raakt 0.4 → 0.5

data/lib/raakt.rb

@@ -1,101 +1,110 @@
-# :title: Ruby Accessibility Analysis Kit
-# =Ruby Accessibility Analysis Kit
+# == The Ruby Accessibility Analysis Kit (RAAKT)
+# :title: Ruby Accessibility Analysis Kit (RAAKT)
+# Author::    Peter Krantz (http://www.peterkrantz.com/)
+# License::   See LICENSE file
 #
-# See README for a full explanation of this library.
-
+# RAAKT is a toolkit to find accessibility issues in HTML documents. RAAKT can be used as part of a an automatic test procedure or as a standalone module for mass validation of all pages in a site.
+# 
+# The ambition has been to provide tests that can be fully automated. Currently, none of the included tests should fail for any web page.
+#
+# Many of the tests included here map to tests defined in the Unified Web Evaluation Methodology (UWEM[http://www.wabcluster.org/uwem/tests/]). See note for each test to find the corresponding UWEM test.
+# 
+# == Output
+# RAAKT output is in the form of an array of Raakt::ErrorMessage objects.
+#
+# == Contributions
+# Thanks to Derek Perrault for refactoring RAAKT to use Hpricot[http://code.whytheluckystiff.net/hpricot/] while at the same time making the code more readable.
+#
+# == Example usage
+# See the examples folder for a small script that shows how to retrieve a remote web page and perform an accessibility test on it.
 module Raakt
-  require 'rubyful_soup'
+  require 'hpricot'
   
   MESSAGES = {
-    "missingtitle"    => "The title element is missing. Provide a descriptive title for your document.",
-    "emptytitle"      => "The title element is empty. Provide a descriptive title for your document.",  
-    "missingalt"      => "Missing alt attribute for image (with src '%s').",
-    "missingheading"  => "Missing first level heading (h1). Provide at least one first level heading describing document content.",
-    "wronghstructure" => "Document heading structure is wrong.",
-    "firsthnoth1"     => "The first heading is not h1.",
-    "hasnestedtables" => "You have one or more nested tables.",
-    "missingsemantics"=> "You have used %s for visual formatting. Use CSS instead.",
-    "hasflicker"      => "You have used <blink> or <marquee>. These may create accessibility issues and should be avoided.",
-    "missinglanginfo" => "Document language information is missing. Use the lang attribute on the html element.",
-    "missingth"       => "Missing table headings (th) for table #%s.",
-    "ambiguouslinktext" => "One or more links have the same link text ('%s'). Make sure each link is unambiguous.",
-    "fieldmissinglabel" => "A field (with id/name '%s') is missing a corresponding label element. Make sure a label exists for all visible fields.",
-    "missingframetitle" => "Missing title attribute for frame with url %s",
-    "hasmetarefresh"  => "Client side redirect (meta refresh) detected. Use server side redirection instead."
+    :missing_title       => "The title element is missing. Provide a descriptive title for your document.",
+    :empty_title         => "The title element is empty. Provide a descriptive title for your document.",  
+    :missing_alt         => "Missing alt attribute for image (with src '%s').",
+    :missing_heading     => "Missing first level heading (h1). Provide at least one first level heading describing document content.",
+    :wrong_h_structure   => "Document heading structure is wrong.",
+    :first_h_not_h1      => "The first heading is not h1.",
+    :has_nested_tables   => "You have one or more nested tables.",
+    :missing_semantics   => "You have used %s for visual formatting. Use CSS instead.",
+    :has_flicker         => "You have used <blink> and/or <marquee>. These may create accessibility issues and should be avoided.",
+    :missing_lang_info   => "Document language information is missing. Use the lang attribute on the html element.",
+    :missing_th          => "Missing table headings (th) for table #%s.",
+    :ambiguous_link_text => "One or more links have the same link text ('%s'). Make sure each link is unambiguous.",
+    :field_missing_label => "A field (with id/name '%s') is missing a corresponding label element. Make sure a label exists for all visible fields.",
+    :missing_frame_title => "Missing title attribute for frame with url %s",
+    :has_meta_refresh    => "Client side redirect (meta refresh) detected. Use server side redirection instead.",
+	:charset_mismatch	 => "The character set specified in the HTTP headers does not match that specified in the markup.",
+	:embed_used			 => "You have used the embed element. It does not provide a way to express a text representation.",
+	:wrong_lang_code	 => "You have used a language code ('%s') not recognized in the ISO 639 standard.",
+	:fieldset_missing_legend => "Missing legend element for fieldset #%s.",
+	: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'."
   } 
   
-  VERSION = "0.4"
+  VERSION = "0.5"
   
   class ErrorMessage
   
     attr_reader :eid, :text, :note
-  
+
     def initialize(eid, note=nil)
       @eid = eid
+      
       if note
-        @text = MESSAGES[eid].sub(/%s/, note)
+        @text = MESSAGES[@eid].sub(/%s/, note)
       else
-        @text = MESSAGES[eid]
+        @text = MESSAGES[@eid]
       end
       @note = note
     end
-  
+
     def to_s
-      @eid + ": " + @text
+      "#{@eid}: #{@text}"
     end
-  end
-
 
+	# Return single error message as an xml element.
+  	def to_xml
+  		"<message id=\"#{@eid}\">#{@text}</message>"
+  	end
+  end
 
 
 
   class Test
 
-    attr_accessor :soup, :html, :user_agent
-  
-    def initialize(html=nil)
+    attr_accessor :html, :headers, :user_agent, :ignore_bi
+
+    def initialize(html=nil, headers=nil)
       @html = html
-      @soup = BeautifulSoup.new(@html) if html
-      @user_agent = "Mozilla/5.0 (RAAKT v#{VERSION}; http://raakt.rubyforge.org; The Ruby Accessibility Analysis Kit)"
-    end
-    
-    def feed(html)
-      @html = html || ""
-      if @html.length > 0
-        @soup = BeautifulSoup.new(@html)
-      else
-        raise "You called feed with no data. There is nothing to check."
-      end
+	  @headers = headers
+      self.doc = @html if html
+	  self.headers = @headers if headers
+	  @ignore_bi = false 
     end
-    
-    
-    
-    def feedurl(url)
-      if url.length == 0
-        raise "You called feedurl with a blank url. There is nothing to check."
-      end
-      
-      #Clean the url and make sure protocol and trailing slash is available
-      url = "http://" + url unless url[0..3] == "http"
-      
-      require 'open-uri'
 
-      open(url, "User-Agent" => @user_agent) { |f|
-        @html = f.read || ""
-      }
-      
-      if @html.length == 0
-        raise "Could not fetch html from the url #{url}. There is nothing to check."
-      else
-        @soup = BeautifulSoup.new(@html)
-      end      
-    
+	# Set the HTML used in the test.
+    def doc=(html)
+	  Hpricot.buffer_size = 262144 #Allow for asp.net bastard-sized viewstate attributes...
+      @doc = Hpricot(html)
     end
     
-    
-    
+    # Set HTML headers to be used in the test. Headers are necessary for some tests (e.g. to check encoding).
+    def headers=(headers)
+		if headers
+      		@headers = downcase_hash_keys(headers)
+		else
+			@headers = nil
+		end
+    end
+
+
+	# Call all check methods.
     def all
-      #Call all check methods
       messages = []
       
       self.methods.each do |method|
@@ -107,69 +116,121 @@ module Raakt
       return messages
     end
 
+
+	# Verify that all fieldset elements have a legend child element. See UWEM 1.0 Test 12.3_HTML_01.
+	def check_fieldset_legend
+		messages = []
+		fieldsets = (@doc/"fieldset")
+		fieldset_instance = 1
+		for fieldset in fieldsets 
+			if (fieldset/"legend").empty?
+				messages << ErrorMessage.new(:fieldset_missing_legend, fieldset_instance.to_s)				
+			end
+			fieldset_instance += 1
+		end
+		messages
+	end
+
+
+	# Verify that the embed element isn't used. See UWEM 1.0 Test 1.1_HTML_06.
+	def check_embed
+		return [ErrorMessage.new(:embed_used)] unless (@doc/'embed').empty?
+		[]
+	end
+
   
+	# Verify that the charater set specified in HTTP headers match that specidied in the HTML meta element.
+	def check_character_set
+		messages = []
+		header_charset = meta_charset = ""
+		if @headers and @headers.length > 0 then
+			if @headers.has_key?("content-type")
+				header_charset = parse_charset(@headers["content-type"].to_s)
+			end
+
+			#get meta element charset
+			meta_elements = @doc.search("//meta[@http-equiv]")
+			for element in meta_elements do
+				if element["http-equiv"].downcase == "content-type" then
+					meta_charset = parse_charset(element["content"])
+				end
+			end
+
+			if header_charset.length > 0 and meta_charset.length > 0
+				unless meta_charset == header_charset
+					messages << ErrorMessage.new(:charset_mismatch) 
+				end
+			end
+		end
+
+		return messages
+
+	end
+
+
+	# Verify that all input type=image elements have an alt attribute.
+	def check_input_type_img
+		#Covers UWEM 1.0 Test 1.1_HTML_01
+
+		messages = []
+		image_input_buttons = @doc.search("input").select { |element| element['type'] =~ /image/i }
+		image_input_buttons.map { |element| 
+			unless element['alt']
+				messages << ErrorMessage.new(:missing_input_alt, element['name'] || element['id'] || "") 
+			else
+				if element['alt'].length == 0
+					messages << ErrorMessage.new(:missing_input_alt_text, element['name'] || element['id'] || "")
+				end
+			end
+		}
+
+		messages
+	end
+
+
+	# Verify that all img elements have an alt attribute.
     def check_images
-      #soup = BeautifulSoup.new(html)
-      images = @soup.find_all("img")
-      messages = []
-    
-      for image in images:
-          if image["alt"] == nil:
-            img_src = image["src"] || ""
-            messages << ErrorMessage.new("missingalt", img_src)
-          end
-      end
-      
-      return messages  
+      no_alt_images = (@doc/"img:not([@alt])")
+      no_alt_images.map { |img| ErrorMessage.new(:missing_alt, img['src']) }
     end
 
   
-    def check_title
-      title = @soup.find("title")
-      messages = []
-    
-      if title
-        titletext = normalize_text(title.string)
-        if titletext.length == 0
-          messages << ErrorMessage.new("emptytitle")
-        end
-      else
-        messages << ErrorMessage.new("missingtitle")
-      end
-      
-      return messages 
-    end
+	# Verify that all area elements have a non-empty alt attribute. See UWEM 1.0 Test 1.1_HTML_01 (together with check_images)
+    def check_areas
+		messages = []
+		area_elements = (@doc/"area")
+		area_elements.map { |element| 
+			unless element['alt']
+				messages << ErrorMessage.new(:missing_area_alt, element['name'] || element['id'] || "unknown") 
+			else
+				if element['alt'].length == 0
+					messages << ErrorMessage.new(:missing_area_alt_text, element['name'] || element['id'] || "unknown")
+				end
+			end
+		}
 
-  
-    def check_has_heading
-      messages = []  
-      
-      if @soup.find_all("h1").length == 0
-        messages << ErrorMessage.new("missingheading")
-      end
-      
-      return messages
+		messages
     end
 
 
-    def headings
-      headings = []
-      headings.push(@soup.find_all("h1")) if @soup.find_all("h1").length > 0
-      headings.push(@soup.find_all("h2")) if @soup.find_all("h2").length > 0
-      headings.push(@soup.find_all("h3")) if @soup.find_all("h3").length > 0
-      headings.push(@soup.find_all("h4")) if @soup.find_all("h4").length > 0
-      headings.push(@soup.find_all("h5")) if @soup.find_all("h5").length > 0
-      headings.push(@soup.find_all("h6")) if @soup.find_all("h6").length > 0
-      
-      return headings.flatten
+
+	# Verify that the document has a non-empty title element.
+    def check_title
+      title = @doc.at('title')
+      return [ErrorMessage.new(:missing_title)] unless title
+      return [ErrorMessage.new(:empty_title)] if normalize_text(title.inner_html).empty?
+      []			
     end
 
-    
-    def level(heading)
-      Integer(heading[1,1])
+  
+	# Verify that the document has at least one h1 element.
+    def check_has_heading
+      return [ErrorMessage.new(:missing_heading)] if (@doc/'h1').empty?
+      []
     end
 
-    
+
+	# Verify that heading elements (h1-h6) appear in the correct order (no levels skipped). See UWEM 1.0 Test 3.5_HTML_03.
     def check_document_structure
       messages = []
       currentitem = 0
@@ -178,11 +239,11 @@ module Raakt
       for heading in docheadings
         if currentitem == 0
           if level(heading.name) != 1
-            messages << ErrorMessage.new("firsthnoth1", "h" + heading.name[1,1])
+            messages << ErrorMessage.new(:first_h_not_h1, "h" + heading.name[1,1])
           end
         else
           if level(heading.name) - level(docheadings[currentitem - 1].name) > 1
-            messages << ErrorMessage.new("wronghstructure")
+            messages << ErrorMessage.new(:wrong_h_structure)
             break
           end  
         end
@@ -191,122 +252,125 @@ module Raakt
         
       end
       
-      return messages
+      messages
     end
 
     
+	# Verify that the document does not have any nested tabled. This is indicative of a table-based layout.
     def check_for_nested_tables
       
       messages = []  
-      tables = @soup.find_all("table")
+      tables = (@doc/"table")
       
       for table in tables
-        if table.find_all("table").length > 0
-          messages << ErrorMessage.new("hasnestedtables")
-          break
+        unless (table/"table").empty?
+          return messages << ErrorMessage.new(:has_nested_tables)
         end
       end
       
-      return messages
+      messages
     end
 
     
+	# Verify that all tables have at least on table header (th) element.
     def check_tables
-      
       messages = []  
-      tables = @soup.find_all("table")   
-      hasth = false
+      tables = (@doc/"table")   
       currenttable = 1
       
       for table in tables     
-        if table.thead
-          if table.thead.tr
-            if table.thead.tr.th   
-              hasth = true
-            end
-          end
-        end
-        
-        if table.tr
-          if table.tr.th
-            hasth = true
-          end
-        end
-        
-        unless hasth
-          messages << ErrorMessage.new("missingth", currenttable.to_s)
-        end
+      	hasth = false
+        hasth = true unless (table/">tr>th").empty?
+        hasth = true unless (table/">thead>tr>th").empty?
         
+        messages << ErrorMessage.new(:missing_th, currenttable.to_s) unless hasth
+                
         currenttable += 1
       end
       
-      return messages
+      messages
     end
     
     
+
+	# Verify that no formatting elements have been used. See UWEM 1.0 Test 7.2_HTML_01 and Test 7.3_HTML_01.
     def check_for_formatting_elements
       
-      messages = []  
-      formatting_items = @soup.find_all(%w(font b i u tt small big strike s))
-      flicker_items = @soup.find_all(["blink", "marquee"])
-      
-      formatting_items.each do |element|
-        messages << ErrorMessage.new("missingsemantics", "<#{element.name}>")  
-      end
+      	messages = []
 
-      if flicker_items.length > 0
-          messages << ErrorMessage.new("hasflicker")  
-      end
-            
-      return messages   
+	  	formatting_elements = %w(font b i u tt small big strike s)
+		formatting_elements = %w(font u tt small big strike s) if @ignore_bi
+	  
+	    formatting_items = (@doc/formatting_elements.join('|'))
+      
+      	unless formatting_items.empty?
+        	messages << ErrorMessage.new(:missing_semantics, "#{formatting_items.join(', ')}")  
+	    end
+	  
+	    flicker_elements = %w(blink marquee)
+	    flicker_items = (@doc/flicker_elements.join('|'))
+	    
+      	unless flicker_items.empty?
+        	messages << ErrorMessage.new(:has_flicker)  
+      	end
+
+      	messages   
     end
     
     
+	# Verify that the root documet html element as a lang attribute.
     def check_for_language_info
-      messages = []
-      
-      htmlelement = @soup.find("html")
-  
-      lang = langinfo(htmlelement) || ""
-      
-      unless lang.length > 1
-        messages << ErrorMessage.new("missinglanginfo")
-      end
-      
-      return messages
+      messages = []  
+	  unless (@doc/'html[@lang]').empty?
+	  	lang_code = (@doc/"html").first["lang"].to_s
+	  	if lang_code.length < 2
+      		messages << ErrorMessage.new(:missing_lang_info) 
+	  	end
+	  else
+      	messages << ErrorMessage.new(:missing_lang_info) 
+	  end
+	  messages
     end
+
+
+	# Verify that the html element has a valid lang code.
+	def check_valid_language_code
+	  messages = []
+	  unless (@doc/"html[@lang]").empty?
+		#load list of valid language codes
+		iso_lang_codes = []
+		IO.foreach(File.dirname(__FILE__) + "/iso_language_codes.txt") { |code| iso_lang_codes << code.chomp }
+
+		doc_main_lang_code = (@doc/"html").first["lang"].to_s.downcase
+		unless iso_lang_codes.include?(doc_main_lang_code[0..1])
+			messages << ErrorMessage.new(:wrong_lang_code, doc_main_lang_code)
+		end
+	  end
+
+	  messages
+	end
     
     
+	# Verify that no link texts are ambiguous. A typical example is the presence of multiple "Read more" links.
     def check_link_text
-      messages = []
       links = get_links
-      linktexts = links.collect { |el| el[3] }
       
-      for link_a in links
-        #compare to other links in collection
-        for link_b in links
-          if link_a[0] != link_b[0]
-            if is_ambiguous_link(link_a, link_b)
-              #add message if not added already for link text
-              unless find_errormsg_with_text(messages, link_a[3])
-                messages << ErrorMessage.new("ambiguouslinktext", link_a[3])
-              end
-            end
-          end
-        end
+      link = links.find do |link|
+        links.find { |cmp_link| is_ambiguous_link(link, cmp_link) }
       end
       
-      return messages      
+      return [] unless link
+      [ErrorMessage.new(:ambiguous_link_text, get_link_text(link))]
     end
         
     
+	# Verify that all form fields have a corresponding label element. See UWEM 1.0 Test 12.4_HTML_02.
     def check_form
       messages = []
       labels = get_labels
       fields = get_editable_fields
       
       #make sure all fields have associated labels
-      
       label_for_ids = []
       for label in labels
         if label["for"]
@@ -320,94 +384,80 @@ module Raakt
         field_id = (field["id"] || "")
         field_identifier = (field["id"] || field["name"] || "unknown")
         if not label_for_ids.include?(field_id)
-          messages << ErrorMessage.new("fieldmissinglabel", field_identifier)
+          messages << ErrorMessage.new(:field_missing_label, field_identifier)
         end
       end   
       
-      return messages
+      messages
     end
     
     
+	# Verify that all frame elements have a title atribute.
     def check_frames
-      #Verify frame titles
-      
-      messages = []  
-      if is_frameset
-        frames = @soup.find_all("frame")
-        frame_title = ""
+	  # Covers UWEM Test 12.1_HTML_01
+      return [] unless is_frameset
       
-        for frame in frames
-          frame_title = frame["title"] || ""
-          if normalize_text(frame_title).length == 0
-            messages << ErrorMessage.new("missingframetitle", frame["src"])
-          end
-        end
-      end
-      
-      return messages            
+      (@doc/"frame").find_all do |frame|
+        frame_title = frame['title'] || ''
+        normalize_text(frame_title).empty?
+      end.map { |frame| ErrorMessage.new(:missing_frame_title, frame['src']) }            
     end
     
     
+	# Verify that the document does not use meta-refresh to redirect the user away after a period of time.
     def check_refresh
+      meta_elements = (@doc/'meta')
       
-      messages = []
-      meta_elements = @soup.find_all("meta")
-      
-      for element in meta_elements
-        if element["http-equiv"] == "refresh"
-          messages << ErrorMessage.new("hasmetarefresh")
-        end
-      end
-      
-      return messages
+      meta_elements.find_all do |element|
+        element["http-equiv"] == "refresh"
+      end.map { ErrorMessage.new(:has_meta_refresh) }
     end
     
     
     #Utility methods    
-    
-    def is_ambiguous_link(link_a, link_b)
-      #Link A and B are ambiguous if:
-      #1. The url differs
-      #2. The link text is identical
-      #3. The title text is identical (if present)
-      if link_a[1] != link_b[1] and 
-         normalize_text(link_a[2]) == normalize_text(link_b[2]) and
-         normalize_text(link_a[3]) == normalize_text(link_b[3]) then
-         return true
+	
+    def headings
+      headings = []
+      1.upto(6) do |i| 
+        headings.push((@doc/"h#{i}")) if (@doc/"h#{i}").length > 0
       end
-      
-      return false
+      headings.flatten
     end
+
     
+    def level(heading)
+      Integer(heading[1].chr)
+    end
+
     
-    def find_errormsg_with_text(messages, text)
-      for errormessage in messages
-        if errormessage.note == text
-          return errormessage
-        end
-      end
+	def downcase_hash_keys(a_hash)
+		downcased_hash = {}
+		a_hash.collect {|key,value| downcased_hash[key.downcase] = value}
+		return downcased_hash
+	end
+
+	def parse_charset(contenttype)
+		# get charset identifier from content type string
+		if contenttype=~/charset=(.*)\w?/ then
+			return $1.downcase.strip
+		end
+
+		return ""
+	end
+
+
+    def is_ambiguous_link(link_a, link_b)
+      return false if links_point_to_same_resource?(link_a, link_b)
+      return true if link_text_identical?(link_a, link_b) &&
+                     link_title_identical?(link_a, link_b)
       
-      return nil
+      false
     end
     
-    
     def get_links      
-      linkelements = @soup.find_all("a")
-      links = []
-      currentlink = 0
-      
-      for element in linkelements
-        title = normalize_text((element['title'] || "").strip)
-        linktext = normalize_text((elements_to_text(element) || "").strip)
-        url = element['href']
-        links << [currentlink, url, title, linktext]
-        currentlink += 1
-      end      
-      
-      return links
+      (@doc/'a')
     end
 
-
     def langinfo(element)
       langval = ""
       
@@ -423,39 +473,33 @@ module Raakt
     end    
     
     
-    def img_to_text(imgtag)
-      return (imgtag['alt'] || "")
+    def alt_to_text(element)
+		if element.kind_of?(Hpricot::Elem) then
+      		element.has_attribute?("alt") ? element['alt'] : ""
+		else
+			""
+		end
     end
-    
-    
+
     def elements_to_text(element)
-      retval = ""
-      
-      for el in element.contents
-        if el.class.to_s == 'NavigableString'
-          retval += el
-        else
-          if el.name == "img"
-            retval += img_to_text(el)
-          else
-            retval += elements_to_text(el)
-          end
-        end        
+      str = ''
+      element.traverse_all_element do |elem|
+        elem.kind_of?(Hpricot::Text) ? str += "#{elem}" : str += alt_to_text(elem)
       end
       
-      return retval
+      str
     end
     
     
     def normalize_text(text)
-      text = (text || "")
-      retval = text.gsub(/&nbsp;/, " ")
-      retval = retval.gsub(/&#160;/, " ")
-      retval = retval.gsub(/\n/, "")
-      retval = retval.gsub(/\r/, "")
-      retval = retval.gsub(/\t/, "")
+      text ||= ''
+      retval = text.gsub(/&nbsp;/, ' ')
+      retval = retval.gsub(/&#160;/, ' ')
+      retval = retval.gsub(/\n/, '')
+      retval = retval.gsub(/\r/, '')
+      retval = retval.gsub(/\t/, '')
       while /  /.match(retval) do
-        retval = retval.gsub(/  /, " ")
+        retval = retval.gsub(/  /, ' ')
       end
       
       retval = retval.strip
@@ -465,12 +509,12 @@ module Raakt
     
     
     def get_labels
-      return @soup.find_all("label")
+      @doc/'label'
     end
 
 
     def get_editable_fields
-      allfields = @soup.find_all(["textarea", "select", "input"])
+      allfields = (@doc/"textarea|select|input")
       fields = []
       field_type = ""
       
@@ -487,9 +531,37 @@ module Raakt
         
         
     def is_frameset
-      return (@soup.find("frameset") != nil)
+      (@doc/"frameset").length > 0
     end    
     
+    
+    def link_text_identical?(link_a, link_b)
+      get_link_text(link_a) == get_link_text(link_b)
+    end
+    
+    def link_title_identical?(link_a, link_b)
+      get_link_title(link_a) == get_link_title(link_b)
+    end
+    
+    def links_point_to_same_resource?(link_a, link_b)
+      (link_a == link_b) ||
+      (get_link_url(link_a) == get_link_url(link_b))
+    end
+    
+    def get_link_text(link)
+      text = (elements_to_text(link) || '').strip
+      normalize_text(text)
+    end
+    
+    def get_link_url(link)
+      link['href']
+    end
+    
+    def get_link_title(link)
+      text = (link['title'] || '').strip
+      normalize_text(text)
+    end
+
   end
 
-end
+end