alexrabarts-term_extraction 0.1.0

data/VERSION.yml

@@ -0,0 +1,4 @@
+--- 
+:minor: 1
+:patch: 0
+:major: 0

data/lib/term_extraction.rb

@@ -0,0 +1,12 @@
+class TermExtraction
+  def initialize(options={})
+    @context = options[:context]
+    @api_key = options[:api_key]
+  end
+
+  def canonical_name
+    self.class.canonical_name
+  end
+end
+
+%w{yahoo zemanta}.each{|t| require "term_extraction/#{t}"}

data/lib/term_extraction/yahoo.rb

@@ -0,0 +1,51 @@
+require 'nokogiri'
+
+class TermExtraction
+  class Yahoo < TermExtraction
+    def terms
+      terms   = []
+
+      unless data.nil?
+        data.search('//s:Result', ns).each do |n|
+          terms << n.text
+        end
+      end
+
+      terms
+    end 
+
+    class << self
+      def canonical_name
+        'yahoo'
+      end
+    end
+
+    private
+      def ns
+        {'s' => 'urn:yahoo:cate'}
+      end
+
+      def gateway
+        'http://search.yahooapis.com/ContentAnalysisService/V1/termExtraction'
+      end
+
+      def url
+        uri = Addressable::URI.parse(gateway)
+        uri.query_values = {
+           # TODO: Change appid to the BMP one
+           'appid'   => @api_key,
+           'output'  => 'xml',
+           'context' => @context
+        }
+        uri
+      end
+
+      def remote_xml
+        open(url).read
+      end
+
+      def data
+        @data ||= Nokogiri::XML.parse(remote_xml)
+      end
+  end
+end

data/lib/term_extraction/zemanta.rb

@@ -0,0 +1,49 @@
+require 'net/http'
+require 'nokogiri'
+
+class TermExtraction
+  class Zemanta < TermExtraction
+    def terms
+      terms = []
+
+      data.search('//name').each do |n|
+        terms << n.text
+      end
+
+      terms
+    end  
+
+    class << self
+      def canonical_name
+        'zemanta'
+      end
+    end
+
+    private
+      def gateway
+        'http://api.zemanta.com/services/rest/0.0/'
+      end
+
+      def url
+        URI.parse(gateway)
+      end
+
+      def post_params
+        {
+          'method'        =>'zemanta.suggest',
+          'api_key'       => @api_key,
+          'return_images' => 0,
+          'text'          => @context,
+          'format'        => 'xml'
+        }
+      end
+
+      def remote_xml
+        Net::HTTP.post_form(url, post_params).body
+      end
+
+      def data
+        @data ||= Nokogiri::XML.parse(remote_xml)
+      end
+  end
+end

data/test/fixtures/yahoo.xml

@@ -0,0 +1,3 @@
+<?xml version="1.0"?>
+<ResultSet xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="urn:yahoo:cate" xsi:schemaLocation="urn:yahoo:cate http://api.search.yahoo.com/ContentAnalysisService/V1/TermExtractionResponse.xsd"><Result>gears of war</Result><Result>gears</Result></ResultSet>
+<!-- ws04.search.scd.yahoo.com uncompressed/chunked Thu Feb 26 21:04:16 PST 2009 -->

data/test/fixtures/zemanta.xml

@@ -0,0 +1,104 @@
+<?xml version="1.0"?>
+<rsp>
+	<status>ok</status>
+	<articles>
+		<article>
+			<url>http://www.crunchgear.com/2009/02/24/nvidia-based-imacs-coming-soon/</url>
+			<confidence>0.033153</confidence>
+			<published_datetime>2009-02-24T18:00:45Z</published_datetime>
+			<zemified>0</zemified>
+			<title>NVIDIA-based iMacs coming soon?</title>
+		</article><article>
+			<url>http://www.tuaw.com/2009/02/24/rumor-new-imacs-around-the-bend/</url>
+			<confidence>0.028595</confidence>
+			<published_datetime>2009-02-24T22:00:00Z</published_datetime>
+			<zemified>0</zemified>
+			<title>Rumor: New iMacs around the bend</title>
+		</article><article>
+			<url>http://www.engadget.com/2009/02/24/a-few-new-rumors-point-to-two-new-nvidia-packing-imacs/</url>
+			<confidence>0.027595</confidence>
+			<published_datetime>2009-02-24T13:03:00Z</published_datetime>
+			<zemified>0</zemified>
+			<title>A few new rumors point to two new NVIDIA-packing iMacs</title>
+		</article><article>
+			<url>http://www.techmeme.com/090126/p31</url>
+			<confidence>0.019056</confidence>
+			<published_datetime>2009-01-26T17:00:22Z</published_datetime>
+			<zemified>0</zemified>
+			<title>Chip complex delaying Apple's new iMac line, says analyst (Zach Spear/AppleInsider)</title>
+		</article><article>
+			<url>http://i.gizmodo.com/5143476/apple-warns-resellers-of-reduced-imac-availability-new-models-on-the-way</url>
+			<confidence>0.018611</confidence>
+			<published_datetime>2009-01-31T16:30:00Z</published_datetime>
+			<zemified>0</zemified>
+			<title>Apple Warns Resellers of Reduced iMac Availability: New Models On the Way? [Apple]</title>
+		</article><article>
+			<url>http://cultofmac.com/analyst-new-imacs-delayed-for-chips-snow-leopard/7549</url>
+			<confidence>0.01685</confidence>
+			<published_datetime>2009-01-26T16:43:44Z</published_datetime>
+			<zemified>0</zemified>
+			<title>Analyst: New iMacs Delayed For Chips, Snow Leopard</title>
+		</article><article>
+			<url>http://www.ubergizmo.com/15/archives/2008/12/new_imac_allinone_pc_confirmed.html</url>
+			<confidence>0.016311</confidence>
+			<published_datetime>2008-12-24T01:06:56Z</published_datetime>
+			<zemified>0</zemified>
+			<title>New iMac All-In-One PC Confirmed</title>
+		</article><article>
+			<url>http://www.labnol.org/gadgets/exchange-windows-pc-for-apple-mac/5775/</url>
+			<confidence>0.016238</confidence>
+			<published_datetime>2008-12-03T08:17:35Z</published_datetime>
+			<zemified>0</zemified>
+			<title>Exchange Your Windows PC or TV for an Apple iMac</title>
+		</article><article>
+			<url>http://cultofmac.com/class-action-lawsuit-over-imac-display-problems/6562</url>
+			<confidence>0.015608</confidence>
+			<published_datetime>2009-01-02T16:17:06Z</published_datetime>
+			<zemified>0</zemified>
+			<title>Class-Action Lawsuit Over iMac Display Problems</title>
+		</article><article>
+			<url>http://www.crunchgear.com/2008/12/29/new-imacs-to-include-hotter-components-new-cooling/</url>
+			<confidence>0.015093</confidence>
+			<published_datetime>2008-12-29T23:50:04Z</published_datetime>
+			<zemified>0</zemified>
+			<title>New iMacs to include hotter components, new cooling?</title>
+		</article>
+	</articles><markup>
+		<text>apple imac</text>
+	</markup><signature>&lt;div class="zemanta-pixie"&gt;&lt;a class="zemanta-pixie-a" href="http://reblog.zemanta.com/zemified/3695b55a-d68e-4328-bf86-c3f01be8b2da/" title="Zemified by Zemanta"&gt;&lt;img class="zemanta-pixie-img" src="http://img.zemanta.com/reblog_e.png?x-id=3695b55a-d68e-4328-bf86-c3f01be8b2da" alt="Reblog this post [with Zemanta]" /&gt;&lt;/a&gt;&lt;/div&gt;</signature>
+	<keywords>
+		<keyword>
+			<confidence>0.404748</confidence>
+			<scheme>general</scheme>
+			<name>Apple</name>
+		</keyword><keyword>
+			<confidence>0.277753</confidence>
+			<scheme>general</scheme>
+			<name>IMac</name>
+		</keyword><keyword>
+			<confidence>0.094392</confidence>
+			<scheme>general</scheme>
+			<name>Rumor</name>
+		</keyword><keyword>
+			<confidence>0.076399</confidence>
+			<scheme>general</scheme>
+			<name>Hardware</name>
+		</keyword><keyword>
+			<confidence>0.072186</confidence>
+			<scheme>general</scheme>
+			<name>Nvidia</name>
+		</keyword><keyword>
+			<confidence>0.069183</confidence>
+			<scheme>general</scheme>
+			<name>Macintosh</name>
+		</keyword><keyword>
+			<confidence>0.050487</confidence>
+			<scheme>general</scheme>
+			<name>Desktops</name>
+		</keyword><keyword>
+			<confidence>0.039205</confidence>
+			<scheme>general</scheme>
+			<name>AllInOne</name>
+		</keyword>
+	</keywords><rid>3695b55a-d68e-4328-bf86-c3f01be8b2da</rid>
+</rsp>

data/test/term_extraction_test.rb

@@ -0,0 +1,15 @@
+require File.dirname(__FILE__) + '/test_helper'
+
+class TermExtractionTest < Test::Unit::TestCase
+  should 'return correct terms from Yahoo!' do
+    yahoo = TermExtraction::Yahoo.new
+    yahoo.stubs(:remote_xml).returns(read_xml_fixture('yahoo'))
+    assert_equal yahoo.terms, ['gears of war', 'gears']
+  end
+  
+  should 'return correct terms from Zemanta' do
+    zemanta = TermExtraction::Zemanta.new
+    zemanta.stubs(:remote_xml).returns(read_xml_fixture('zemanta'))
+    assert_equal zemanta.terms, ['Apple', 'IMac', 'Rumor', 'Hardware', 'Nvidia', 'Macintosh', 'Desktops', 'AllInOne']
+  end
+end

data/test/test_helper.rb

@@ -0,0 +1,13 @@
+require 'rubygems'
+require 'test/unit'
+require 'shoulda'
+require 'mocha'
+
+$LOAD_PATH.unshift(File.dirname(__FILE__))
+require 'term_extraction'
+
+class Test::Unit::TestCase
+  def read_xml_fixture(name)
+    File.read("test/fixtures/#{name}.xml")
+  end
+end

metadata

@@ -0,0 +1,72 @@
+--- !ruby/object:Gem::Specification 
+name: alexrabarts-term_extraction
+version: !ruby/object:Gem::Version 
+  version: 0.1.0
+platform: ruby
+authors: 
+- alex
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2009-02-27 00:00:00 -08:00
+default_executable: 
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: nokogiri
+  type: :runtime
+  version_requirement: 
+  version_requirements: !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: 1.0.7
+    version: 
+description: Term extraction library
+email: alexrabarts@gmail.com
+executables: []
+
+extensions: []
+
+extra_rdoc_files: []
+
+files: 
+- VERSION.yml
+- lib/term_extraction
+- lib/term_extraction/yahoo.rb
+- lib/term_extraction/zemanta.rb
+- lib/term_extraction.rb
+- test/fixtures
+- test/fixtures/yahoo.xml
+- test/fixtures/zemanta.xml
+- test/term_extraction_test.rb
+- test/test_helper.rb
+has_rdoc: true
+homepage: http://github.com/alexrabarts/term_extraction
+post_install_message: 
+rdoc_options: 
+- --inline-source
+- --charset=UTF-8
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+requirements: []
+
+rubyforge_project: 
+rubygems_version: 1.2.0
+signing_key: 
+specification_version: 2
+summary: Provides access to term extraction APIs such as Yahoo! Term Extraction API and Zemanta.
+test_files: []
+