philologic-client 0.0.10 → 0.0.11

data/HISTORY.rdoc

@@ -1,5 +1,9 @@
 = Philologic::Client History
 
+== 2012-02-09 v0.0.11
+
+* Added +Philologic::Link#text()+
+
 == 2012-01-18 v0.0.10
 
 * Added +Philologic::TableRow+

data/README.rdoc

@@ -68,7 +68,6 @@ https://github.com/blairc/philologic-client/
 
 == To Do
 
-* Add collocation table searching
 * Add +children+ to Philologic::Document
 * Add +parent+ to Philologic::Document
 * Unify (as much as possible) Philologic::Document, Philologic::Link, Philologic::TableRow

data/lib/philologic-client/version.rb

@@ -5,6 +5,6 @@ module Philologic
     #
     # Philologic::Client version
     #
-    VERSION = '0.0.10'
+    VERSION = '0.0.11'
   end
 end

data/lib/philologic-client.rb

@@ -76,7 +76,6 @@ require 'open-uri'
 #
 # == To Do
 #
-# * Add collocation table searching
 # * Add +children+ to Philologic::Document
 # * Add +parent+ to Philologic::Document
 # * Unify (as much as possible) Philologic::Document, Philologic::Link, Philologic::TableRow
@@ -200,7 +199,7 @@ module Philologic # :nodoc:
       # TODO Use +URI+
       raise('no endpoint specified') if @endpoint.nil?
       uri = "#{ @endpoint }#{ @endpoint.end_with?('/') ? '' : '/' }#{path}".gsub( %r{/+$}, '/' )
-      _parse( Nokogiri::HTML( open(uri).read, nil, @encoding ) )
+      _parse( Nokogiri::XML( open(uri).read, nil, @encoding ) )
     end
 
     #
@@ -216,8 +215,8 @@ module Philologic # :nodoc:
         return children
       elsif !doc.css('div.philologic_concordance').first.nil?
         results = []
-        doc.css('div.philologic_concordance').css('a.philologic_cite').each do |cite|
-          results << Philologic::Link.new(self, cite) # TODO Why not a select?
+        doc.css('div.philologic_concordance').css('li.philologic_occurence').each do |occurence|
+          results << Philologic::Link.new(self, occurence) # TODO Why not a select?
         end
         return results
       elsif !doc.css('div.philologic_object').first.nil?
@@ -300,10 +299,20 @@ module Philologic # :nodoc:
     # +document+:: Nokogiri document
     #
     def initialize(client, document)
+      raise('nil client') if client.nil?
+      raise('nil document') if document.nil?
       @client     = client
       @doc        = document
-      # TODO Prefer lazy initialization?
       @properties = {}
+
+      if !@doc.css('a.philologic_cite').first.nil?
+        @doc.css('a.philologic_cite').each do |a|
+          a.attributes.keys.each { |k| @properties[k] = a.attributes[k].value }
+        end
+      end
+      if @doc.respond_to?(:attributes)
+        @doc.attributes.keys.each { |k| @properties[k] = @doc.attributes[k].value }
+      end
       @doc.css('span.philologic_property').each do |span|
         @properties[ span.attributes['title'].value ] = span.children.text
       end
@@ -313,7 +322,7 @@ module Philologic # :nodoc:
     # Returns value of attribute +key+ or +nil+.
     #
     def [](key)
-      @doc.attributes.key?(key) ? @doc.attributes[key].value : @properties[key]
+      @properties[key]
     end
 
     #
@@ -323,6 +332,13 @@ module Philologic # :nodoc:
       @client.document( self['href'] )
     end
 
+    #
+    # Returns Philologic::Link text.
+    #
+    def text
+      @doc.text.strip
+    end
+
     #
     # Return String representation of object.
     #

data/test/test_philologic_client.rb

@@ -7,6 +7,7 @@ require 'philologic-client'
 require 'test/unit'
 require 'mocha'
 
+
 class TestPhilologicClient < Test::Unit::TestCase
 
   def setup
@@ -95,7 +96,8 @@ class TestPhilologicClient < Test::Unit::TestCase
 
   def test_get_children
     Philologic::Client.new do |client|
-      children = client.send(:_parse, Nokogiri::HTML( open(@root_file) ) )
+      h = Nokogiri::HTML( open(@root_file) )
+      children = client.send( :_parse, Nokogiri::HTML( open(@root_file) ) )
       client.stubs(:_get).with('/').returns(children)
       assert_equal children, client.children
     end
@@ -124,21 +126,21 @@ class TestPhilologicClient < Test::Unit::TestCase
         assert_not_nil  child['author']
         assert_not_nil  child['filename']
         assert_not_nil  child['title']
-        assert_raise(NoMethodError) { child.text }
+        assert_not_nil  child.text
       end
 
       assert_equal  './1/0/0/0/0/0/0',                          children.first['href']
       assert_equal  'Conon, de Béthune, ca. 1160-1219 or 20.',  children.first['author']
       assert_equal  'Les Chansons',                             children.first['title']
       assert_equal  'TLF.0.tei',                                children.first['filename']
-      assert_raise(NoMethodError) { children.first.text }
+      assert_match  /^Conon, de Béthune, ca. 1160-1219 or 20./, children.first.text
       assert_equal  './1/0/0/0/0/0/0 | Les Chansons | Conon, de Béthune, ca. 1160-1219 or 20.', children.first.to_s
 
-      assert_equal  './1845/0/0/0/0/0/0',                       children.last['href']
-      assert_equal  'Sartre J.-P.',                             children.last['author']
-      assert_equal  'Les Mots',                                 children.last['title']
-      assert_equal  'TLF.1879.tei',                             children.last['filename']
-      assert_raise(NoMethodError) { children.last.text }
+      assert_equal  './1845/0/0/0/0/0/0',                           children.last['href']
+      assert_equal  'Sartre J.-P.',                                 children.last['author']
+      assert_equal  'Les Mots',                                     children.last['title']
+      assert_equal  'TLF.1879.tei',                                 children.last['filename']
+      assert_equal  "Sartre J.-P., Les Mots: TLF.1879.tei",         children.last.text
       assert_equal  './1845/0/0/0/0/0/0 | Les Mots | Sartre J.-P.', children.last.to_s
 
     end
@@ -230,29 +232,28 @@ class TestPhilologicClient < Test::Unit::TestCase
       assert_kind_of  Array, results
       assert_equal    50,    results.size
       results.each do |result|
-        assert_kind_of Philologic::Link,  result
-        assert_equal   'philologic_cite', result['class']
-        assert_not_nil result['href']
-        assert_nil     result['some random attribute']
-        assert_not_nil result['author']
-        assert_nil     result['filename']
-        assert_not_nil result['title']
-        assert_raise(NoMethodError) { result.text }
-      end
+        assert_kind_of  Philologic::Link,  result
+        assert_equal   'philologic_occurence', result['class']
+        assert_not_nil  result['href']
+        assert_nil      result['some random attribute']
+        assert_not_nil  result['author']
+        assert_nil      result['filename']
+        assert_not_nil  result['title']
+        assert_not_nil  result.text
+     end
 
-      assert_equal './54/1/0/0',             results.first['href']
-      assert_equal 'Peiresc N. de',          results.first['author']
-      assert_equal 'Lettres T. 6 1602-1637', results.first['title']
-      assert_nil   results.first['filename']
-      assert_raise(NoMethodError) { results.first.text }
-      assert_equal './54/1/0/0 | Lettres T. 6 1602-1637 | Peiresc N. de', results.first.to_s
+      assert_equal  './54/1/0/0',             results.first['href']
+      assert_equal  'Peiresc N. de',          results.first['author']
+      assert_equal  'Lettres T. 6 1602-1637', results.first['title']
+      assert_match  /^Peiresc N. de, Lettre/, results.first.text
+      assert_equal  './54/1/0/0 | Lettres T. 6 1602-1637 | Peiresc N. de', results.first.to_s
 
-      assert_equal './1792/11/0/0',          results.last['href']
-      assert_equal 'Gide A.',                results.last['author']
-      assert_equal 'Journal 1939-1949',      results.last['title']
-      assert_nil   results.last['filename']
-      assert_raise(NoMethodError) { results.last.text }
-      assert_equal './1792/11/0/0 | Journal 1939-1949 | Gide A.', results.last.to_s
+      assert_equal  './1792/11/0/0',          results.last['href']
+      assert_equal  'Gide A.',                results.last['author']
+      assert_equal  'Journal 1939-1949',      results.last['title']
+      assert_nil    results.last['filename']
+      assert_match  /^Gide A., Journal 193/,  results.last.text
+      assert_equal  './1792/11/0/0 | Journal 1939-1949 | Gide A.', results.last.to_s
 
     end
   end
@@ -268,29 +269,29 @@ class TestPhilologicClient < Test::Unit::TestCase
       assert_kind_of  Array, results
       assert_equal    50,    results.size
       results.each do |result|
-        assert_kind_of Philologic::Link,  result
-        assert_equal   'philologic_cite', result['class']
-        assert_not_nil result['href']
-        assert_nil     result['some random attribute']
-        assert_not_nil result['author']
-        assert_nil     result['filename']
-        assert_not_nil result['title']
-        assert_raise(NoMethodError) { result.text }
+        assert_kind_of  Philologic::Link,  result
+        assert_equal   'philologic_occurence', result['class']
+        assert_not_nil  result['href']
+        assert_nil      result['some random attribute']
+        assert_not_nil  result['author']
+        assert_nil      result['filename']
+        assert_not_nil  result['title']
+        assert_not_nil  result.text
       end
 
-      assert_equal './54/1/0/0',             results.first['href']
-      assert_equal 'Peiresc N. de',          results.first['author']
-      assert_equal 'Lettres T. 6 1602-1637', results.first['title']
-      assert_nil   results.first['filename']
-      assert_raise(NoMethodError) { results.first.text }
-      assert_equal './54/1/0/0 | Lettres T. 6 1602-1637 | Peiresc N. de', results.first.to_s
+      assert_equal  './54/1/0/0',             results.first['href']
+      assert_equal  'Peiresc N. de',          results.first['author']
+      assert_equal  'Lettres T. 6 1602-1637', results.first['title']
+      assert_nil    results.first['filename']
+      assert_match  /^Peiresc N. de/,         results.first.text
+      assert_equal  './54/1/0/0 | Lettres T. 6 1602-1637 | Peiresc N. de', results.first.to_s
 
-      assert_equal './1792/11/0/0',          results.last['href']
-      assert_equal 'Gide A.',                results.last['author']
-      assert_equal 'Journal 1939-1949',      results.last['title']
-      assert_nil   results.last['filename']
-      assert_raise(NoMethodError) { results.last.text }
-      assert_equal './1792/11/0/0 | Journal 1939-1949 | Gide A.', results.last.to_s
+      assert_equal  './1792/11/0/0',          results.last['href']
+      assert_equal  'Gide A.',                results.last['author']
+      assert_equal  'Journal 1939-1949',      results.last['title']
+      assert_nil    results.last['filename']
+      assert_match  /^Gide A., Journal /,     results.last.text
+      assert_equal  './1792/11/0/0 | Journal 1939-1949 | Gide A.', results.last.to_s
 
     end
   end
@@ -397,29 +398,29 @@ class TestPhilologicClient < Test::Unit::TestCase
       assert_equal    48,    links.size
 
       links.each do |link|
-        assert_kind_of Philologic::Link, link
-        assert_equal   'philologic_cite', link['class']
-        assert_not_nil link['href']
-        assert_nil     link['some random attribute']
-        assert_not_nil link['author']
-        assert_nil     link['filename']
-        assert_not_nil link['title']
-        assert_raise(NoMethodError) { link.text }
+        assert_kind_of  Philologic::Link, link
+        assert_equal    'philologic_occurence', link['class']
+        assert_not_nil  link['href']
+        assert_nil      link['some random attribute']
+        assert_not_nil  link['author']
+        assert_nil      link['filename']
+        assert_not_nil  link['title']
+        assert_not_nil  link.text
       end
 
-      assert_equal './1839/3/0/0',                links.first['href']
-      assert_equal 'Beauvoir S. de',              links.first['author']
-      assert_equal 'Memoires Jeune Fille Rangee', links.first['title']
-      assert_nil   links.first['filename']
-      assert_raise(NoMethodError) { links.first.text }
-      assert_equal './1839/3/0/0 | Memoires Jeune Fille Rangee | Beauvoir S. de', links.first.to_s
+      assert_equal  './1839/3/0/0',                 links.first['href']
+      assert_equal  'Beauvoir S. de',               links.first['author']
+      assert_equal  'Memoires Jeune Fille Rangee',  links.first['title']
+      assert_nil    links.first['filename']
+      assert_match  /^Beauvoir S. de/,              links.first.text
+      assert_equal  './1839/3/0/0 | Memoires Jeune Fille Rangee | Beauvoir S. de', links.first.to_s
 
-      assert_equal './1839/4/0/0',                links.last['href']
-      assert_equal 'Beauvoir S. de',              links.last['author']
-      assert_equal 'Memoires Jeune Fille Rangee', links.last['title']
-      assert_nil   links.last['filename']
-      assert_raise(NoMethodError) { links.last.text }
-      assert_equal './1839/4/0/0 | Memoires Jeune Fille Rangee | Beauvoir S. de', links.last.to_s
+      assert_equal  './1839/4/0/0',                 links.last['href']
+      assert_equal  'Beauvoir S. de',               links.last['author']
+      assert_equal  'Memoires Jeune Fille Rangee',  links.last['title']
+      assert_nil    links.last['filename']
+      assert_match  /^Beauvoir S. de,/,             links.last.text
+      assert_equal  './1839/4/0/0 | Memoires Jeune Fille Rangee | Beauvoir S. de', links.last.to_s
 
     end
   end
@@ -526,29 +527,29 @@ class TestPhilologicClient < Test::Unit::TestCase
       assert_equal    48,    links.size
 
       links.each do |link|
-        assert_kind_of Philologic::Link, link
-        assert_equal   'philologic_cite', link['class']
-        assert_not_nil link['href']
-        assert_nil     link['some random attribute']
-        assert_not_nil link['author']
-        assert_nil     link['filename']
-        assert_not_nil link['title']
-        assert_raise(NoMethodError) { link.text }
+        assert_kind_of  Philologic::Link, link
+        assert_equal    'philologic_occurence', link['class']
+        assert_not_nil  link['href']
+        assert_nil      link['some random attribute']
+        assert_not_nil  link['author']
+        assert_nil      link['filename']
+        assert_not_nil  link['title']
+        assert_not_nil  link.text
       end
 
-      assert_equal './1839/3/0/0',                links.first['href']
-      assert_equal 'Beauvoir S. de',              links.first['author']
-      assert_equal 'Memoires Jeune Fille Rangee', links.first['title']
-      assert_nil   links.first['filename']
-      assert_raise(NoMethodError) { links.first.text }
-      assert_equal './1839/3/0/0 | Memoires Jeune Fille Rangee | Beauvoir S. de', links.first.to_s
-
-      assert_equal './1839/4/0/0',                links.last['href']
-      assert_equal 'Beauvoir S. de',              links.last['author']
-      assert_equal 'Memoires Jeune Fille Rangee', links.last['title']
-      assert_nil   links.last['filename']
-      assert_raise(NoMethodError) { links.last.text }
-      assert_equal './1839/4/0/0 | Memoires Jeune Fille Rangee | Beauvoir S. de', links.last.to_s
+      assert_equal  './1839/3/0/0',                 links.first['href']
+      assert_equal  'Beauvoir S. de',               links.first['author']
+      assert_equal  'Memoires Jeune Fille Rangee',  links.first['title']
+      assert_nil    links.first['filename']
+      assert_match  /^Beauvoir S. de/,              links.first.text
+      assert_equal  './1839/3/0/0 | Memoires Jeune Fille Rangee | Beauvoir S. de', links.first.to_s
+
+      assert_equal  './1839/4/0/0',                 links.last['href']
+      assert_equal  'Beauvoir S. de',               links.last['author']
+      assert_equal  'Memoires Jeune Fille Rangee',  links.last['title']
+      assert_nil    links.last['filename']
+      assert_match  /^Beauvoir S. de/,              links.last.text
+      assert_equal  './1839/4/0/0 | Memoires Jeune Fille Rangee | Beauvoir S. de', links.last.to_s
 
     end
   end

data/test/test_philologic_link.rb

@@ -0,0 +1,101 @@
+# encoding: utf-8
+
+require 'simplecov'
+SimpleCov.start
+
+require 'philologic-client'
+require 'test/unit'
+
+
+class TestPhilologicLink < Test::Unit::TestCase
+
+  def setup
+    @endpoint = File.expand_path( File.join( File.dirname(__FILE__), 'data' ) )
+    @client   = Philologic::Client.new(@endpoint)
+    @doc      = Nokogiri::HTML("")
+  end
+
+
+  def test_initialization_with_nil_client
+    assert_raise(RuntimeError, 'nil client') {  Philologic::Link.new( nil, @doc ) }
+  end
+
+  def test_initialization_with_nil_document
+    assert_raise(RuntimeError, 'nil document') {  Philologic::Link.new( @client, nil ) }
+  end
+
+  def test_philologic_cite
+    html = <<EOHTML;
+<a class='philologic_cite' href='./1/0/0/0/0/0/0'>
+  <span class='philologic_property' title='author'>Conon, de Béthune, ca. 1160-1219 or 20.</span>,
+  <span class='philologic_property' title='title'>Les Chansons</span>:
+  <span class='philologic_property' title='filename'>TLF.0.tei</span>
+</a>
+EOHTML
+    link = Philologic::Link.new( @client, Nokogiri::XML(html).css('a.philologic_cite').first )
+
+    assert_kind_of  Philologic::Link,   link
+
+    assert_not_nil  link['class']
+    assert_equal    'philologic_cite',  link['class']
+
+    assert_not_nil  link['href']
+    assert_equal    './1/0/0/0/0/0/0',  link['href']
+
+    assert_not_nil  link['author']
+    assert_equal    'Conon, de Béthune, ca. 1160-1219 or 20.',  link['author']
+  
+    assert_not_nil  link['title']
+    assert_equal    'Les Chansons',     link['title']
+
+    assert_not_nil  link['filename']
+    assert_equal    'TLF.0.tei',        link['filename']
+
+    assert_nil      link['some random attribute']
+
+    assert_equal    "Conon, de Béthune, ca. 1160-1219 or 20.,\n  Les Chansons:\n  TLF.0.tei",   link.text
+
+    assert_equal    './1/0/0/0/0/0/0 | Les Chansons | Conon, de Béthune, ca. 1160-1219 or 20.', link.to_s
+  end
+
+  def test_philologic_occurence
+    html = <<EOHTML;
+<li class='philologic_occurence'>
+  <a class='philologic_cite' href='./54/1/0/0'>
+    <span class='philologic_property' title='author'>Peiresc N. de</span>,
+    <span class='philologic_property' title='title'>Lettres T.  6 1602-1637</span>
+    <span class='philologic_property' title='date'>(1625)</span>
+  </a>
+  ...re pas l'adresse directement à nous ny à voz lettres ny au pacquet, mais à Monsieur
+    <span class="hilite">Sartre</span>
+  respondent de Monsr Segnier qui nous les rendra gentilment. Car je...
+</li>
+EOHTML
+    link = Philologic::Link.new( @client, Nokogiri::XML(html).css('li.philologic_occurence').first )
+
+    assert_kind_of  Philologic::Link,   link
+
+    assert_not_nil  link['class']
+    assert_equal    'philologic_occurence',     link['class']
+
+    assert_not_nil  link['href']
+    assert_equal    './54/1/0/0',               link['href']
+
+    assert_not_nil  link['author']
+    assert_equal    'Peiresc N. de',            link['author']
+  
+    assert_not_nil  link['title']
+    assert_equal    'Lettres T.  6 1602-1637',  link['title']
+
+    assert_nil      link['filename']
+
+    assert_nil      link['some random attribute']
+
+    assert_match    /^Peiresc N. de,/,          link.text
+    assert_match    /ntilment. Car je\.\.\.$/,  link.text
+
+    assert_equal    './54/1/0/0 | Lettres T.  6 1602-1637 | Peiresc N. de', link.to_s
+  end
+
+end
+

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: philologic-client
 version: !ruby/object:Gem::Version
-  version: 0.0.10
+  version: 0.0.11
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-01-18 00:00:00.000000000Z
+date: 2012-02-09 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
-  requirement: &70093876599220 !ruby/object:Gem::Requirement
+  requirement: &70128325064720 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,10 +21,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70093876599220
+  version_requirements: *70128325064720
 - !ruby/object:Gem::Dependency
   name: rdoc
-  requirement: &70093876864820 !ruby/object:Gem::Requirement
+  requirement: &70128325064280 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -32,10 +32,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70093876864820
+  version_requirements: *70128325064280
 - !ruby/object:Gem::Dependency
   name: mocha
-  requirement: &70093879163480 !ruby/object:Gem::Requirement
+  requirement: &70128325063780 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -43,10 +43,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70093879163480
+  version_requirements: *70128325063780
 - !ruby/object:Gem::Dependency
   name: rdoc-readme
-  requirement: &70093879782440 !ruby/object:Gem::Requirement
+  requirement: &70128325052240 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -54,10 +54,10 @@ dependencies:
         version: 0.1.2
   type: :development
   prerelease: false
-  version_requirements: *70093879782440
+  version_requirements: *70128325052240
 - !ruby/object:Gem::Dependency
   name: simplecov
-  requirement: &70093879916440 !ruby/object:Gem::Requirement
+  requirement: &70128325051400 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -65,10 +65,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70093879916440
+  version_requirements: *70128325051400
 - !ruby/object:Gem::Dependency
   name: nokogiri
-  requirement: &70093880007360 !ruby/object:Gem::Requirement
+  requirement: &70128325050940 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -76,7 +76,7 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *70093880007360
+  version_requirements: *70128325050940
 description: Ruby client for interacting with the Philologic API.
 email:
 - blair.christensen@gmail.com
@@ -101,6 +101,7 @@ files:
 - test/data/query_sartre.html
 - test/data/root_file.html
 - test/test_philologic_client.rb
+- test/test_philologic_link.rb
 homepage: https://github.com/blairc/philologic-client/
 licenses: []
 post_install_message: 
@@ -115,7 +116,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -1051070454611838584
+      hash: -3858874117340008545
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
@@ -124,7 +125,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -1051070454611838584
+      hash: -3858874117340008545
 requirements: []
 rubyforge_project: philologic-client
 rubygems_version: 1.8.7
@@ -140,3 +141,4 @@ test_files:
 - test/data/query_sartre.html
 - test/data/root_file.html
 - test/test_philologic_client.rb
+- test/test_philologic_link.rb