converter.rb in asciidoctor-epub3-1.5.0.alpha.8

- old
+ new

@@ -1,6 +1,6 @@
-# encoding: UTF-8
+# encoding: utf-8
 require_relative 'spine_item_processor'
 require_relative 'font_icon_map'
 
 module Asciidoctor
 module Epub3
@@ -56,12 +56,10 @@
   RightAngleQuote = '&#x203a;'
   CalloutStartNum = %(\u2460)
 
   XmlElementRx = /<\/?.+?>/
   CharEntityRx = /&#(\d{2,6});/
-  NamedEntityRx = /&([A-Z]+);/
-  UppercaseTagRx = /<(\/)?([A-Z]+)>/
 
   FromHtmlSpecialCharsMap = {
     '&lt;' => '<',
     '&gt;' => '>',
     '&amp;' => '&'
@@ -98,26 +96,21 @@
   end
 
   def document node
     docid = node.id
 
-    if (doctitle = node.doctitle partition: true, sanitize: true, use_fallback: true).subtitle?
-      title = doctitle.main
-      title_upper = title.upcase
+    if (doctitle = node.doctitle partition: true, use_fallback: true).subtitle?
+      title = %(#{doctitle.main} )
       subtitle = doctitle.subtitle
     else
       # HACK until we get proper handling of title-only in CSS
-      title = title_upper = ''
+      title = ''
       subtitle = doctitle.combined
     end
 
-    doctitle_sanitized = doctitle.combined
+    doctitle_sanitized = (node.doctitle sanitize: true, use_fallback: true).to_s
     subtitle_formatted = subtitle.split.map {|w| %(<b>#{w}</b>) } * ' '
-    # FIXME use uppercase pcdata helper to make less fragile (see logic in Asciidoctor PDF)
-    subtitle_formatted_upper = subtitle_formatted.upcase
-        .gsub(UppercaseTagRx) { %(<#{$1}#{$2.downcase}>) }
-        .gsub(NamedEntityRx) { %(&#{$1.downcase};) }
 
     if (node.attr 'publication-type', 'book') == 'book'
       byline = nil
     else
       author = node.attr 'author'
@@ -169,11 +162,11 @@
 </head>
 <body>
 <section class="chapter" title="#{doctitle_sanitized.gsub '"', '&quot;'}" epub:type="chapter" id="#{docid}">
 #{icon_css_scoped}<header>
 <div class="chapter-header">
-#{byline}<h1 class="chapter-title">#{title_upper}#{subtitle ? %[ <small class="subtitle">#{subtitle_formatted_upper}</small>] : nil}</h1>
+#{byline}<h1 class="chapter-title">#{title}#{subtitle ? %[<small class="subtitle">#{subtitle_formatted}</small>] : ''}</h1>
 </div>
 </header>
 #{content})]
 
     if node.footnotes?
@@ -272,10 +265,11 @@
       content
     end
   end
 
   def admonition node
+    id_attr = node.id ? %( id="#{node.id}") : nil
     if node.title?
       title = node.title
       title_sanitized = xml_sanitize title
       title_attr = %( title="#{node.caption}: #{title_sanitized}")
       title_el = %(<h2>#{title}</h2>
@@ -292,21 +286,22 @@
     when 'note'
       'note'
     when 'important', 'warning', 'caution'
       'warning'
     end
-    %(<aside class="admonition #{type}"#{title_attr} epub:type="#{epub_type}">
+    %(<aside#{id_attr} class="admonition #{type}"#{title_attr} epub:type="#{epub_type}">
 #{title_el}<div class="content">
 #{convert_content node}
 </div>
 </aside>)
   end
 
   def example node
+    id_attr = node.id ? %( id="#{node.id}") : nil
     title_div = node.title? ? %(<div class="example-title">#{node.title}</div>
 ) : nil
-    %(<div class="example">
+    %(<div#{id_attr} class="example">
 #{title_div}<div class="example-content">
 #{convert_content node}
 </div>
 </div>)
   end
@@ -342,10 +337,13 @@
   def thematic_break node
     '<hr class="thematicbreak"/>'
   end
 
   def quote node
+    id_attr = %( id="#{node.id}") if node.id
+    class_attr = (role = node.role) ? %( class="blockquote #{role}") : ' class="blockquote"'
+
     footer_content = []
     if (attribution = node.attr 'attribution')
       footer_content << attribution
     end
 
@@ -361,18 +359,21 @@
     footer_tag = footer_content.empty? ? nil : %(
 <footer>~ #{footer_content * ' '}</footer>)
     content = (convert_content node).strip.
       sub(OpenParagraphTagRx, '<p><span class="open-quote">“</span>').
       sub(CloseParagraphTagRx, '<span class="close-quote">”</span></p>')
-    %(<div class="blockquote">
+    %(<div#{id_attr}#{class_attr}>
 <blockquote>
 #{content}#{footer_tag}
 </blockquote>
 </div>)
   end
 
   def verse node
+    id_attr = %( id="#{node.id}") if node.id
+    class_attr = (role = node.role) ? %( class="verse #{role}") : ' class="verse"'
+
     footer_content = []
     if (attribution = node.attr 'attribution')
       footer_content << attribution
     end
 
@@ -381,11 +382,11 @@
       footer_content << %(<cite title="#{citetitle_sanitized}">#{citetitle}</cite>)
     end
 
     footer_tag = footer_content.size > 0 ? %(
 <span class="attribution">~ #{footer_content * ', '}</span>) : nil
-    %(<div class="verse">
+    %(<div#{id_attr}#{class_attr}>
 <pre>#{node.content}#{footer_tag}</pre>
 </div>)
   end
 
   def sidebar node
@@ -393,13 +394,11 @@
     if node.title?
       classes << 'titled'
       title = node.title
       title_sanitized = xml_sanitize title
       title_attr = %( title="#{title_sanitized}")
-      # FIXME use uppercase pcdata helper to make less fragile (see logic in Asciidoctor PDF)
-      title_upper = title.upcase.gsub(NamedEntityRx) { %(&#{$1.downcase};) }
-      title_el = %(<h2>#{title_upper}</h2>
+      title_el = %(<h2>#{title}</h2>
 )
     else
       title_attr = nil
       title_el = nil
     end
@@ -416,18 +415,20 @@
     lines << %(<div class="content">)
     table_id_attr = node.id ? %( id="#{node.id}") : nil
     frame_class = {
       'all' => 'table-framed',
       'topbot' => 'table-framed-topbot',
-      'sides' => 'table-framed-sides'
+      'sides' => 'table-framed-sides',
+      'none' => ''
     }
     grid_class = {
       'all' => 'table-grid',
       'rows' => 'table-grid-rows',
-      'cols' => 'table-grid-cols'
+      'cols' => 'table-grid-cols',
+      'none' => ''
     }
-    table_classes = %W(table #{frame_class[(node.attr 'frame')] || frame_class['topbot']} #{grid_class[(node.attr 'grid')] || grid_class['rows']})
+    table_classes = %W(table #{frame_class[node.attr 'frame'] || frame_class['topbot']} #{grid_class[node.attr 'grid'] || grid_class['rows']})
     if (role = node.role)
       table_classes << role
     end
     table_class_attr = %( class="#{table_classes * ' '}")
     table_styles = []
@@ -623,20 +624,18 @@
   end
 
   def image node
     target = node.attr 'target'
     type = (::File.extname target)[1..-1]
+    id_attr = node.id ? %( id="#{node.id}") : ''
     img_attrs = [%(alt="#{node.attr 'alt'}")]
     case type
     when 'svg'
       img_attrs << %(style="width: #{node.attr 'scaledwidth', '100%'}")
       # TODO make this a convenience method on document
-      epub_properties = (node.document.attr 'epub-properties') || []
-      unless epub_properties.include? 'svg'
-        epub_properties << 'svg'
-        node.document.attributes['epub-properties'] = epub_properties
-      end
+      epub_properties = (node.document.attributes['epub-properties'] ||= [])
+      epub_properties << 'svg' unless epub_properties.include? 'svg'
     else
       if node.attr? 'scaledwidth'
         img_attrs << %(style="width: #{node.attr 'scaledwidth'}")
       end
     end
@@ -655,11 +654,11 @@
         # Aldiko doesn't not scale width to 100% by default
         img_attrs << %(width="100%")
       end
     end
 =end
-    %(<figure class="image">
+    %(<figure#{id_attr} class="image#{prepend_space node.role}">
 <div class="content">
 <img src="#{node.image_uri node.attr('target')}" #{img_attrs * ' '}/>
 </div>#{node.title? ? %[
 <figcaption>#{node.captioned_title}</figcaption>] : nil}
 </figure>)
@@ -690,24 +689,35 @@
           id_attr = %( id="xref--#{refdoc_id}")
         end
         id_attr = nil unless @xrefs_seen.add? refid
         refdoc = doc.references[:spine_items].find {|it| refdoc_id == (it.id || (it.attr 'docname')) }
         if refdoc
-          if (reftext = refdoc.references[:ids][refdoc_refid])
-            text ||= reftext
+          # QUESTION should we invoke xreftext for references in other documents?
+          if (refs = refdoc.references[:refs]) && ::Asciidoctor::Document === (ref = refs[refdoc_refid])
+            text ||= (ref.attr 'docreftext') || ref.doctitle
+          elsif (xreftext = refdoc.references[:ids][refdoc_refid])
+            text ||= xreftext
           else
             warn %(asciidoctor: WARNING: #{::File.basename(doc.attr 'docfile')}: invalid reference to unknown anchor in #{refdoc_id} chapter: #{refdoc_refid})
           end
         else
           warn %(asciidoctor: WARNING: #{::File.basename(doc.attr 'docfile')}: invalid reference to anchor in unknown chapter: #{refdoc_id})
         end
       else
         id_attr = (@xrefs_seen.add? refid) ? %( id="xref-#{refid}") : nil
-        if (reftext = doc.references[:ids][refid])
-          text ||= reftext
+        if (refs = doc.references[:refs])
+          if ::Asciidoctor::AbstractNode === (ref = refs[refid])
+            xreftext = text || ref.xreftext((@xrefstyle ||= (doc.attr 'xrefstyle')))
+          end
         else
-          # FIXME we get false negatives for reference to bibref
+          xreftext = doc.references[:ids][refid]
+        end
+
+        if xreftext
+          text ||= xreftext
+        else
+          # FIXME we get false negatives for reference to bibref when using Asciidoctor < 1.5.6
           warn %(asciidoctor: WARNING: #{::File.basename(doc.attr 'docfile')}: invalid reference to unknown local anchor (or valid bibref): #{refid})
         end
       end
       %(<a#{id_attr} href="#{target}" class="xref">#{text || "[#{refid}]"}</a>)
     when :ref
@@ -755,12 +765,20 @@
       i_classes << %(icon-rotate-#{node.attr 'rotate'}) if node.attr? 'rotate'
       i_classes << node.role if node.role?
       %(<i class="#{i_classes * ' '}"></i>)
     else
       target = node.image_uri node.target
-      class_attr = %( class="#{node.role}") if node.role?
-      %(<img src="#{target}" alt="#{node.attr 'alt'}"#{class_attr}/>)
+      img_attrs = [%(alt="#{node.attr 'alt'}"), %(class="inline#{node.role? ? " #{node.role}" : ''}")]
+      if target.end_with? '.svg'
+        img_attrs << %(style="width: #{node.attr 'scaledwidth', '100%'}")
+        # TODO make this a convenience method on document
+        epub_properties = (node.document.attributes['epub-properties'] ||= [])
+        epub_properties << 'svg' unless epub_properties.include? 'svg'
+      elsif node.attr? 'scaledwidth'
+        img_attrs << %(style="width: #{node.attr 'scaledwidth'}")
+      end
+      %(<img src="#{target}" #{img_attrs * ' '}/>)
     end
   end
 
   def inline_indexterm node
     node.type == :visible ? node.text : ''
@@ -837,14 +855,19 @@
     if last_block.context == :paragraph
       last_block.attributes['role'] = last_block.role? ? %(#{last_block.role} last) : 'last'
     end
     nil
   end
+
+  # Prepend a space to the value if it's non-nil, otherwise return empty string.
+  def prepend_space value
+    value ? %( #{value}) : ''
+  end
 end
 
 class DocumentIdGenerator
   ReservedIds = %w(cover nav ncx)
-  CharRefRx = /&(?:([a-zA-Z]{2,})|#(\d{2,6})|#x([a-fA-F0-9]{2,5}));/
+  CharRefRx = /&(?:([a-zA-Z][a-zA-Z]+\d{0,2})|#(\d\d\d{0,4})|#x([\da-fA-F][\da-fA-F][\da-fA-F]{0,3}));/
   if defined? __dir__
     InvalidIdCharsRx = /[^\p{Word}]+/
     LeadingDigitRx = /^\p{Nd}/
   else
     InvalidIdCharsRx = /[^[:word:]]+/