RubyGems - raakt - Versions diffs - 0.5.2 → 0.5.4 - Mend

raakt 0.5.2 → 0.5.4

Files changed (11) hide show

data/lib/raakt.rb CHANGED Viewed

@@ -44,10 +44,11 @@ module Raakt
 	:missing_input_alt	 => "Missing alt attribute for image button with id/name '%s'.",
 	:missing_input_alt_text	 => "Missing alt text for image button with id/name '%s'.",
 	:missing_area_alt	 => "Missing alt attribute for area with id/name '%s'.",
-	:missing_area_alt_text	 => "Missing alt text for area with id/name '%s'."
+	:missing_area_alt_text	 => "Missing alt text for area with id/name '%s'.",
+	:difficult_word => "Vocabulary: %s"
   }
-  VERSION = "0.5.2"
+  VERSION = "0.5.4"
   class ErrorMessage
@@ -78,19 +79,21 @@ module Raakt
   class Test
-    attr_accessor :html, :headers, :user_agent, :ignore_bi
+    attr_accessor :html, :headers, :ignore_bi, :wordlist
-    def initialize(html=nil, headers=nil)
+    def initialize(html=nil, headers=nil, wordlist=nil)
       @html = html
 	  @headers = headers
+	  @wordlist = wordlist
       self.doc = @html if html
 	  self.headers = @headers if headers
+	  self.wordlist = @wordlist if wordlist
 	  @ignore_bi = false
     end
 	# Set the HTML used in the test.
     def doc=(html)
-	  Hpricot.buffer_size = 262144 #Allow for asp.net bastard-sized viewstate attributes...
+	  Hpricot.buffer_size = 524288 #Allow for asp.net bastard-sized viewstate attributes...
       @doc = Hpricot(html)
     end
@@ -226,7 +229,7 @@ module Raakt
 	# Verify that the document has at least one h1 element.
     def check_has_heading
-      return [ErrorMessage.new(:missing_heading)] if (@doc/'h1').empty?
+      return [ErrorMessage.new(:missing_heading)] if (@doc/"h1").empty?
       []
     end
@@ -257,7 +260,7 @@ module Raakt
     end
-	# Verify that the document does not have any nested tabled. This is indicative of a table-based layout.
+	# Verify that the document does not have any nested tables. This is indicative of a table-based layout.
     def check_for_nested_tables
       messages = []
@@ -283,6 +286,7 @@ module Raakt
       	hasth = false
         hasth = true unless (table/">tr>th").empty?
         hasth = true unless (table/">thead>tr>th").empty?
+        hasth = true unless (table/">tbody>tr>th").empty?
         messages << ErrorMessage.new(:missing_th, currenttable.to_s) unless hasth
@@ -418,8 +422,45 @@ module Raakt
       end.map { ErrorMessage.new(:has_meta_refresh) }
     end
+	def check_difficult_words
+      messages = []
+		if @wordlist
+			# get document text (and all title and ait attributes but remove blockquote and q elements)
+			# remove q and blockquotes
+			@doc.search("blockquote").remove
+			@doc.search("q").remove
+			doctext = @doc.inner_text
+			#add alt texts
+			@doc.search("*[@alt]").each { |item|
+				doctext += " " + item['alt']
+				doctext += ", "
+			}
+			#add title texts
+			@doc.search("*[@title]").each { |item|
+				doctext += " " + item['title']
+				doctext += ", "
+			}
+			@wordlist.each { |key, value|
+				re = Regexp.new("\\b" + key.sub(/ /, "\\s+") + "\\b", true)
+				if doctext =~ re
+					# loop over all keys in wordlist
+					messages << ErrorMessage.new(:difficult_word, value)
+				end
+			}
+		end
+		return messages
+	end
-    #Utility methods
+    # Utility methods
     def headings
       headings = []

data/tests/difficult_words1.htm ADDED Viewed

@@ -0,0 +1,18 @@
+<html>
+  <head>
+    <title>This is a document with some difficult words</title>
+  </head>
+  <body>
+	<h1>Difficult words. Reticent and taciturn should be found</h1>
+	<h2><a href="/news1">New sitem 1</a></h2>
+    <p>He was reticent. This is the text for the first news item. <a href="/news1">Read more </a></p>
+	<h2><a href="/news2" title="This is the word taciturn in a title attribute">New sitem 2</a></h2>
+    <p>This is the text for the second news item. <a href="/news2">
+	Read more</a></p>
+	<h2><a href="/news1">New sitem 3</a></h2>
+    <p>This is the text for the third news item. <a href="/news3">Read
+	more</a></p>
+	<h2><a href="/news1">New sitem 4</a></h2>
+    <p>This is the text for the fourth news item. <a href="/news4">Read&nbsp;more</a></p>
+  </body>
+</html>

data/tests/difficult_words2.htm ADDED Viewed

@@ -0,0 +1,18 @@
+<html>
+  <head>
+    <title>This is a document with some difficult words</title>
+  </head>
+  <body>
+	<h1>Difficult words, on ein the alt attribute</h1>
+	<h2><a href="/news1">New sitem 1</a></h2>
+    <p>This is the text for the first news item. <a href="/news1">Read more </a></p>
+	<h2><img src="/news2" alt="This is the word taciturn in an alt attribute">New sitem 2</h2>
+    <p>This is the text for the second news item. <a href="/news2">
+	Read more</a></p>
+	<h2><a href="/news1">New sitem 3</a></h2>
+    <p>This is the text for the third news item. <a href="/news3">Read
+	more</a></p>
+	<h2><a href="/news1">New sitem 4</a></h2>
+    <p>This is the text for the fourth news item. <a href="/news4">Read&nbsp;more</a></p>
+  </body>
+</html>

data/tests/difficult_words3.htm ADDED Viewed

@@ -0,0 +1,18 @@
+<html>
+  <head>
+    <title>This is a document with some difficult words</title>
+  </head>
+  <body>
+	<h1>Difficult words, on in the title attribute</h1>
+	<h2><a href="/news1">New sitem 1</a></h2>
+    <p>This is the text for the first news item. <a href="/news1">Read more </a></p>
+	<h2><img src="/news2" alt="This is the word taciturn in an alt attribute">New sitem 2</h2>
+    <p title="And here is reticent.">This is the text for the second news item. <a href="/news2">
+	Read more</a></p>
+	<h2><a href="/news1">New sitem 3</a></h2>
+    <p>This is the text for the third news item. <a href="/news3">Read
+	more</a></p>
+	<h2><a href="/news1">New sitem 4</a></h2>
+    <p>This is the text for the fourth news item. <a href="/news4">Read&nbsp;more</a></p>
+  </body>
+</html>

data/tests/difficult_words4.htm ADDED Viewed

@@ -0,0 +1,16 @@
+<html>
+  <head>
+    <title>This is a document with some difficult words</title>
+  </head>
+  <body>
+	<h1>Difficult words, on in the title attribute</h1>
+	<h2><a href="/news1">New sitem 1</a></h2>
+    <p>This is the text for the first news item. <a href="/news1">Read more </a></p>
+	<h2><img src="/news2" alt="This is the word taciturn and reticent in an alt attribute">New sitem 2</h2>
+	<h2><a href="/news1">New sitem 3</a></h2>
+    <p>This is the text for the third news item. <a href="/news3">Read
+	more</a></p>
+	<h2><a href="/news1">New sitem 4</a></h2>
+    <p>This is the text for the fourth news item. <a href="/news4">Read&nbsp;more</a></p>
+  </body>
+</html>

data/tests/difficult_words5.htm ADDED Viewed

@@ -0,0 +1,16 @@
+<html>
+  <head>
+    <title>This is a document with some difficult words</title>
+  </head>
+  <body>
+	<h1>Words that are not in the list</h1>
+	<h2><a href="/news1">New sitem 1</a></h2>
+    <p>This is the superreticent text for the first news item. <a href="/news1">Read more </a></p>
+	<h2><img src="/news2" alt="">New sitem 2</h2>
+	<h2><a href="/news1">New sitem 3</a></h2>
+    <p>This is the text for the third news item. <a href="/news3">Read
+	more</a></p>
+	<h2><a href="/news1">New sitem 4</a></h2>
+    <p>This is the text for the fourth news item. <a href="/news4">Read&nbsp;more</a></p>
+  </body>
+</html>

data/tests/difficult_words6.htm ADDED Viewed

@@ -0,0 +1,18 @@
+<html>
+  <head>
+    <title>This is a document with some difficult words</title>
+  </head>
+  <body>
+	<h1>Checking for a  phrase.</h1>
+	<h2><a href="/news1">New sitem 1</a></h2>
+	<p><q>Reticent</q> he said. <a href="/news1">Read more </a></p>
+	<h2><img src="/news2" alt="">New sitem 2</h2>
+	<h2><a href="/news1">New sitem 3</a></h2>
+    <p>Here are some accented chars: "anhängiggöra". <a href="/news3">Read
+	more</a></p>
+	<h2><a href="/news1">New sitem 4</a></h2>
+	<blockquote>
+    <p>This is the word reticent inside of a blockquote.</p>
+	</blockquote>
+  </body>
+</html>

data/tests/difficult_words7.htm ADDED Viewed

@@ -0,0 +1,8 @@
+<html>
+  <head>
+    <title>This is a document with some difficult words</title>
+  </head>
+  <body>
+    <p>This is the ReTicent text for the first news item. <a href="/news1">Read more </a></p>
+  </body>
+</html>

data/tests/raakt_test.rb CHANGED Viewed

@@ -205,6 +205,9 @@ class RaaktTest < Test::Unit::TestCase
     @raakt.doc = data_tabledoc2
     assert_equal 2, @raakt.check_tables.length
+    @raakt.doc = data_tabledoc7
+    assert_equal 0, @raakt.check_tables.length
     # More accurate count here due to hpricot
     @raakt.doc = data_full_berg
     assert_equal 21, @raakt.check_tables.length
@@ -412,5 +415,98 @@ class RaaktTest < Test::Unit::TestCase
     @raakt.doc = data_xhtmldoc1
     assert_equal 0, @raakt.check_refresh.length
   end
+  def test_difficult_words_combinations
+	  #words tat are part of other words should not be flagged
+	  @raakt.wordlist = {
+		  "reticent" => "Use a 'did not want to say' instead.",
+		  "taciturn" => "Use a different phrase.",
+	  }
+	  @raakt.doc = data_difficult_words5
+	  assert_equal 0, @raakt.check_difficult_words.length
+  end
+  def test_difficult_words_ignore_case
+	  #words tat are part of other words should not be flagged
+	  @raakt.wordlist = {
+		  "reticent" => "Use a 'did not want to say' instead.",
+	  }
+	  @raakt.doc = data_difficult_words7
+	  assert_equal 1, @raakt.check_difficult_words.length
+  end
+  def test_difficult_words_exclude_blockquoted_text
+	  #words that are part of a blockquote should not be flagged
+	  @raakt.wordlist = {
+		  "reticent" => "Use a 'did not want to say' instead.",
+	  }
+	  @raakt.doc = data_difficult_words6
+	  assert_equal 0, @raakt.check_difficult_words.length
+  end
+  def test_difficult_words_exclude_text_in_q_elements
+	  #words that are part of a q-element should not be flagged
+	  @raakt.wordlist = {
+		  "reticent" => "Use a 'did not want to say' instead.",
+	  }
+	  @raakt.doc = data_difficult_words6
+	  assert_equal 0, @raakt.check_difficult_words.length
+  end
+  def test_difficult_words_accented_chars
+	  #Phrases
+	  @raakt.wordlist = {
+		  "anhängiggöra" => "Använd 'väcka/inleda' istället",
+	  }
+	  @raakt.doc = data_difficult_words6
+	  assert_equal 1, @raakt.check_difficult_words.length
+  end
+  def test_difficult_words_phrases
+	  #Phrases
+	  @raakt.wordlist = {
+		  "a phrase" => "Use something else instead.",
+	  }
+	  @raakt.doc = data_difficult_words6
+	  assert_equal 1, @raakt.check_difficult_words.length
+  end
+  def test_difficult_words
+	  @raakt.wordlist = {
+		  "reticent" => "Use a 'did not want to say' instead.",
+		  "taciturn" => "Use a different phrase.",
+	  }
+	  @raakt.doc = data_xhtmldoc1
+	  assert_equal 0, @raakt.check_difficult_words.length
+	  @raakt.doc = data_difficult_words1
+	  assert_equal 2, @raakt.check_difficult_words.length
+	  #Make sure text in alt attribute is found
+	  @raakt.doc = data_difficult_words2
+	  assert_equal 1, @raakt.check_difficult_words.length
+	  #Make sure text in title attribute is found
+	  @raakt.doc = data_difficult_words3
+	  assert_equal 2, @raakt.check_difficult_words.length
+	  @raakt.doc = data_difficult_words4
+	  assert_equal 2, @raakt.check_difficult_words.length
+  end
 end

data/tests/tabledoc7.htm ADDED Viewed

	@@ -0,0 +1 @@
1	+ <table><tbody><tr><th>Fondkapital</th><td>227,5 mdkr</td></tr></tbody></table>

metadata CHANGED Viewed

@@ -1,33 +1,35 @@
 --- !ruby/object:Gem::Specification
-rubygems_version: 0.9.0
-specification_version: 1
 name: raakt
 version: !ruby/object:Gem::Version
-  version: 0.5.2
-date: 2007-03-31 00:00:00 +02:00
-summary: A toolkit to find accessibility issues in HTML documents.
-require_paths:
-- lib
-email: peter.krantzNODAMNSPAM@gmail.com
-homepage: http://raakt.rubyforge.org
-rubyforge_project: raakt
-description:
-autorequire: raakt
-default_executable:
-bindir: bin
-has_rdoc: true
-required_ruby_version: !ruby/object:Gem::Version::Requirement
-  requirements:
-  - - ">="
-    - !ruby/object:Gem::Version
-      version: 1.8.2
-  version:
+  version: 0.5.4
 platform: ruby
-signing_key:
-cert_chain:
-post_install_message:
 authors:
 - Peter Krantz
+autorequire: raakt
+bindir: bin
+cert_chain: []
+date: 2009-01-19 00:00:00 +01:00
+default_executable:
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: hpricot
+  type: :runtime
+  version_requirement:
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: "0.5"
+    version:
+description:
+email: peter.krantzNODAMNSPAM@gmail.com
+executables: []
+extensions: []
+extra_rdoc_files: []
 files:
 - lib/iso_language_codes.rb
 - lib/raakt.rb
@@ -37,6 +39,13 @@ files:
 - tests/bdoc.htm
 - tests/charset_nocharset_specified.htm
 - tests/charset_utf8.htm
+- tests/difficult_words1.htm
+- tests/difficult_words2.htm
+- tests/difficult_words3.htm
+- tests/difficult_words4.htm
+- tests/difficult_words5.htm
+- tests/difficult_words6.htm
+- tests/difficult_words7.htm
 - tests/embeddoc1.htm
 - tests/empty.htm
 - tests/emptytitledoc.htm
@@ -90,28 +99,35 @@ files:
 - tests/tabledoc4.htm
 - tests/tabledoc5.htm
 - tests/tabledoc6.htm
+- tests/tabledoc7.htm
 - tests/tablelayoutdoc.htm
 - tests/test_helper.rb
 - tests/xhtmldoc1.htm
-test_files:
-- tests/raakt_test.rb
+has_rdoc: true
+homepage: http://raakt.rubyforge.org
+post_install_message:
 rdoc_options: []
-extra_rdoc_files: []
-executables: []
-extensions: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: 1.8.2
+  version:
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - ">="
+    - !ruby/object:Gem::Version
+      version: "0"
+  version:
 requirements: []
-dependencies:
-- !ruby/object:Gem::Dependency
-  name: hpricot
-  version_requirement:
-  version_requirements: !ruby/object:Gem::Version::Requirement
-    requirements:
-    - - ">="
-      - !ruby/object:Gem::Version
-        version: "0.5"
-    version:
+rubyforge_project: raakt
+rubygems_version: 1.3.1
+signing_key:
+specification_version: 2
+summary: A toolkit to find accessibility issues in HTML documents.
+test_files:
+- tests/raakt_test.rb