RubyGems - mida - Versions diffs - 0.2.0 → 0.3.0 - Mend

mida 0.2.0 → 0.3.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (35) hide show

data/CHANGELOG.rdoc +18 -2
data/README.rdoc +22 -11
data/Rakefile +2 -2
data/lib/mida.rb +1 -1
data/lib/mida/datatype.rb +15 -0
data/lib/mida/datatype/boolean.rb +18 -0
data/lib/mida/datatype/float.rb +15 -0
data/lib/mida/datatype/integer.rb +15 -0
data/lib/mida/datatype/iso8601date.rb +17 -0
data/lib/mida/datatype/number.rb +15 -0
data/lib/mida/datatype/text.rb +13 -0
data/lib/mida/document.rb +21 -19
data/lib/mida/genericvocabulary.rb +13 -0
data/lib/mida/item.rb +83 -71
data/lib/mida/itemprop.rb +55 -30
data/lib/mida/itemscope.rb +82 -0
data/lib/mida/propertydesc.rb +36 -0
data/lib/mida/vocabulary.rb +60 -6
data/spec/datatype/boolean_spec.rb +27 -0
data/spec/datatype/float_spec.rb +23 -0
data/spec/datatype/integer_spec.rb +23 -0
data/spec/datatype/iso8601date_spec.rb +20 -0
data/spec/datatype/number_spec.rb +23 -0
data/spec/datatype/text_spec.rb +14 -0
data/spec/document_spec.rb +31 -487
data/spec/item_spec.rb +163 -472
data/spec/itemprop_spec.rb +40 -45
data/spec/itemscope_spec.rb +287 -0
data/spec/propertydesc_spec.rb +56 -0
data/spec/spec_helper.rb +13 -36
data/spec/vocabulary_spec.rb +148 -0
metadata +22 -6
data/lib/mida/vocabulary/generic.rb +0 -15
data/lib/mida/vocabularydesc.rb +0 -57
data/spec/vocabularydesc_spec.rb +0 -106

data/spec/datatype/text_spec.rb ADDED Viewed

@@ -0,0 +1,14 @@
+require 'mida/datatype'
+describe Mida::DataType::Text do
+  it '#extract should return an empty string if an empty string passed' do
+    value = ''
+    Mida::DataType::Text.extract(value).should == value
+  end
+  it '#extract should return the input value' do
+    value = 'A Test Value'
+    Mida::DataType::Text.extract(value).should == value
+  end
+end

data/spec/document_spec.rb CHANGED Viewed

@@ -1,5 +1,5 @@
-require_relative 'spec_helper'
-require_relative '../lib/mida'
+require 'spec_helper'
+require 'mida'
 def test_parsing(md, vocabulary, expected_results)
   items = md.search(vocabulary)
@@ -40,292 +40,45 @@ shared_examples_for 'one root itemscope' do
   end
 end
-describe Mida::Document, 'when run with a document containing textContent and non textContent itemprops' do
-  before do
-    @html = '
-      <html>
-        <head itemscope>
-          <link itemprop="link_field" rel="stylesheet" type="text/css" href="stylesheet.css" />
-        </head>
-        <body>
-          There is some text here
-          <div>
-            and also some here
-            <div itemscope>
-              <span itemprop="span_field">Some span content</span>
-              <time itemprop="dtreviewed" datetime="2009-01-06">Jan 6</time>.
-              <meta itemprop="meta_field" content="Some meta content">
-              <a itemprop="a_field1" href="http://example.com">non content</a>
-              <a itemprop="a_field2" href="welcome/index.html">non content</a>
-              <a itemprop="a_field3" href="/intro">non content</a>
-              <a itemprop="a_field4" href="/intro/index.html">non content</a>
-              <map name="somemap">
-                <area shape="rect" coords="0,0,50,120" href="left.html" />
-                <area itemprop="area_right" shape="rect" coords="51,0,120,120" href="right.html" />
-              </map>
-              <audio itemprop="audio_field" src="asound.ogg" controls="controls">
-                Audio tag not supported by your browser.
-              </audio>
-              <embed itemprop="embed_field" src="helloworld.swf" />
-              <iframe itemprop="iframe_field" src="http://www.example.com/iframe_test"></iframe>
-              <img itemprop="img_field" src="animage.png" width="120" height="120" usemap="#planetmap" />
-              <object itemprop="object_field" data="object.png" type="image/png" />
-              <audio controls="controls">
-                <source itemprop="source_field" src="song.ogg" type="audio/ogg" />
-                <track itemprop="track_field" src="atrack.ogg" />
-                Audio tag not supported by your browser.
-              </audio>
-              <video itemprop="video_field" src="movie.ogg" controls="controls">
-                Video tag not supported by your browser.
-              </video>
-            </div>
-          </div>
-        </body>
-      </html>
-    '
-  end
-  context 'when not given a page_url' do
-    before do
-      @md = Mida::Document.new(@html)
-    end
-    it 'should return all the properties and types with the correct values' do
-      expected_results = [
-        { vocabulary: Mida::Vocabulary::Generic,
-          type: nil, id: nil, properties: {'link_field' => ['']} },
-        { vocabulary: Mida::Vocabulary::Generic,
-          type: nil,
-          id: nil,
-          properties: {
-            'span_field' => ['Some span content'],
-            'dtreviewed' => ['2009-01-06'],
-            'meta_field' => ['Some meta content'],
-            'a_field1' => ['http://example.com'],
-            'a_field2' => [''],
-            'a_field3' => [''],
-            'a_field4' => [''],
-            'area_right' => [''],
-            'audio_field' => [''],
-            'embed_field' => [''],
-            'iframe_field' => ['http://www.example.com/iframe_test'],
-            'img_field' => [''],
-            'object_field' => [''],
-            'source_field' => [''],
-            'track_field' => [''],
-            'video_field' => ['']
-          }
-        }
-      ]
-      test_parsing(@md, %r{}, expected_results)
-    end
-  end
-  context 'when given a page_url' do
-    before do
-      @md = Mida::Document.new(@html, 'http://example.com/start/')
-    end
-    it 'should return all the properties and types with the correct values' do
-      expected_results = [
-        { vocabulary: Mida::Vocabulary::Generic,
-          type: nil, id: nil, properties: {
-          'link_field' => ['http://example.com/start/stylesheet.css']
-          }
-        },
-        { vocabulary: Mida::Vocabulary::Generic,
-          type: nil,
-          id: nil,
-          properties: {
-            'span_field' => ['Some span content'],
-            'dtreviewed' => ['2009-01-06'],
-            'meta_field' => ['Some meta content'],
-            'a_field1' => ['http://example.com'],
-            'a_field2' => ['http://example.com/start/welcome/index.html'],
-            'a_field3' => ['http://example.com/intro'],
-            'a_field4' => ['http://example.com/intro/index.html'],
-            'area_right' => ['http://example.com/start/right.html'],
-            'audio_field' => ['http://example.com/start/asound.ogg'],
-            'embed_field' => ['http://example.com/start/helloworld.swf'],
-            'iframe_field' => ['http://www.example.com/iframe_test'],
-            'img_field' => ['http://example.com/start/animage.png'],
-            'object_field' => ['http://example.com/start/object.png'],
-            'source_field' => ['http://example.com/start/song.ogg'],
-            'track_field' => ['http://example.com/start/atrack.ogg'],
-            'video_field' => ['http://example.com/start/movie.ogg']
-          }
-        }
-      ]
-      test_parsing(@md, %r{}, expected_results)
-    end
-  end
-end
-describe Mida::Document, 'when run against a full html document containing one itemscope with no itemtype' do
+describe Mida::Document do
   before do
     html = '
       <html><body>
         There is some text here
         <div>
           and also some here
-          <div itemscope>
+          <div itemscope itemtype="http://data-vocabulary.org/Review">
             <span itemprop="itemreviewed">Romeo Pizza</span>
-            Reviewed by <span itemprop="reviewer">Ulysses Grant</span> on
-            <time itemprop="dtreviewed" datetime="2009-01-06">Jan 6</time>.
-            <meta itemprop="fielda" content="a5482">
-            <span itemprop="summary">Delicious, tasty pizza in Eastlake!</span>
-            <span itemprop="description">This is a very nice pizza place.</span>
             Rating: <span itemprop="rating">4.5</span>
           </div>
-        </div>
-      </body></html>
-    '
-    @md = Mida::Document.new(html)
-  end
-  it_should_behave_like 'one root itemscope'
-  it 'should return all the properties and types with the correct values' do
-    expected_results = [{
-      vocabulary: Mida::Vocabulary::Generic,
-      type: nil,
-      id: nil,
-      properties: {
-        'itemreviewed' => ['Romeo Pizza'],
-        'reviewer' => ['Ulysses Grant'],
-        'dtreviewed' => ['2009-01-06'],
-        'fielda' => ['a5482'],
-        'summary' => ['Delicious, tasty pizza in Eastlake!'],
-        'description' => ['This is a very nice pizza place.'],
-        'rating' => ['4.5']
-      }
-    }]
-    test_parsing(@md, %r{}, expected_results)
-  end
-end
-describe Mida::Document, 'when run against a full html document containing one itemscope nested within another' do
-  before do
-    html = '
-      <html><body>
-        There is some text here
-        <div>
-          and also some here
-          <div itemscope>
-            <span itemprop="itemreviewed">Romeo Pizza</span>
-            <div itemprop="address" itemscope>
-              <span itemprop="firstline">237 Italian Way</span>
-              <span itemprop="country">United Kingdom</span>
-            </div>
-            Rating: <span itemprop="rating">4.5</span>
+          <div itemscope itemtype="http://data-vocabulary.org/Organization">
+            <span itemprop="name">An org name</span>
+            <span itemprop="url">http://example.com</span>
           </div>
         </div>
       </body></html>
     '
     @md = Mida::Document.new(html)
   end
-  it_should_behave_like 'one root itemscope'
-  it 'should return all the properties and types with the correct values' do
-    expected_results = [{
-      vocabulary: Mida::Vocabulary::Generic,
-      type: nil,
-      id: nil,
-      properties: {
-        'itemreviewed' => ['Romeo Pizza'],
-        'address' => [{
-          vocabulary: Mida::Vocabulary::Generic,
-          type: nil, id: nil, properties: {
-            'firstline' => ['237 Italian Way'],
-            'country' => ['United Kingdom']
-           }
-        }],
-        'rating' => ['4.5']
-      }
-    }]
-    test_parsing(@md, %r{}, expected_results)
-  end
-end
-describe Mida::Document, 'when run against a full html document containing one itemscope nested within another within another' do
-  before do
-    html = '
-      <html><body>
-        There is some text here
-        <div>
-          and also some here
-          <div itemscope>
-            <span itemprop="itemreviewed">Romeo Pizza</span>
-            <div itemprop="address" itemscope>
-              <div itemprop="firstline" itemscope>
-                <span itemprop="number">237</span>
-                <span itemprop="road">Italian Way</span>
-              </div>
-              <span itemprop="country">United Kingdom</span>
-            </div>
-            Rating: <span itemprop="rating">4.5</span>
-          </div>
-        </div>
-      </body></html>
-    '
-    @md = Mida::Document.new(html)
+  it '#each should pass each item to the block' do
+    item_num = 0
+    @md.each {|item| item.should == @md.items[item_num]; item_num += 1}
   end
-  it_should_behave_like 'one root itemscope'
+  it 'should have access to the Enumerable mixin methods such as #find' do
+    review = @md.find {|item| item.type == 'http://data-vocabulary.org/Review'}
+    review.type.should == 'http://data-vocabulary.org/Review'
+    review.properties['itemreviewed'].should == ["Romeo Pizza"]
-  it 'should return all the properties and types with the correct values' do
-    expected_results = [{
-      vocabulary: Mida::Vocabulary::Generic,
-      type: nil,
-      id: nil,
-      properties: {
-        'itemreviewed' => ['Romeo Pizza'],
-        'address' => [{
-          vocabulary: Mida::Vocabulary::Generic,
-          type: nil,
-          id: nil,
-          properties: {
-            'firstline' => [{
-              vocabulary: Mida::Vocabulary::Generic,
-              type: nil,
-              id: nil,
-              properties: {
-                'number' => ['237'],
-                'road' => ['Italian Way']
-              },
-            }],
-            'country' => ['United Kingdom']
-          },
-        }],
-        'rating' => ['4.5']
-      }
-    }]
-    test_parsing(@md, %r{^$}, expected_results)
+    organization = @md.find {|item| item.type == 'http://data-vocabulary.org/Organization'}
+    organization.type.should == 'http://data-vocabulary.org/Organization'
+    organization.properties['name'].should == ["An org name"]
   end
 end
-describe Mida::Document, 'when run against a full html document containing one itemscope with an itemtype' do
+describe Mida::Document, 'when run against a full html document containing itemscopes with and without itemtypes' do
   before do
     html = '
@@ -335,12 +88,12 @@ describe Mida::Document, 'when run against a full html document containing one i
           and also some here
           <div itemscope itemtype="http://data-vocabulary.org/Review">
             <span itemprop="itemreviewed">Romeo Pizza</span>
-            Reviewed by <span itemprop="reviewer">Ulysses Grant</span> on
-            <time itemprop="dtreviewed" datetime="2009-01-06">Jan 6</time>.
-            <span itemprop="summary">Delicious, tasty pizza in Eastlake!</span>
-            <span itemprop="description">This is a very nice pizza place.</span>
             Rating: <span itemprop="rating">4.5</span>
           </div>
+          <div itemscope>
+            <span itemprop="name">An org name</span>
+            <span itemprop="url">http://example.com</span>
+          </div>
         </div>
       </body></html>
     '
@@ -349,33 +102,17 @@ describe Mida::Document, 'when run against a full html document containing one i
   end
-  it_should_behave_like 'one root itemscope'
-  it 'should find the correct number of itemscopes if outer specified' do
-    @md.search(%r{http://data-vocabulary.org/Review}).size.should == 1
-  end
-  it 'should specify the correct type' do
-    @md.search(%r{http://data-vocabulary.org/Review}).first.type.should == 'http://data-vocabulary.org/Review'
+  it '#search should be able to match against items without an itemtype' do
+    items = @md.search(%r{^$})
+    items.size.should == 1
+    items[0].properties['name'].should == ['An org name']
   end
-  it 'should return all the properties and types with the correct values' do
-    expected_results = [{
-      vocabulary: Mida::Vocabulary::Generic,
-      type: 'http://data-vocabulary.org/Review',
-      id: nil,
-      properties: {
-        'itemreviewed' => ['Romeo Pizza'],
-        'reviewer' => ['Ulysses Grant'],
-        'dtreviewed' => ['2009-01-06'],
-        'summary' => ['Delicious, tasty pizza in Eastlake!'],
-        'description'  => ['This is a very nice pizza place.'],
-        'rating' => ['4.5']
-      }
-    }]
-    test_parsing(@md, %r{http://data-vocabulary.org/Review}, expected_results)
+  it '#search should be able to match against items with an itemtype' do
+    items = @md.search(%r{^.+$})
+    items.size.should == 1
+    items[0].type.should == 'http://data-vocabulary.org/Review'
   end
 end
 describe Mida::Document, 'when run against a full html document containing two non-nested itemscopes with itemtypes' do
@@ -423,7 +160,6 @@ describe Mida::Document, 'when run against a full html document containing two n
   it 'should return all the properties and types with the correct values for 1st itemscope' do
     expected_results = [{
-      vocabulary: Mida::Vocabulary::Generic,
       type: 'http://data-vocabulary.org/Review',
       id: nil,
       properties: {
@@ -436,7 +172,6 @@ describe Mida::Document, 'when run against a full html document containing two n
   it 'should return all the properties from the text for 2nd itemscope' do
     expected_results = [{
-      vocabulary: Mida::Vocabulary::Generic,
       type: 'http://data-vocabulary.org/Organization',
       id: nil,
       properties: {
@@ -482,19 +217,18 @@ describe Mida::Document, 'when run against a full html document containing one
       %r{http://data-vocabulary.org/Review-aggregate}
     ]
     vocabularies.each {|vocabulary| @md.search(vocabulary).size.should == 1}
   end
   context "when looking at the outer vocabulary" do
     it 'should return all the properties from the text with the correct values' do
       expected_results = [{
-        vocabulary: Mida::Vocabulary::Generic,
         type: 'http://data-vocabulary.org/Product',
         id: nil,
         properties: {
           'name' => ['DC07'],
           'brand' => ['Dyson'],
           'review' => [{
-            vocabulary: Mida::Vocabulary::Generic,
             type: 'http://data-vocabulary.org/Review-aggregate',
             id: nil,
             properties: {
@@ -572,193 +306,3 @@ describe Mida::Document, 'when run against a document containing an itemscope
     end
   end
 end
-describe Mida::Document, 'when run against a document using itemrefs' do
-  before do
-    html = '
-      <html><body>
-        <div itemscope id="amanda" itemref="a b">
-          <span itemprop="age">30</span>
-        </div>
-        <p id="a">Name: <span itemprop="name">Amanda</span></p>
-        <div id="b" itemprop="band" itemscope itemref="c"></div>
-        <div id="c">
-          <p>Band: <span itemprop="name">Jazz Band</span></p>
-          <p>Size: <span itemprop="size">12</span> players</p>
-        </div>
-      </body></html>
-    '
-    @md = Mida::Document.new(html)
-  end
-  it 'should return all the properties from the text with the correct values' do
-    expected_results = [{
-      vocabulary: Mida::Vocabulary::Generic,
-      type: nil,
-      id: nil,
-      properties: {
-        'name' => ['Amanda'],
-        'band' => [{
-          vocabulary: Mida::Vocabulary::Generic,
-          type: nil,
-          id: nil,
-          properties: {
-            'name' => ['Jazz Band'],
-            'size' => ['12']
-          }
-        }],
-        'age' => ['30']
-      }
-    }]
-    test_parsing(@md, %r{}, expected_results)
-  end
-end
-describe Mida::Document, 'when run against a document using multiple itemprops with the same name' do
-  before do
-    html = '
-      <html><body>
-        <div itemscope itemtype="icecreams">
-          <p>Flavours in my favourite ice cream:</p>
-          <ul>
-            <li itemprop="flavour">Lemon sorbet</li>
-            <li itemprop="flavour">Apricot sorbet</li>
-            <li itemprop="flavour" itemscope itemtype="icecream-type">
-              <span itemprop="fruit">Strawberry</span>
-              <span itemprop="style">Homemade</span>
-            </li>
-          </ul>
-        </div>
-      </body></html>
-    '
-    @md = Mida::Document.new(html)
-  end
-  it_should_behave_like 'one root itemscope'
-  it 'should return the correct number of itemscopes' do
-    vocabularies = [
-      %r{icecreams},
-      %r{icecream-type}
-    ]
-    vocabularies.each {|vocabulary| @md.search(vocabulary).size.should == 1}
-  end
-  it 'should return all the properties from the text with the correct values' do
-    expected_results = [{
-      vocabulary: Mida::Vocabulary::Generic,
-      type: 'icecreams',
-      id: nil,
-      properties: {
-        'flavour' => [
-          'Lemon sorbet',
-          'Apricot sorbet',
-          { vocabulary: Mida::Vocabulary::Generic,
-            type: 'icecream-type',
-            id: nil,
-            properties: {
-              'fruit' => ['Strawberry'],
-              'style' => ['Homemade']
-            }
-          }
-        ]
-      }
-    }]
-    test_parsing(@md, %r{icecreams}, expected_results)
-  end
-end
-describe Mida::Document, 'when run against a document using an itemprop with multiple properties' do
-  before do
-    html = '
-      <html><body>
-        <div itemscope>
-          <span itemprop="favourite-colour favourite-fruit">orange</span>
-        </div>
-      </body></html>
-    '
-    @md = Mida::Document.new(html)
-  end
-  it 'should return all the properties from the text with the correct values' do
-    expected_results = [{
-      vocabulary: Mida::Vocabulary::Generic,
-      type: nil,
-      id: nil,
-      properties: {
-        'favourite-colour' => ['orange'],
-        'favourite-fruit' => ['orange']
-      }
-    }]
-    test_parsing(@md, %r{}, expected_results)
-  end
-end
-describe Mida::Document, 'when run against a full html document containing an itemtype that matches a registered vocabulary' do
-  before do
-    html = '
-      <html><body>
-        There is some text here
-        <div>
-          and also some here
-          <div itemscope itemtype="http://data-vocabulary.org/Review">
-            <span itemprop="itemreviewed">Romeo Pizza</span>
-            Reviewed by <span itemprop="reviewer">Ulysses Grant</span> on
-            <time itemprop="dtreviewed" datetime="2009-01-06">Jan 6</time>.
-            <span itemprop="summary">Delicious, tasty pizza in Eastlake!</span>
-            <span itemprop="description">This is a very nice pizza place.</span>
-            Rating: <span itemprop="rating">4.5</span>
-          </div>
-        </div>
-      </body></html>
-    '
-    class Review < Mida::VocabularyDesc
-      itemtype %r{http://data-vocabulary.org/Review}
-      has_one 'itemreviewed', 'reviewer', 'dtreviewed', 'summary'
-      has_one 'rating', 'description'
-    end
-    Mida::Vocabulary.register(Review)
-    @md = Mida::Document.new(html)
-  end
-  it_should_behave_like 'one root itemscope'
-  it '#search should match against Review' do
-    @md.search(Review).size.should == 1
-  end
-  it 'should specify the correct type' do
-    @md.search(Review).first.type.should == 'http://data-vocabulary.org/Review'
-  end
-  it 'should return all the properties and types with the correct values' do
-    expected_results = [{
-      vocabulary: Review,
-      type: 'http://data-vocabulary.org/Review',
-      id: nil,
-      properties: {
-        'itemreviewed' => ['Romeo Pizza'],
-        'reviewer' => ['Ulysses Grant'],
-        'dtreviewed' => ['2009-01-06'],
-        'summary' => ['Delicious, tasty pizza in Eastlake!'],
-        'description'  => ['This is a very nice pizza place.'],
-        'rating' => ['4.5']
-      }
-    }]
-    test_parsing(@md, Review, expected_results)
-  end
-end