microformats 0.3 → 4.0.0

data/lib/microformats/parser.rb

@@ -0,0 +1,105 @@
+module Microformats
+
+  class Parser < ParserCore
+    attr_reader :http_headers, :http_body
+    def initialize
+      @http_headers = {}
+      super
+    end
+
+    def parse(html, base: nil, headers:{})
+
+      @http_headers = {}
+
+      @items = []
+      @rels = {}
+      @rel_urls = {}
+
+      @alternates = []
+
+      @base = base
+
+      html = read_html(html, headers: headers)
+      document = Nokogiri::HTML(html)
+
+      found_base = parse_base(document)
+      @base = found_base unless found_base.nil?
+
+      document.traverse do |node|
+        if not node.attribute('src').nil?
+          absolute_url = Microformats::AbsoluteUri.new(node.attribute('src').value.to_s, base: @base).absolutize
+          node.attribute('src').value = absolute_url
+
+        elsif not node.attribute('href').nil?
+          absolute_url = Microformats::AbsoluteUri.new(node.attribute('href').value.to_s, base: @base).absolutize
+          node.attribute('href').value = absolute_url
+        end
+      end
+      parse_node(document)
+      parse_rels(document)
+
+      Collection.new({'items' => @items, 'rels' => @rels, 'rel-urls' =>  @rel_urls})
+    end
+
+    def read_html(html, headers:{})
+      open(html, headers) do |response|
+        @http_headers = response.meta if response.respond_to?(:meta)
+        @http_body = response.read
+      end
+      @http_body
+    rescue Errno::ENOENT, Errno::ENAMETOOLONG => e
+      @http_body = html
+    end
+
+    private
+
+    def parse_element(element)
+
+      fmt_classes = format_classes(element)
+      bc_fmt_classes = backcompat_format_classes(element).reject do |format_class|
+        fmt_classes.include? format_class
+      end
+      joined_classes =  fmt_classes + bc_fmt_classes
+
+      if bc_fmt_classes.length >= 1
+        @items << FormatParser.new.parse(element, base: @base, format_class_array: joined_classes, backcompat: true)
+      elsif fmt_classes.length >= 1
+        @items << FormatParser.new.parse(element, base: @base, format_class_array: fmt_classes )
+      else
+        parse_nodeset(element.children)
+      end
+    end
+
+    def parse_base(document)
+      base = document.search('base').first
+      base.values.first unless base.nil?
+    end
+
+    def parse_rels(element)
+      element.search('*[@rel]').each do |rel|
+        unless rel.attribute('href').nil?
+          url = Microformats::AbsoluteUri.new(rel.attribute('href').text, base: @base).absolutize
+
+          rel_values = rel.attribute('rel').text.split(' ')
+          rel_values.each do |rel_value|
+            @rels[rel_value] = [] unless @rels.has_key?(rel_value)
+            @rels[rel_value] << Microformats::AbsoluteUri.new(rel.attribute('href').text, base: @base).absolutize
+            @rels[rel_value].uniq!
+          end
+
+          @rel_urls[url] = {} unless @rel_urls.has_key?(url)
+
+          @rel_urls[url]['hreflang'] = rel.attribute('hreflang').value if @rel_urls[url]['hreflang'].nil? and not rel.attribute('hreflang').nil?
+          @rel_urls[url]['media'] = rel.attribute('media').value if @rel_urls[url]['media'].nil? and not rel.attribute('media').nil?
+          @rel_urls[url]['title'] = rel.attribute('title').value if @rel_urls[url]['title'].nil? and not rel.attribute('title').nil?
+          @rel_urls[url]['type'] = rel.attribute('type').value if @rel_urls[url]['type'].nil? and not rel.attribute('type').nil?
+          @rel_urls[url]['text'] = rel.text.strip if  @rel_urls[url]['text'].nil? and not rel.text.empty?
+          @rel_urls[url]['rels'] = rel_values
+
+        end
+      end
+    end
+
+  end
+end
+

data/lib/microformats/parser_core.rb

@@ -0,0 +1,343 @@
+module Microformats
+  class ParserCore
+
+    VALUE_CLASS_REG_EXP = /^value$/
+    VALUE_TITLE_CLASS_REG_EXP = /^value-title$/
+    FORMAT_CLASS_REG_EXP = /^h-[a-z-]+$/
+    PROPERTY_CLASS_REG_EXP = /^(p-|u-|dt-|e-)[a-z-]+$/
+
+
+    def initialize
+      @mode_backcompat = false
+      @fmt_classes = []
+    end
+
+    private
+
+    def parse_node(node)
+      if node.is_a?(Nokogiri::HTML::Document) then
+        parse_node(node.children)
+      elsif node.is_a?(Nokogiri::XML::NodeSet) then
+        parse_nodeset(node)
+      elsif node.is_a?(Nokogiri::XML::Element) then
+        parse_element(node)
+      else
+        nil
+      end
+    end
+
+
+    def parse_nodeset(nodeset)
+      nodeset.each do |node|
+        parse_node(node)
+      end
+    end
+
+    def format_classes(element)
+      element.attribute('class').to_s.split.select do |html_class|
+        html_class =~ FORMAT_CLASS_REG_EXP
+      end
+    end
+
+    def backcompat_format_classes(element)
+      result_set = []
+      html_classes = element.attribute('class').to_s.split
+
+      html_classes.each do |html_class|
+        case html_class
+
+        when  /^adr$/
+          result_set << 'h-adr'
+        when  /^geo$/
+          result_set << 'h-geo'
+        when  /^h[eE]ntry$/
+          result_set << 'h-entry'
+        when  /^h[pP]roduct$/
+          result_set << 'h-product'
+        when  /^h[rR]ecipe$/
+          result_set << 'h-recipe'
+        when  /^h[rR]esume$/
+          result_set << 'h-resume'
+        when  /^h[rR]eview$/
+          result_set << 'h-review'
+        when  /^h[rR]eview-aggregate$/
+          result_set << 'h-review-aggregate'
+        when  /^[vh][eE]vent$/
+          result_set << 'h-event'
+        when  /^[vh][cC]ard$/
+          result_set << 'h-card'
+
+        #these aren't actually specified by the backcompat faq, but probably should parse them
+        when  /^h[fF]eed$/
+          result_set << 'h-feed'
+        when  /^h[nN]ews$/
+          result_set << 'h-news'
+
+        else
+          if @fmt_classes.include? 'h-entry' and html_class == 'author'
+            result_set << 'h-card'
+          end
+          if @fmt_classes.include? 'h-product' and html_class == 'review'
+            result_set << 'h-review'
+          end
+          if @fmt_classes.include? 'h-recipe' and html_class == 'author'
+            result_set << 'h-card'
+          end
+          if @fmt_classes.include? 'h-resume' and html_class == 'contact'
+            result_set << 'h-card'
+          end
+          if @fmt_classes.include? 'h-resume' and html_class == 'education'
+            result_set << 'h-event'
+          end
+          if @fmt_classes.include? 'h-resume' and html_class == 'experience'
+            result_set << 'h-event'
+          end
+          if @fmt_classes.include? 'h-resume' and html_class == 'affiliation'
+            result_set << 'h-card'
+          end
+          if @fmt_classes.include? 'h-review' and html_class == 'reviewer'
+            result_set << 'h-card'
+          end
+          if @fmt_classes.include? 'h-review' and html_class == 'item'
+            if not html_classes.include? 'vcard' and not html_classes.include? 'vevent' and not html_classes.include? 'hproduct'
+              result_set << 'h-item'
+            end
+          end
+          if @fmt_classes.include? 'h-review-aggregate' and html_class == 'item'
+            if not html_classes.include? 'vcard' and not html_classes.include? 'vevent' and not html_classes.include? 'hproduct'
+              result_set << 'h-item'
+            end
+          end
+          if @fmt_classes.include? 'h-review-aggregate' and html_class == 'reviewer'
+            result_set << 'h-card'
+          end
+          if @fmt_classes.include? 'h-entry' and html_class == 'location'
+            result_set << 'h-adr'
+            result_set << 'h-card'
+          end
+          if @fmt_classes.include? 'h-feed' and html_class == 'author'
+            result_set << 'h-card'
+          end
+        end
+
+      end
+      result_set.uniq
+    end
+
+    def property_classes(element)
+      element.attribute('class').to_s.split.select do |html_class|
+        html_class =~ PROPERTY_CLASS_REG_EXP
+      end
+    end
+
+    def backcompat_property_classes(element)
+      result_set = []
+      rels = element.attribute('rel').to_s.split
+
+      if @fmt_classes.include? 'h-entry' and rels.include? 'bookmark'
+        result_set << 'u-url'
+      end
+      if @fmt_classes.include? 'h-entry' and rels.include? 'tag'
+        result_set << 'p-category'
+      end
+      if @fmt_classes.include? 'h-recipe' and rels.include? 'tag'
+        result_set << 'p-category'
+      end
+      if @fmt_classes.include? 'h-review' and rels.include? 'tag'
+        result_set << 'p-category'
+      end
+      if @fmt_classes.include? 'h-feed' and rels.include? 'tag'
+        result_set << 'p-category'
+      end
+      if @fmt_classes.include? 'h-review' and rels.include? 'self' and rels.include? 'bookmark'
+        result_set << 'u-url'
+      end
+      if @fmt_classes.include? 'h-news' and rels.include? 'principles'
+        result_set << 'u-principles'
+      end
+
+      #TODO PROPOSED convert time.entry-date[datetime] to dt-published  see wiki/h-entry
+      #TODO PROPOSED convert rel=author to u-author  see wiki/h-entry
+
+      element.attribute('class').to_s.split.each do |html_class|
+
+        if @fmt_classes.include? 'h-adr'
+          if [ 'post-office-box', 'extended-address', 'street-address', 'locality', 'region', 'postal-code', 'country-name' ].include? html_class
+            result_set << 'p-' + html_class
+          end
+        end
+
+        if @fmt_classes.include? 'h-geo'
+          if [ 'longitude', 'latitude' ].include? html_class
+            result_set << 'p-' + html_class
+          end
+        end
+
+        if @fmt_classes.include? 'h-entry'
+          if html_class ==  'entry-title'
+            result_set << 'p-name'
+          elsif html_class ==  'entry-summary'
+            result_set << 'p-summary'
+          elsif html_class ==  'entry-content'
+            result_set << 'e-content'
+          elsif ['updated', 'published'].include? html_class
+            result_set << 'dt-' + html_class
+          elsif [ 'category', 'author' ].include? html_class
+            result_set << 'p-' + html_class
+          end
+        end
+
+        #h-news isn't even listed in backcompat list, adding to follow test suite
+        if @fmt_classes.include? 'h-news'
+          if [ 'source-org', 'entry', 'dateline', 'geo' ].include? html_class
+            result_set << 'p-' + html_class
+          end
+        end
+
+        if @fmt_classes.include? 'h-feed'
+          if html_class ==  'fn'
+            result_set << 'p-name'
+          elsif [ 'author', 'summary' ].include? html_class
+            result_set << 'p-' + html_class
+          elsif ['photo', 'url'].include? html_class
+            result_set << 'u-' + html_class
+          end
+        end
+
+        if @fmt_classes.include? 'h-item'
+          if html_class ==  'fn'
+            result_set << 'p-name'
+          elsif ['photo', 'url'].include? html_class
+            result_set << 'u-' + html_class
+          end
+        end
+
+        if @fmt_classes.include? 'h-product'
+          if html_class ==  'fn'
+            result_set << 'p-name'
+          elsif ['photo', 'url', 'identifier'].include? html_class
+            result_set << 'u-' + html_class
+          elsif [ 'brand', 'category', 'price', 'review', 'description' ].include? html_class
+            result_set << 'p-' + html_class
+          end
+        end
+
+        if @fmt_classes.include? 'h-recipe'
+          if html_class ==  'fn'
+            result_set << 'p-name'
+          elsif html_class ==  'instructions'
+            result_set << 'e-instructions'
+          elsif html_class ==  'duration'
+            result_set << 'dt-duration'
+          elsif html_class ==  'photo'
+            result_set << 'u-photo'
+          elsif [ 'ingredient', 'category', 'yield', 'summary', 'nutrition', 'author' ].include? html_class
+            result_set << 'p-' + html_class
+          end
+        end
+
+        if @fmt_classes.include? 'h-resume'
+          if [ 'skill', 'summary', 'contact', 'education', 'experience', 'affiliation' ].include? html_class
+            result_set << 'p-' + html_class
+          end
+        end
+
+        if @fmt_classes.include? 'h-review'
+          if html_class ==  'summary'
+            result_set << 'p-name'
+          elsif html_class ==  'fn'
+            result_set << 'p-name'
+          elsif html_class ==  'reviewer'
+            result_set << 'p-author'
+          elsif html_class ==  'dtreviewed'
+            result_set << 'dt-published'
+          elsif html_class ==  'description'
+            result_set << 'e-content'
+          elsif ['photo', 'url', 'identifier'].include? html_class
+            result_set << 'u-' + html_class
+          elsif [ 'rating', 'best', 'worst', 'item'].include? html_class
+            result_set << 'p-' + html_class
+          end
+        end
+
+        if @fmt_classes.include? 'h-review-aggregate'
+          if html_class ==  'summary'
+            result_set << 'p-name'
+          elsif html_class ==  'fn'
+            result_set << 'p-name'
+          elsif html_class ==  'reviewer'
+            result_set << 'p-author'
+          elsif html_class ==  'dtreviewed'
+            result_set << 'dt-published'
+          elsif html_class ==  'description'
+            result_set << 'e-content'
+          elsif ['photo', 'url', 'identifier'].include? html_class
+            result_set << 'u-' + html_class
+          elsif [ 'rating', 'best', 'worst', 'item', 'count', 'votes', 'average' ].include? html_class
+            result_set << 'p-' + html_class
+          end
+        end
+
+        if @fmt_classes.include? 'h-event'
+          if html_class ==  'summary'
+            result_set << 'p-name'
+          elsif html_class ==  'dtstart'
+            result_set << 'dt-start'
+          elsif html_class ==  'dtend'
+            result_set << 'dt-end'
+          elsif html_class ==  'duration'
+            result_set << 'dt-duration'
+          elsif html_class ==  'geo'
+            result_set << 'p-location'
+          elsif ['url'].include? html_class
+            result_set << 'u-' + html_class
+          elsif [ 'description', 'category', 'location', 'attendee'].include? html_class
+            result_set << 'p-' + html_class
+          end
+        end
+
+        if @fmt_classes.include? 'h-card'
+          if html_class ==  'fn'
+            result_set << 'p-name'
+          elsif html_class ==  'bday'
+            result_set << 'dt-bday'
+          elsif html_class ==  'title'
+            result_set << 'p-job-title'
+          elsif html_class ==  'rev'
+            result_set << 'dt-rev'
+          elsif ['email', 'logo', 'photo', 'url', 'uid', 'key'].include? html_class
+            result_set << 'u-' + html_class
+          elsif [ 'honorific-prefix', 'given-name', 'additional-name', 'family-name', 'honorific-suffix', 'nickname',
+                  'category', 'adr', 'extended-address', 'street-address', 'locality', 'region', 'postal-code', 'country-name',
+                  'label', 'geo', 'latitude', 'longitude', 'tel', 'note', 'org', 'organization-name', 'organization-unit', 'role', 'tz' ].include? html_class
+            result_set << 'p-' + html_class
+
+          #these aren't listed in the wiki, may be removed
+          elsif ['sound'].include? html_class
+            result_set << 'u-' + html_class
+          #these aren't listed in the wiki, may be removed
+          elsif ['agent', 'mailer', 'sort-string', 'class'].include? html_class
+            result_set << 'p-' + html_class
+
+          end
+        end
+
+      end
+      result_set.uniq
+    end
+
+    def value_classes(element)
+      element.attribute('class').to_s.split.select do |html_class|
+        html_class =~ VALUE_CLASS_REG_EXP
+      end
+    end
+    def value_title_classes(element)
+      element.attribute('class').to_s.split.select do |html_class|
+        html_class =~ VALUE_TITLE_CLASS_REG_EXP
+      end
+    end
+
+  end
+
+end
+

data/lib/microformats/property_parser.rb

@@ -0,0 +1,122 @@
+module Microformats
+  class PropertyParser < ParserCore
+
+    def parse(element, base: nil, element_type: , format_class_array: [], backcompat: nil)
+      @base = base
+      @value = nil
+      @property_type = element_type
+
+      @fmt_classes = format_class_array
+      @mode_backcompat = backcompat
+
+      if element_type == 'p'
+        parse_value_class_pattern(element)
+
+        if @value.nil?
+          if element.name == 'abbr' and not element.attribute('title').nil?
+            @value = element.attribute('title').value.strip
+          elsif (element.name == 'data' or element.name == 'input') and not element.attribute('value').nil?
+            @value = element.attribute('value').value.strip
+          elsif (element.name == 'img' or element.name == 'area') and not element.attribute('alt').nil?
+            @value = element.attribute('alt').value.strip
+          else
+            @value = render_text_and_replace_images(element, base: @base)
+          end
+        end
+
+      elsif element_type == 'e'
+        @value = {
+          value: render_text(element, base: @base), #TODO the spec doesn't say to remove script and style tags, assuming this to be in error
+          html: element.inner_html.gsub(/\A +/, '').gsub(/ +\Z/, '')
+        }
+
+      elsif element_type == 'u'
+        if ['a', 'area'].include? element.name and not element.attribute('href').nil?
+          @value = element.attribute('href').value.strip
+        elsif ['img', 'audio', 'video', 'source'].include? element.name and not element.attribute('src').nil?
+          @value = element.attribute('src').value.strip
+        elsif element.name == 'video' and not element.attribute('poster').nil?
+          @value = element.attribute('poster').value.strip
+        elsif element.name == 'object' and not element.attribute('data').nil?
+          @value = element.attribute('data').value.strip
+        end
+
+        if not @value.nil?
+          @value = Microformats::AbsoluteUri.new(@value, base: @base).absolutize
+        else
+
+          parse_value_class_pattern(element)
+
+          if @value.nil?
+            if element.name == 'abbr' and not element.attribute('title').nil?
+              @value = element.attribute('title').value.strip
+            elsif (element.name == 'data' or element.name == 'input') and not element.attribute('value').nil?
+              @value = element.attribute('value').value.strip
+            else
+              @value = render_text(element, base: @base)
+            end
+
+          end
+        end
+
+      elsif element_type == 'dt'
+        @value = Microformats::TimePropertyParser.new.parse(element, base: base, element_type: element_type, format_class_array: format_class_array, backcompat: backcompat)
+
+      end
+
+      @value
+    end
+
+    def parse_value_class_pattern(element)
+      @value_class_pattern_value = []
+      parse_node(element.children)
+      @value = @value_class_pattern_value.join unless @value_class_pattern_value.empty?
+    end
+
+    def parse_element(element)
+      if value_title_classes(element).length >= 1
+        @value_class_pattern_value << element.attribute('title').value.strip
+
+      elsif value_classes(element).length >= 1
+        if element.name == 'img' or element.name == 'area' and not element.attribute('alt').nil?
+          @value_class_pattern_value << element.attribute('alt').value.strip
+        elsif element.name == 'data' and not element.attribute('value').nil?
+          @value_class_pattern_value << element.attribute('value').value.strip
+        elsif element.name == 'abbr' and not element.attribute('title').nil?
+          @value_class_pattern_value << element.attribute('title').value.strip
+        else
+          @value_class_pattern_value << element.text.strip
+        end
+      else
+        p_classes = property_classes(element)
+        p_classes = backcompat_property_classes(element) if @mode_backcompat
+        if p_classes.length == 0 and format_classes(element).length == 0
+          parse_node(element.children)
+        end
+      end
+    end
+
+    def render_text_and_replace_images(node, base: nil)
+      new_doc = Nokogiri::HTML(node.inner_html)
+      new_doc.xpath('//script').remove
+      new_doc.xpath('//style').remove
+      new_doc.traverse do |node|
+        if node.name == 'img' and not node.attribute('alt').nil?
+          node.replace(' ' + node.attribute('alt').value.to_s + ' ')
+        elsif node.name == 'img' and not node.attribute('src').nil?
+          absolute_url = Microformats::AbsoluteUri.new(node.attribute('src').value.to_s, base: @base).absolutize
+          node.replace(' ' + absolute_url  + ' ')
+        end
+      end
+      new_doc.text.strip
+    end
+
+    def render_text(node, base: nil)
+      new_doc = Nokogiri::HTML(node.inner_html)
+      new_doc.xpath('//script').remove
+      new_doc.xpath('//style').remove
+      new_doc.text.strip
+    end
+
+  end
+end