'
-
-module PageRankr
-  class Alexa
-    class << self
-      def lookup(site)
-        url = "http://data.alexa.com/data?cli=10&dat=snbamz&url="
-        Nokogiri::HTML(open(url + CGI.escape(site))).search("//reach/@rank").to_s.to_i
-      end
-    end
-  end
-end

data/lib/page_rankr/google/checksum.rb

@@ -1,74 +0,0 @@
-module PageRankr
-  module Google
-    class Checksum
-      class << self
-        def generate(site)
-          bytes  = byte_array('info:' + site)
-          length = bytes.length
-          a = b = 0x9E3779B9
-          c = 0xE6359A60
-          k, len = 0, length
-          while(len >= 12)
-            a = m(a + bytes[k + 0] + (bytes[k + 1] << 8) + (bytes[k +  2] << 16) + (bytes[k +  3] << 24))
-            b = m(b + bytes[k + 4] + (bytes[k + 5] << 8) + (bytes[k +  6] << 16) + (bytes[k +  7] << 24))
-            c = m(c + bytes[k + 8] + (bytes[k + 9] << 8) + (bytes[k + 10] << 16) + (bytes[k + 11] << 24))
-
-            a, b, c = mix(a, b, c)
-            k += 12
-            len -= 12
-          end
-
-          c = c + length
-
-          c = mix(*toss(a, b, c, bytes, len, k))[2]
-          "6" + c.to_s
-        end
-
-        private
-
-        def byte_array(site)
-          bytes = []
-          site.each_byte {|b| bytes << b}
-          bytes
-        end
-
-        # Need to keep numbers in the unsigned int 32 range
-        def m(v)
-          v % 0x100000000
-        end
-
-        def mix(a, b, c)
-          a, b, c = m(a), m(b), m(c)
-
-          a = m(a-b-c) ^ m(c >> 13)
-          b = m(b-c-a) ^ m(a << 8)
-          c = m(c-a-b) ^ m(b >> 13)
-
-          a = m(a-b-c) ^ m(c >> 12)
-          b = m(b-c-a) ^ m(a << 16)
-          c = m(c-a-b) ^ m(b >> 5)
-
-          a = m(a-b-c) ^ m(c >> 3)
-          b = m(b-c-a) ^ m(a << 10)
-          c = m(c-a-b) ^ m(b >> 15)
-
-          [a, b, c]
-        end
-
-        def toss(a, b, c, bytes, len, k)
-          case len
-            when 9..11
-              c = c + (bytes[k+len-1] << ((len % 8) * 8))
-            when 5..8
-              b = b + (bytes[k+len-1] << ((len % 5) * 8))
-            when 1..4
-              a = a + (bytes[k+len-1] << ((len - 1) * 8))
-            else
-              return [a, b, c]
-          end
-          toss(a, b, c, bytes, len-1, k)
-        end
-      end
-    end
-end
-end