semantic_hacker 0.0.1

data/test/fixtures/concept.json

@@ -0,0 +1,27 @@
+{
+    "about":     {
+        "requestId": "EEDC78572D5F839FD115E0250FA00ADE",
+        "docId": "2627917C26F7F284C8364209B0B21354",
+        "systemType": "concept",
+        "configId": "odp_2007_l1_1.7k",
+        "contentType": "text/html",
+        "contentDigest": "C7DBE9045BF3F905EA8E7273EC1A4540",
+        "requestDate": "2010-05-27T19:19:46+00:00",
+        "systemVersion": "2.0",
+        "sourceUri": "http://washingtondc.craigslist.org/nva/apa/1760484413.html"
+    },
+    "conceptExtractor": {"conceptExtractorResponse": {"concepts":     [
+                {
+            "weight": "0.042842045",
+            "label": "camden"
+        },
+                {
+            "weight": "0.002559835",
+            "label": "lake"
+        },
+                {
+            "weight": "0.0018040233",
+            "label": "fair"
+        }
+    ]}}
+}

data/test/fixtures/signature.json

@@ -0,0 +1,135 @@
+{
+    "about":     {
+        "requestId": "96696110A319CDAE224908941C7AA941",
+        "docId": "2627917C26F7F284C8364209B0B21354",
+        "systemType": "signature",
+        "configId": "odp_2007_l1_1.7k",
+        "contentType": "text/html",
+        "contentDigest": "C7DBE9045BF3F905EA8E7273EC1A4540",
+        "requestDate": "2010-05-27T19:19:26+00:00",
+        "systemVersion": "2.0",
+        "sourceUri": "http://washingtondc.craigslist.org/nva/apa/1760484413.html"
+    },
+    "siggen": {"siggenResponse": {"signature":     [
+                {
+            "weight": "0.28168735",
+            "index": "1096"
+        },
+                {
+            "weight": "0.26628992",
+            "index": "1184"
+        },
+                {
+            "weight": "0.2640163",
+            "index": "1572"
+        },
+                {
+            "weight": "0.26159292",
+            "index": "1098"
+        },
+                {
+            "weight": "0.21498927",
+            "index": "1177"
+        },
+                {
+            "weight": "0.21138062",
+            "index": "1130"
+        },
+                {
+            "weight": "0.21044247",
+            "index": "1182"
+        },
+                {
+            "weight": "0.19739439",
+            "index": "1129"
+        },
+                {
+            "weight": "0.19653957",
+            "index": "1135"
+        },
+                {
+            "weight": "0.18718655",
+            "index": "1216"
+        },
+                {
+            "weight": "0.18387607",
+            "index": "254"
+        },
+                {
+            "weight": "0.1741769",
+            "index": "305"
+        },
+                {
+            "weight": "0.16953631",
+            "index": "658"
+        },
+                {
+            "weight": "0.16824634",
+            "index": "1491"
+        },
+                {
+            "weight": "0.15728694",
+            "index": "1045"
+        },
+                {
+            "weight": "0.15663064",
+            "index": "909"
+        },
+                {
+            "weight": "0.15609612",
+            "index": "854"
+        },
+                {
+            "weight": "0.15229575",
+            "index": "932"
+        },
+                {
+            "weight": "0.15147352",
+            "index": "1475"
+        },
+                {
+            "weight": "0.14810044",
+            "index": "1199"
+        },
+                {
+            "weight": "0.14544998",
+            "index": "1188"
+        },
+                {
+            "weight": "0.14537416",
+            "index": "1036"
+        },
+                {
+            "weight": "0.14495832",
+            "index": "1086"
+        },
+                {
+            "weight": "0.14386787",
+            "index": "1133"
+        },
+                {
+            "weight": "0.14327037",
+            "index": "1301"
+        },
+                {
+            "weight": "0.14155875",
+            "index": "1230"
+        },
+                {
+            "weight": "0.14054182",
+            "index": "890"
+        },
+                {
+            "weight": "0.13867372",
+            "index": "1197"
+        },
+                {
+            "weight": "0.13715525",
+            "index": "1069"
+        },
+                {
+            "weight": "0.13608803",
+            "index": "1062"
+        }
+    ]}}
+}

data/test/semantic_hacker_test.rb

@@ -0,0 +1,32 @@
+require 'test_helper'
+
+class SemanticHackerTest < Test::Unit::TestCase
+
+  should "have category method for category content" do
+    stub_get('/category?uri=http%3A%2F%2Fwashingtondc.craigslist.org%2Fnva%2Fapa%2F1760484413.html&format=json', 'category.json')
+    cat = SemanticHacker::Category.uri("http://washingtondc.craigslist.org/nva/apa/1760484413.html")
+    cat[:categories].size.should == 1
+    cat[:about].system_type.should == "category"
+    cat[:about].config_id.should == "odp_2007_l1_1.7k"
+    cat[:about].source_uri.should == "http://washingtondc.craigslist.org/nva/apa/1760484413.html"
+  end
+
+  should "have concept method for concept content" do
+    stub_get('/concept?uri=http%3A%2F%2Fwashingtondc.craigslist.org%2Fnva%2Fapa%2F1760484413.html&format=json', 'concept.json')
+    con = SemanticHacker::Concept.uri("http://washingtondc.craigslist.org/nva/apa/1760484413.html")
+    con[:concepts].size.should == 3
+    con[:about].system_type.should == "concept"
+    con[:about].config_id.should == "odp_2007_l1_1.7k"
+    con[:about].source_uri.should == "http://washingtondc.craigslist.org/nva/apa/1760484413.html"
+  end
+
+  should "have signature method for signature content" do
+    stub_get('/signature?uri=http%3A%2F%2Fwashingtondc.craigslist.org%2Fnva%2Fapa%2F1760484413.html&format=json', 'signature.json')
+    sig = SemanticHacker::Signature.uri("http://washingtondc.craigslist.org/nva/apa/1760484413.html")
+    sig[:signature].size.should == 30
+    sig[:about].system_type.should == "signature"
+    sig[:about].config_id.should == "odp_2007_l1_1.7k"
+    sig[:about].source_uri.should == "http://washingtondc.craigslist.org/nva/apa/1760484413.html"
+  end
+
+end

data/test/test_helper.rb

@@ -0,0 +1,38 @@
+require 'test/unit'
+require 'pathname'
+require 'rubygems'
+
+gem 'shoulda', '>= 2.10.1'
+gem 'jnunemaker-matchy', '0.4.0'
+gem 'mocha', '0.9.4'
+gem 'fakeweb', '>= 1.2.5'
+
+require 'shoulda'
+require 'matchy'
+require 'mocha'
+require 'fakeweb'
+
+FakeWeb.allow_net_connect = false
+
+dir = (Pathname(__FILE__).dirname + '../lib').expand_path
+
+require dir + 'semantic_hacker'
+
+class Test::Unit::TestCase
+end
+
+def fixture_file(filename)
+  return '' if filename == ''
+  file_path = File.expand_path(File.dirname(__FILE__) + '/fixtures/' + filename)
+  File.read(file_path)
+end
+
+def semantichacker_url(url)
+  url =~ /^http/ ? url : "http://api.semantichacker.com/55u2g9gs#{url}"
+end
+
+def stub_get(url, filename, status=nil)
+  options = {:body => fixture_file(filename)}
+  options.merge!({:status => status}) unless status.nil?
+  FakeWeb.register_uri(:get, semantichacker_url(url), options)
+end

data/version.txt

@@ -0,0 +1 @@
+0.0.1

metadata

@@ -0,0 +1,100 @@
+--- !ruby/object:Gem::Specification 
+name: semantic_hacker
+version: !ruby/object:Gem::Version 
+  prerelease: false
+  segments: 
+  - 0
+  - 0
+  - 1
+  version: 0.0.1
+platform: ruby
+authors: 
+- Thomas Gallaway
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2010-05-27 00:00:00 -04:00
+default_executable: semantic_hacker
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: bones
+  prerelease: false
+  requirement: &id001 !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        segments: 
+        - 3
+        - 4
+        - 1
+        version: 3.4.1
+  type: :development
+  version_requirements: *id001
+description: "Semantic Signatures\xC2\xAE  are a new way of representing and analyzing semantic information (meaning) in text. Semantic Signatures, produced by TextWise\xE2\x80\x99s Trainable Semantic Vectors (TSV) technology, provide a rich semantic representation of the multiple concepts and topics contained in a body of text. Semantic Signatures can be constructed for a wide range of texts including individual words, phrases, word lists (e.g. metadata), short passages (such as text advertisements or image labels), web pages, or full text documents (e.g. technical articles)."
+email: atomist@atomlab.us
+executables: 
+- semantic_hacker
+extensions: []
+
+extra_rdoc_files: 
+- History.txt
+- README.txt
+- bin/semantic_hacker
+- version.txt
+files: 
+- History.txt
+- README.txt
+- Rakefile
+- bin/semantic_hacker
+- data/categories.yaml
+- data/labels.yaml
+- lib/semantic_hacker.rb
+- lib/semantic_hacker/api.rb
+- lib/semantic_hacker/category.rb
+- lib/semantic_hacker/concept.rb
+- lib/semantic_hacker/setup.rb
+- lib/semantic_hacker/signature.rb
+- semantic_hacker.gemspec
+- spec/semantic_hacker_spec.rb
+- spec/spec_helper.rb
+- test/fixtures/1760484413.html
+- test/fixtures/category.json
+- test/fixtures/concept.json
+- test/fixtures/signature.json
+- test/semantic_hacker_test.rb
+- test/test_helper.rb
+- version.txt
+has_rdoc: true
+homepage: http://www.atomlab.us/semantic_hacker
+licenses: []
+
+post_install_message: 
+rdoc_options: 
+- --main
+- README.txt
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      segments: 
+      - 0
+      version: "0"
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      segments: 
+      - 0
+      version: "0"
+requirements: []
+
+rubyforge_project: semantic_hacker
+rubygems_version: 1.3.6
+signing_key: 
+specification_version: 3
+summary: "Semantic Signatures\xC2\xAE  are a new way of representing and analyzing semantic information (meaning) in text"
+test_files: 
+- test/test_helper.rb