monkeyhelper-muddyit_fu 0.1.1 → 0.2.0

data/README.rdoc

@@ -2,7 +2,7 @@
 
 == Installation
 
-sudo gem install muddyit-fu
+  sudo gem install monkeyhelper-muddyit-fu --source http://gems.github.com
 
 == Getting started
 
@@ -56,7 +56,7 @@ can be stored in a yml file, an example of which is provided below.
                          :access_token_secret => 'ddd')
   site = muddyit.sites.first
   site.pages.index do |page|
-    puts page.content_data.title
+    puts page.title
     page.results.each do |result|
       puts result.uri
     end
@@ -76,7 +76,7 @@ can be stored in a yml file, an example of which is provided below.
   require 'muddyit_fu'
   muddyit = Muddyit.new('muddyit.yml')
   site = muddyit.sites.first
-  site.pages.related_entities('http://dbpedia.org/resource/Gordon_Brown').each do |entity|
+  site.entities.related('http://dbpedia.org/resource/Gordon_Brown').each do |entity|
     puts "#{entity.uri} : #{entity.confidence}"
   end
 
@@ -86,9 +86,9 @@ can be stored in a yml file, an example of which is provided below.
   muddyit = Muddyit.new('muddyit.yml')
   site = muddyit.sites.first
   page = site.pages.find('http://news.bbc.co.uk/1/hi/uk_politics/7878418.stm')
-  puts "Our page : #{page.content_data.title}\n\n"
+  puts "Our page : #{page.title}\n\n"
   page.related_content.each do |results|
-    puts "#{results[:page].content_data.title} #{results[:count]}"
+    puts "#{results[:page].title} #{results[:count]}"
   end
 
 == Authorising clients using irb

data/Rakefile

@@ -10,7 +10,7 @@ begin
     gem.homepage = "http://github.com/monkeyhelper/muddyit_fu"
     gem.authors = ["robl"]
     gem.add_dependency('json', '>= 0.0.0')
-    gem.add_dependency('oauth', '>= 0.0.0')
+#    gem.add_dependency('oauth', '>= 0.0.0')
 
     # gem is a Gem::Specification... see http://www.rubygems.org/read/chapter/20 for additional settings
   end

data/VERSION

@@ -1 +1 @@
-0.1.1
+0.2.0

data/lib/muddyit/entities.rb

@@ -0,0 +1,3 @@
+class Entities
+  # Placeholder
+end

data/lib/muddyit/sites/entities/entity.rb

@@ -0,0 +1,35 @@
+class Muddyit::Sites::Site::Entities::Entity < Muddyit::Generic
+
+  def classification
+    unless @attributes[:type]
+      # We merge here as we don't want to overwrite a entity specific confidence score
+      @attributes.merge!(self.fetch)
+    end
+    @attributes[:type]
+  end
+
+  # retrieve entities related to the specified entity within the site entities collection
+  #
+  # Params
+  # * options (Optional)
+  #
+  def related(options = {})
+    api_url = "/sites/#{self.site.attributes[:token]}/entities/#{Digest::MD5.hexdigest(URI.encode(@attributes[:uri]))}/related"
+    response = @muddyit.send_request(api_url, :get, options)
+
+    results = []
+    response.each { |result|
+      # The return format needs sorting out here .......
+      results.push Muddyit::Sites::Site::Entities::Entity.new(@muddyit, result)
+    }
+    return results
+  end
+
+  protected
+  def fetch
+    api_url = "/sites/#{@attributes[:site][:token]}/entities/#{Digest::MD5.hexdigest(URI.encode(@attributes[:uri]))}"
+    response = @muddyit.send_request(api_url, :get)
+    response.nested_symbolize_keys!
+  end
+  
+end

data/lib/muddyit/sites/entities.rb

@@ -0,0 +1,22 @@
+class Muddyit::Sites::Site::Entities
+  # Placeholder
+
+  # retrieve entities related to the specified entity within the site entities collection
+  #
+  # Params
+  # * options (Optional)
+  #
+  def related(uri, options = {})
+    raise if uri.nil?
+    api_url = "/sites/#{self.site.attributes[:token]}/entities/#{Digest::MD5.hexdigest(URI.encode(uri))}/related"
+    response = @muddyit.send_request(api_url, :get, options)
+
+    results = []
+    response.each { |result|
+      # The return format needs sorting out here .......
+      results.push Muddyit::Sites::Site::Entities::Entity.new(@muddyit, result)
+    }
+    return results
+  end
+  
+end

data/lib/muddyit/{content_data.rb → sites/pages/page/extracted_content.rb}

@@ -1,4 +1,4 @@
-class Muddyit::Sites::Site::Page::ContentData < Muddyit::Generic
+class Muddyit::Sites::Site::Pages::Page::ExtractedContent < Muddyit::Generic
 
   def initialize(muddyit, attributes)
     super(muddyit, attributes)

data/lib/muddyit/{page.rb → sites/pages/page.rb}

@@ -1,10 +1,10 @@
-class Muddyit::Sites::Site::Page < Muddyit::Generic
+class Muddyit::Sites::Site::Pages::Page < Muddyit::Generic
 
   # Create a set of entities from the categorisation results
   def initialize(muddyit, attributes = {})
     super(muddyit, attributes)
     create_entities
-    @content_data_cache = nil
+    @extracted_content_cache = nil
   end
 
   # submit a page or text for re-categorisation
@@ -12,38 +12,38 @@ class Muddyit::Sites::Site::Page < Muddyit::Generic
   # Params
   # * options (Required)
   #
-  def refresh(options = {})
+  def update(options = {})
 
-    # Ensure we get content_data as well
+    # Ensure we get extracted_content as well
     options[:include_content] = true
 
     body = { :page => { :uri => self.uri }, :options => options }
 
-    api_url = "/sites/#{self.site.attributes[:token]}/pages/#{URI.escape(CGI.escape(self.identifier),'.')}"
+    api_url = "/sites/#{self.site.attributes[:token]}/pages/#{self.identifier}"
     response = @muddyit.send_request(api_url, :put, {}, body.to_json)
-    return Muddyit::Sites::Site::Page.new(@muddyit, response.merge!(:site => self.site))
+    return Muddyit::Sites::Site::Pages::Page.new(@muddyit, response['page'].merge!(:site => self.site))
   end
 
 
-  # get content_data for page
+  # get extracted_content for page
   #
-  def content_data
-    if @content_data_cache.nil?
-      if @attributes[:content_data]
-        @content_data_cache = Muddyit::Sites::Site::Page::ContentData.new(@muddyit, @attributes[:content_data])
+  def extracted_content
+    if @extracted_content_cache.nil?
+      if @attributes[:extracted_content]
+        @extracted_content_cache = Muddyit::Sites::Site::Pages::Page::ExtractedContent.new(@muddyit, @attributes[:extracted_content])
       else
         r = self.fetch
-        @content_data_cache = Muddyit::Sites::Site::Page::ContentData.new(@muddyit, r[:content_data])
+        @extracted_content_cache = Muddyit::Sites::Site::Pages::Page::ExtractedContent.new(@muddyit, r[:extracted_content])
       end
     end
-    @content_data_cache
+    @extracted_content_cache
   end
 
   
   # delete the page
   #
   def destroy
-    api_url = "/sites/#{self.site.attributes[:token]}/pages/#{URI.escape(CGI.escape(@attributes[:identifier]),'.')}"
+    api_url = "/sites/#{self.site.attributes[:token]}/pages/#{@attributes[:identifier]}"
     response = @muddyit.send_request(api_url, :delete, {})
     # Is this the correct thing to return ?
     return true
@@ -55,8 +55,8 @@ class Muddyit::Sites::Site::Page < Muddyit::Generic
   # * options (Optional)
   #
   def related_content(options = {})
-    api_url = "/sites/#{self.site.attributes[:token]}/pages/#{URI.escape(CGI.escape(@attributes[:identifier]),'.')}/related/content"
-    response = @muddyit.send_request(api_url, :get, options)
+    api_url = "/sites/#{self.site.attributes[:token]}/pages/#{@attributes[:identifier]}/related"
+    response = @muddyit.send_request(api_url, :get, options, nil)
     results = []
     response.each { |result|
       # The return format needs sorting out here .......
@@ -67,8 +67,10 @@ class Muddyit::Sites::Site::Page < Muddyit::Generic
 
   protected
   def fetch
-    api_url = "/sites/#{self.site.attributes[:token]}/pages/#{URI.escape(CGI.escape(@attributes[:identifier]),'.')}"
-    response = @muddyit.send_request(api_url, :get, {:include_content => true })
+    api_url = "/sites/#{self.site.attributes[:token]}/pages/#{@attributes[:identifier]}"
+    
+    response = @muddyit.send_request(api_url, :get, {:include_content => true}, nil)
+
     response.nested_symbolize_keys!
   end
 
@@ -77,7 +79,7 @@ class Muddyit::Sites::Site::Page < Muddyit::Generic
     results = []
     if @attributes.has_key?(:entities)
       @attributes[:entities].each do |result|
-         results.push Muddyit::Entity.new(@muddyit, result)
+         results.push Muddyit::Sites::Site::Entities::Entity.new(@muddyit, result)
       end
       @attributes[:entities] = results
     end

data/lib/muddyit/{pages.rb → sites/pages.rb}

@@ -18,7 +18,7 @@ class Muddyit::Sites::Site::Pages < Muddyit::Generic
           begin
           response = @muddyit.send_request(api_url, :get, options.merge!(:page => token))
           response['pages'].each { |page|
-            yield Muddyit::Sites::Site::Page.new(@muddyit, page.merge!(:site => self.site))
+            yield Muddyit::Sites::Site::Pages::Page.new(@muddyit, page.merge!(:site => self.site))
           }
           token = response['next_page']
           # Need to figure out which of the below actually occurs
@@ -28,7 +28,7 @@ class Muddyit::Sites::Site::Pages < Muddyit::Generic
           response = @muddyit.send_request(api_url, :get, options)
 
           pages = []
-          response['pages'].each { |page| pages.push Muddyit::Sites::Site::Page.new(@muddyit, page.merge!(:site => self.site)) }
+          response['pages'].each { |page| pages.push Muddyit::Sites::Site::Pages::Page.new(@muddyit, page.merge!(:site => self.site)) }
           return { :next_page => response['next_page'], :pages => pages }
         end
       else
@@ -36,36 +36,20 @@ class Muddyit::Sites::Site::Pages < Muddyit::Generic
       end
 
     elsif type.is_a? String
-      api_url = "/sites/#{self.site.attributes[:token]}/pages/#{URI.escape(CGI.escape(type),'.')}"
+      api_url = "/sites/#{self.site.attributes[:token]}/pages/#{type}"
       response = @muddyit.send_request(api_url, :get, {})
-      response.has_key?('identifier') ? Muddyit::Sites::Site::Page.new(@muddyit, response.merge!(:site => self.site)) : nil
+      response.has_key?('identifier') ? Muddyit::Sites::Site::Pages::Page.new(@muddyit, response.merge!(:site => self.site)) : nil
     end
   end
 
-  # retrieve entities related to the specified entity within the site pages collection
-  #
-  # Params
-  # * options (Optional)
-  #
-  def related_entities(uri, options = {})
-    raise "no uri supplied" if uri.nil?
-    api_url = "/sites/#{self.site.attributes[:token]}/related/entities/#{URI.escape(CGI.escape(uri),'.')}"
-    response = @muddyit.send_request(api_url, :get, options)
-
-    results = []
-    response.each { |result|
-      # The return format needs sorting out here .......
-      results.push Muddyit::Entity.new(@muddyit, result)
-    }
-    return results
-  end
+
 
   # submit a page or text for categorisation
   #
   # Params
   # * options (Required)
   #
-  def categorise(doc = {}, options = {})
+  def create(doc = {}, options = {})
 
     # Ensure we get content_data as well
     options[:include_content] = true
@@ -79,7 +63,7 @@ class Muddyit::Sites::Site::Pages < Muddyit::Generic
 
     api_url = "/sites/#{self.site.attributes[:token]}/pages/"
     response = @muddyit.send_request(api_url, :post, {}, body.to_json)
-    return Muddyit::Sites::Site::Page.new(@muddyit, response.merge!(:site => self.site))
+    return Muddyit::Sites::Site::Pages::Page.new(@muddyit, response['page'].merge!(:site => self.site))
   end
 
   # find all pages with specified entity
@@ -141,7 +125,7 @@ class Muddyit::Sites::Site::Pages < Muddyit::Generic
   #     must contain uri parameter which corresponds to dbpedia uri
   #
   def queryAllWithURI(uri, options, &block)
-    api_url = "/sites/#{self.site.attributes[:token]}/entities/#{URI.escape(CGI.escape(uri),'.')}"
+    api_url = "/sites/#{self.site.attributes[:token]}/entities/#{Digest::MD5.hexdigest(URI.encode(uri))}"
     query_page(api_url, options, &block)
   end
 
@@ -171,7 +155,7 @@ class Muddyit::Sites::Site::Pages < Muddyit::Generic
         options.merge!(:page => token) unless token.nil?
         response = @muddyit.send_request(api_url, :get, options.merge!(:page => token))
         response['pages'].each { |page|
-          yield Muddyit::Sites::Site::Page.new(@muddyit, page.merge!(:site => self.site))
+          yield Muddyit::Sites::Site::Pages::Page.new(@muddyit, page.merge!(:site => self.site))
         }
         token = response['next_page']
         # Need to figure out which of the below actually occurs
@@ -180,7 +164,7 @@ class Muddyit::Sites::Site::Pages < Muddyit::Generic
       response = @muddyit.send_request(api_url, :get, {})
 
       pages = []
-      response['pages'].each { |page| pages.push Muddyit::Sites::Site::Page.new(@muddyit, page.merge!(:site => self.site)) }
+      response['pages'].each { |page| pages.push Muddyit::Sites::Site::Pages::Page.new(@muddyit, page.merge!(:site => self.site)) }
       return { :next_page => response[:next_page], :pages => pages }
     end    
   end

data/lib/muddyit_fu.rb

@@ -3,7 +3,9 @@ require 'net/http'
 require 'cgi'
 require 'json'
 #require 'json/ext'
+#gem 'monkeyhelper-oauth', :lib => 'lib/oauth'
 require 'oauth/consumer'
+require 'digest/md5'
 
 require 'pp'
 
@@ -44,7 +46,7 @@ class Hash
 end
 
 # base must load first
-%w(base errors generic sites entity site pages page content_data).each do |file|
+%w(base errors generic sites entity sites/site sites/pages sites/pages/page sites/pages/page/extracted_content sites/entities sites/entities/entity).each do |file|
   require File.join(File.dirname(__FILE__), 'muddyit', file)
 end
 

data/muddyit_fu.gemspec

@@ -2,11 +2,11 @@
 
 Gem::Specification.new do |s|
   s.name = %q{muddyit_fu}
-  s.version = "0.1.1"
+  s.version = "0.2.0"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["robl"]
-  s.date = %q{2009-06-23}
+  s.date = %q{2009-08-03}
   s.email = %q{robl[at]monkeyhelper.com}
   s.extra_rdoc_files = [
     "LICENSE",
@@ -20,21 +20,23 @@ Gem::Specification.new do |s|
      "Rakefile",
      "VERSION",
      "lib/muddyit/base.rb",
-     "lib/muddyit/content_data.rb",
-     "lib/muddyit/entity.rb",
+     "lib/muddyit/entities.rb",
      "lib/muddyit/errors.rb",
      "lib/muddyit/generic.rb",
-     "lib/muddyit/page.rb",
-     "lib/muddyit/pages.rb",
-     "lib/muddyit/site.rb",
      "lib/muddyit/sites.rb",
+     "lib/muddyit/sites/entities.rb",
+     "lib/muddyit/sites/entities/entity.rb",
+     "lib/muddyit/sites/pages.rb",
+     "lib/muddyit/sites/pages/page.rb",
+     "lib/muddyit/sites/pages/page/extracted_content.rb",
+     "lib/muddyit/sites/site.rb",
      "lib/muddyit_fu.rb",
      "muddyit_fu.gemspec"
   ]
   s.homepage = %q{http://github.com/monkeyhelper/muddyit_fu}
   s.rdoc_options = ["--charset=UTF-8"]
   s.require_paths = ["lib"]
-  s.rubygems_version = %q{1.3.4}
+  s.rubygems_version = %q{1.3.5}
   s.summary = %q{Provides a ruby interface to muddy.it}
 
   if s.respond_to? :specification_version then
@@ -43,13 +45,10 @@ Gem::Specification.new do |s|
 
     if Gem::Version.new(Gem::RubyGemsVersion) >= Gem::Version.new('1.2.0') then
       s.add_runtime_dependency(%q<json>, [">= 0.0.0"])
-      s.add_runtime_dependency(%q<oauth>, [">= 0.0.0"])
     else
       s.add_dependency(%q<json>, [">= 0.0.0"])
-      s.add_dependency(%q<oauth>, [">= 0.0.0"])
     end
   else
     s.add_dependency(%q<json>, [">= 0.0.0"])
-    s.add_dependency(%q<oauth>, [">= 0.0.0"])
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: monkeyhelper-muddyit_fu
 version: !ruby/object:Gem::Version 
-  version: 0.1.1
+  version: 0.2.0
 platform: ruby
 authors: 
 - robl
@@ -9,7 +9,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2009-06-23 00:00:00 -07:00
+date: 2009-08-03 00:00:00 -07:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -22,16 +22,6 @@ dependencies:
       - !ruby/object:Gem::Version 
         version: 0.0.0
     version: 
-- !ruby/object:Gem::Dependency 
-  name: oauth
-  type: :runtime
-  version_requirement: 
-  version_requirements: !ruby/object:Gem::Requirement 
-    requirements: 
-    - - ">="
-      - !ruby/object:Gem::Version 
-        version: 0.0.0
-    version: 
 description: 
 email: robl[at]monkeyhelper.com
 executables: []
@@ -49,18 +39,21 @@ files:
 - Rakefile
 - VERSION
 - lib/muddyit/base.rb
-- lib/muddyit/content_data.rb
-- lib/muddyit/entity.rb
+- lib/muddyit/entities.rb
 - lib/muddyit/errors.rb
 - lib/muddyit/generic.rb
-- lib/muddyit/page.rb
-- lib/muddyit/pages.rb
-- lib/muddyit/site.rb
 - lib/muddyit/sites.rb
+- lib/muddyit/sites/entities.rb
+- lib/muddyit/sites/entities/entity.rb
+- lib/muddyit/sites/pages.rb
+- lib/muddyit/sites/pages/page.rb
+- lib/muddyit/sites/pages/page/extracted_content.rb
+- lib/muddyit/sites/site.rb
 - lib/muddyit_fu.rb
 - muddyit_fu.gemspec
 has_rdoc: false
 homepage: http://github.com/monkeyhelper/muddyit_fu
+licenses: 
 post_install_message: 
 rdoc_options: 
 - --charset=UTF-8
@@ -81,7 +74,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
 requirements: []
 
 rubyforge_project: 
-rubygems_version: 1.2.0
+rubygems_version: 1.3.5
 signing_key: 
 specification_version: 3
 summary: Provides a ruby interface to muddy.it

data/lib/muddyit/entity.rb

@@ -1,18 +0,0 @@
-class Muddyit::Entity < Muddyit::Generic
-
-  def classification
-    unless @attributes[:type]
-      # We merge here as we don't want to overwrite a entity specific confidence score
-      @attributes.merge!(self.fetch)
-    end
-    @attributes[:type]
-  end
-
-  protected
-  def fetch
-    api_url = "/entities/#{URI.escape(CGI.escape(@attributes[:uri]),'.')}"
-    response = @muddyit.send_request(api_url, :get)
-    response.nested_symbolize_keys!
-  end
-  
-end