mime.rb in html2doc-0.5.0

- old
+ new

@@ -1,127 +1,9 @@
 require "uuidtools"
 require "nokogiri"
 
 module Html2Doc
-  def self.process(result, filename, header_file, dir)
-    docxml = Nokogiri::XML(xhtml(result))
-    cleanup(docxml, dir)
-    define_head(docxml, dir, filename, header_file)
-    result = self.msword_fix(docxml.to_xml)
-    system "cp #{header_file} #{dir}/header.html" unless header_file.nil?
-    generate_filelist(filename, dir)
-    File.open("#{filename}.htm", "w") { |f| f.write(result) }
-    mime_package result, filename, dir
-  end
-
-  def self.cleanup(docxml, dir)
-    image_cleanup(docxml, dir)
-    msonormal(docxml)
-  end
-
-  # preserve HTML escapes
-  def self.xhtml(result)
-    unless /<!DOCTYPE html/.match? result
-      result.gsub!(/<\?xml version="1.0"\?>/, "")
-      result = "<!DOCTYPE html SYSTEM " +
-        "'http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd'>" + result
-    end
-    result
-  end
-
-  def self.msword_fix(r)
-    # brain damage in MSWord parser
-    r.gsub!(%r{<span style="mso-special-character:footnote"/>},
-            '<span style="mso-special-character:footnote"></span>')
-    r.gsub!(%r{(<a style="mso-comment-reference:[^>/]+)/>}, "\\1></a>")
-    r.gsub!(%r{<link rel="File-List"}, "<link rel=File-List")
-    r.gsub!(%r{<meta http-equiv="Content-Type"},
-            "<meta http-equiv=Content-Type")
-    r.gsub!(%r{&tab;|&amp;tab;},
-            '<span style="mso-tab-count:1">&#xA0; </span>')
-    r
-  end
-
-  def self.image_resize(orig_filename)
-    image_size = ImageSize.path(orig_filename).size
-    # max width for Word document is 400, max height is 680
-    if image_size[0] > 400
-      image_size[1] = (image_size[1] * 400 / image_size[0]).ceil
-      image_size[0] = 400
-    end
-    if image_size[1] > 680
-      image_size[0] = (image_size[0] * 680 / image_size[1]).ceil
-      image_size[1] = 680
-    end
-    image_size
-  end
-
-  def self.image_cleanup(docxml, dir)
-    docxml.xpath("//*[local-name() = 'img']").each do |i|
-      matched = /\.(?<suffix>\S+)$/.match i["src"]
-      uuid = UUIDTools::UUID.random_create.to_s
-      new_full_filename = File.join(dir, "#{uuid}.#{matched[:suffix]}")
-      # presupposes that the image source is local
-      system "cp #{i['src']} #{new_full_filename}"
-      i["width"], i["height"] = image_resize(i["src"])
-      i["src"] = new_full_filename
-    end
-    docxml
-  end
-
-  def self.define_head1(docxml, dir)
-    docxml.xpath("//*[local-name() = 'head']").each do |h|
-      h.children.first.add_previous_sibling <<~XML
-        <!--[if gte mso 9]>
-        <xml>
-        <w:WordDocument>
-        <w:View>Print</w:View>
-        <w:Zoom>100</w:Zoom>
-        <w:DoNotOptimizeForBrowser/>
-        </w:WordDocument>
-        </xml>
-        <![endif]-->
-        <meta http-equiv=Content-Type content="text/html; charset=utf-8"/>
-        <link rel="File-List" href="#{dir}/filelist.xml"/>
-      XML
-    end
-  end
-
-  def self.stylesheet(filename, header_filename)
-    fn = File.join(File.dirname(__FILE__), "wordstyle.css")
-    stylesheet = File.read(fn, encoding: "UTF-8")
-    if header_filename.nil?
-      stylesheet.gsub!(/\n[^\n]*FILENAME[^\n]*i\n/, "\n")
-    else
-      stylesheet.gsub!(/FILENAME/, filename)
-    end
-    xml = Nokogiri::XML("<style/>")
-    xml.children.first << Nokogiri::XML::Comment.new(xml, "\n#{stylesheet}\n")
-    xml.root.to_s
-  end
-
-  def self.define_head(docxml, dir, filename, header_file)
-    title = docxml.at("//*[local-name() = 'head']/*[local-name() = 'title']")
-    head = docxml.at("//*[local-name() = 'head']")
-    if title.nil?
-      head.children.first.add_previous_sibling stylesheet(filename, header_file)
-    else
-      title.add_next_sibling stylesheet(filename, header_file)
-    end
-    define_head1(docxml, dir)
-    namespace(docxml.root)
-  end
-
-  def self.namespace(root)
-    {
-      o: "urn:schemas-microsoft-com:office:office",
-      w: "urn:schemas-microsoft-com:office:word",
-      m: "http://schemas.microsoft.com/office/2004/12/omml",
-    }.each { |k, v| root.add_namespace_definition(k.to_s, v) }
-    root.add_namespace(nil, "http://www.w3.org/TR/REC-html40")
-  end
-
   def self.mime_preamble(boundary, filename, result)
     <<~"PREAMBLE"
     MIME-Version: 1.0
     Content-Type: multipart/related; boundary="#{boundary}"
 
@@ -168,30 +50,7 @@
       next if item == "." || item == ".." || /^\./.match(item)
       mhtml += mime_attachment(boundary, filename, item, dir)
     end
     mhtml += "--#{boundary}--"
     File.open("#{filename}.doc", "w") { |f| f.write mhtml }
-  end
-
-  def self.generate_filelist(filename, dir)
-    File.open(File.join(dir, "filelist.xml"), "w") do |f|
-      f.write(<<~"XML")
-        <xml xmlns:o="urn:schemas-microsoft-com:office:office">
-        <o:MainFile HRef="../#{filename}.htm"/>
-      XML
-      Dir.foreach(dir) do |item|
-        next if item == "." || item == ".." || /^\./.match(item)
-        f.write %{  <o:File HRef="#{item}"/>\n}
-      end
-      f.write("</xml>\n")
-    end
-  end
-
-  def self.msonormal(docxml)
-    docxml.xpath("//*[local-name() = 'p'][not(self::*[@class])]").each do |p|
-      p["class"] = "MsoNormal"
-    end
-    docxml.xpath("//*[local-name() = 'li'][not(self::*[@class])]").each do |p|
-      p["class"] = "MsoNormal"
-    end
   end
 end