html2doc 1.0.7 → 1.1.3

data/lib/html2doc/math.rb

@@ -9,23 +9,34 @@ module Html2Doc
     Nokogiri::XSLT(File.read(File.join(File.dirname(__FILE__), "mml2omml.xsl"),
                              encoding: "utf-8"))
 
-  def self.asciimath_to_mathml1(x)
-    AsciiMath::MathMLBuilder.new(:msword => true).append_expression(
-      AsciiMath.parse(HTMLEntities.new.decode(x)).ast).to_s.
-        gsub(/<math>/, "<math xmlns='http://www.w3.org/1998/Math/MathML'>")
+  def self.asciimath_to_mathml1(expr)
+    AsciiMath::MathMLBuilder.new(msword: true).append_expression(
+      AsciiMath.parse(HTMLEntities.new.decode(expr)).ast,
+    ).to_s
+      .gsub(/<math>/, "<math xmlns='http://www.w3.org/1998/Math/MathML'>")
+  rescue StandardError => e
+    puts "parsing: #{expr}"
+    puts e.message
+    raise e
   end
 
   def self.asciimath_to_mathml(doc, delims)
     return doc if delims.nil? || delims.size < 2
+
     m = doc.split(/(#{Regexp.escape(delims[0])}|#{Regexp.escape(delims[1])})/)
     m.each_slice(4).map.with_index do |(*a), i|
-      i % 500 == 0 && m.size > 1000 && i > 0 and
-        warn "MathML #{i} of #{(m.size / 4).floor}"
+      progress_conv(i, 500, (m.size / 4).floor, 1000, "AsciiMath")
       a[2].nil? || a[2] = asciimath_to_mathml1(a[2])
       a.size > 1 ? a[0] + a[2] : a[0]
     end.join
   end
 
+  def self.progress_conv(idx, step, total, threshold, msg)
+    return unless (idx % step).zero? && total > threshold && idx.positive?
+
+    warn "#{msg} #{idx} of #{total}"
+  end
+
   def self.unwrap_accents(doc)
     doc.xpath("//*[@accent = 'true']").each do |x|
       x.elements.length > 1 or next
@@ -36,106 +47,124 @@ module Html2Doc
   end
 
   # random fixes to MathML input that OOXML needs to render properly
-  def self.ooxml_cleanup(m, docnamespaces)
-    m = unwrap_accents(mathml_preserve_space(
-      mathml_insert_rows(m, docnamespaces), docnamespaces))
-    m.add_namespace(nil, "http://www.w3.org/1998/Math/MathML")
-    m
+  def self.ooxml_cleanup(math, docnamespaces)
+    math = unwrap_accents(
+      mathml_preserve_space(
+        mathml_insert_rows(math, docnamespaces), docnamespaces
+      ),
+    )
+    math.add_namespace(nil, "http://www.w3.org/1998/Math/MathML")
+    math
   end
 
-  def self.mathml_insert_rows(m, docnamespaces)
-    m.xpath(%w(msup msub msubsup munder mover munderover).
-            map { |m| ".//xmlns:#{m}" }.join(" | "), docnamespaces).each do |x|
+  def self.mathml_insert_rows(math, docnamespaces)
+    math.xpath(%w(msup msub msubsup munder mover munderover)
+            .map { |m| ".//xmlns:#{m}" }.join(" | "), docnamespaces).each do |x|
       next unless x.next_element && x.next_element != "mrow"
+
       x.next_element.wrap("<mrow/>")
     end
-    m
+    math
   end
 
-  def self.mathml_preserve_space(m, docnamespaces)
-    m.xpath(".//xmlns:mtext", docnamespaces).each do |x|
+  def self.mathml_preserve_space(math, docnamespaces)
+    math.xpath(".//xmlns:mtext", docnamespaces).each do |x|
       x.children = x.children.to_xml.gsub(/^\s/, "&#xA0;").gsub(/\s$/, "&#xA0;")
     end
-    m
+    math
   end
 
-  def self.unitalic(m)
-    m.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:scr)]/xmlns:sty[@m:val = 'p']]").each do |x|
-      x.wrap("<span style='font-style:normal;'></span>")
+  HTML_NS = 'xmlns="http://www.w3.org/1999/xhtml"'.freeze
+
+  def self.unitalic(math)
+    math.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:scr)]/xmlns:sty[@m:val = 'p']]").each do |x|
+      x.wrap("<span #{HTML_NS} style='font-style:normal;'></span>")
     end
-    m.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:scr)]/xmlns:sty[@m:val = 'bi']]").each do |x|
-      x.wrap("<span class='nostem' style='font-weight:bold;'><em></em></span>")
+    math.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:scr)]/xmlns:sty[@m:val = 'bi']]").each do |x|
+      x.wrap("<span #{HTML_NS} class='nostem' style='font-weight:bold;'><em></em></span>")
     end
-    m.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:scr)]/xmlns:sty[@m:val = 'i']]").each do |x|
-      x.wrap("<span class='nostem'><em></em></span>")
+    math.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:scr)]/xmlns:sty[@m:val = 'i']]").each do |x|
+      x.wrap("<span #{HTML_NS} class='nostem'><em></em></span>")
     end
-    m.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:scr)]/xmlns:sty[@m:val = 'b']]").each do |x|
-      x.wrap("<span style='font-style:normal;font-weight:bold;'></span>")
+    math.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:scr)]/xmlns:sty[@m:val = 'b']]").each do |x|
+      x.wrap("<span #{HTML_NS} style='font-style:normal;font-weight:bold;'></span>")
     end
-    m.xpath(".//xmlns:r[xmlns:rPr/xmlns:scr[@m:val = 'monospace']]").each do |x|
-      toPlane1(x, :monospace)
+    math.xpath(".//xmlns:r[xmlns:rPr/xmlns:scr[@m:val = 'monospace']]").each do |x|
+      to_plane1(x, :monospace)
     end
-    m.xpath(".//xmlns:r[xmlns:rPr/xmlns:scr[@m:val = 'double-struck']]").each do |x|
-      toPlane1(x, :doublestruck)
+    math.xpath(".//xmlns:r[xmlns:rPr/xmlns:scr[@m:val = 'double-struck']]").each do |x|
+      to_plane1(x, :doublestruck)
     end
-    m.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:sty) or xmlns:sty/@m:val = 'p']/xmlns:scr[@m:val = 'script']]").each do |x|
-      toPlane1(x, :script)
+    math.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:sty) or xmlns:sty/@m:val = 'p']/xmlns:scr[@m:val = 'script']]").each do |x|
+      to_plane1(x, :script)
     end
-    m.xpath(".//xmlns:r[xmlns:rPr[xmlns:sty/@m:val = 'b']/xmlns:scr[@m:val = 'script']]").each do |x|
-      toPlane1(x, :scriptbold)
+    math.xpath(".//xmlns:r[xmlns:rPr[xmlns:sty/@m:val = 'b']/xmlns:scr[@m:val = 'script']]").each do |x|
+      to_plane1(x, :scriptbold)
     end
-    m.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:sty) or xmlns:sty/@m:val = 'p']/xmlns:scr[@m:val = 'fraktur']]").each do |x|
-      toPlane1(x, :fraktur)
+    math.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:sty) or xmlns:sty/@m:val = 'p']/xmlns:scr[@m:val = 'fraktur']]").each do |x|
+      to_plane1(x, :fraktur)
     end
-    m.xpath(".//xmlns:r[xmlns:rPr[xmlns:sty/@m:val = 'b']/xmlns:scr[@m:val = 'fraktur']]").each do |x|
-      toPlane1(x, :frakturbold)
+    math.xpath(".//xmlns:r[xmlns:rPr[xmlns:sty/@m:val = 'b']/xmlns:scr[@m:val = 'fraktur']]").each do |x|
+      to_plane1(x, :frakturbold)
     end
-    m.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:sty) or xmlns:sty/@m:val = 'p']/xmlns:scr[@m:val = 'sans-serif']]").each do |x|
-      toPlane1(x, :sans)
+    math.xpath(".//xmlns:r[xmlns:rPr[not(xmlns:sty) or xmlns:sty/@m:val = 'p']/xmlns:scr[@m:val = 'sans-serif']]").each do |x|
+      to_plane1(x, :sans)
     end
-    m.xpath(".//xmlns:r[xmlns:rPr[xmlns:sty/@m:val = 'b']/xmlns:scr[@m:val = 'sans-serif']]").each do |x|
-      toPlane1(x, :sansbold)
+    math.xpath(".//xmlns:r[xmlns:rPr[xmlns:sty/@m:val = 'b']/xmlns:scr[@m:val = 'sans-serif']]").each do |x|
+      to_plane1(x, :sansbold)
     end
-    m.xpath(".//xmlns:r[xmlns:rPr[xmlns:sty/@m:val = 'i']/xmlns:scr[@m:val = 'sans-serif']]").each do |x|
-      toPlane1(x, :sansitalic)
+    math.xpath(".//xmlns:r[xmlns:rPr[xmlns:sty/@m:val = 'i']/xmlns:scr[@m:val = 'sans-serif']]").each do |x|
+      to_plane1(x, :sansitalic)
     end
-    m.xpath(".//xmlns:r[xmlns:rPr[xmlns:sty/@m:val = 'bi']/xmlns:scr[@m:val = 'sans-serif']]").each do |x|
-      toPlane1(x, :sansbolditalic)
+    math.xpath(".//xmlns:r[xmlns:rPr[xmlns:sty/@m:val = 'bi']/xmlns:scr[@m:val = 'sans-serif']]").each do |x|
+      to_plane1(x, :sansbolditalic)
     end
-    m
+    math
   end
 
-  def self.toPlane1(x, font)
-    x.traverse do |n|
+  def self.to_plane1(xml, font)
+    xml.traverse do |n|
       next unless n.text?
+
       n.replace(Plane1Converter.conv(HTMLEntities.new.decode(n.text), font))
     end
-    x
+    xml
   end
 
   def self.mathml_to_ooml(docxml)
     docnamespaces = docxml.collect_namespaces
     m = docxml.xpath("//*[local-name() = 'math']")
     m.each_with_index do |x, i|
-      i % 100 == 0 && m.size > 500 && i > 0 and
-        warn "Math OOXML #{i} of #{m.size}"
-      element = ooxml_cleanup(x, docnamespaces)
-      doc = Nokogiri::XML::Document::new()
-      doc.root = element
-      ooxml = (unitalic(esc_space(@xsltemplate.transform(doc)))).to_s.
-        gsub(/<\?[^>]+>\s*/, "").
-        gsub(/ xmlns(:[^=]+)?="[^"]+"/, "").
-        gsub(%r{<(/)?(?!span)(?!em)([a-z])}, "<\\1m:\\2")
-      ooxml = uncenter(x, ooxml)
-      x.swap(ooxml)
+      progress_conv(i, 100, m.size, 500, "Math OOXML")
+      mathml_to_ooml1(x, docnamespaces)
     end
   end
 
-  # escape space as &#x32;; we are removing any spaces generated by 
+  # We need span and em not to be namespaced. Word can't deal with explicit 
+  # namespaces.
+  # We will end up stripping them out again under Nokogiri 1.11, which correctly
+  # insists on inheriting namespace from parent.
+  def self.ooml_clean(xml)
+    xml.to_s
+      .gsub(/<\?[^>]+>\s*/, "")
+      .gsub(/ xmlns(:[^=]+)?="[^"]+"/, "")
+      .gsub(%r{<(/)?(?!span)(?!em)([a-z])}, "<\\1m:\\2")
+  end
+
+  def self.mathml_to_ooml1(xml, docnamespaces)
+    doc = Nokogiri::XML::Document::new
+    doc.root = ooxml_cleanup(xml, docnamespaces)
+      ooxml = ooml_clean(unitalic(esc_space(@xsltemplate.transform(doc))))
+    ooxml = uncenter(xml, ooxml)
+    xml.swap(ooxml)
+  end
+
+  # escape space as &#x32;; we are removing any spaces generated by
   # XML indentation
   def self.esc_space(xml)
     xml.traverse do |n|
       next unless n.text?
+
       n = n.text.gsub(/ /, "&#x32;")
     end
     xml
@@ -143,17 +172,15 @@ module Html2Doc
 
   # if oomml has no siblings, by default it is centered; override this with
   # left/right if parent is so tagged
-  def self.uncenter(m, ooxml)
-    if m.next == nil && m.previous == nil
-      alignnode = m.at(".//ancestor::*[@style][local-name() = 'p' or "\
-                       "local-name() = 'div' or local-name() = 'td']/@style")
-      return ooxml unless alignnode
-      if alignnode.text.include? ("text-align:left")
-        ooxml = "<m:oMathPara><m:oMathParaPr><m:jc "\
-          "m:val='left'/></m:oMathParaPr>#{ooxml}</m:oMathPara>"
-      elsif alignnode.text.include? ("text-align:right")
+  def self.uncenter(math, ooxml)
+    alignnode = math.at(".//ancestor::*[@style][local-name() = 'p' or "\
+                        "local-name() = 'div' or local-name() = 'td']/@style")
+    return ooxml unless alignnode && (math.next == nil && math.previous == nil)
+
+    %w(left right).each do |dir|
+      if alignnode.text.include? ("text-align:#{dir}")
         ooxml = "<m:oMathPara><m:oMathParaPr><m:jc "\
-          "m:val='right'/></m:oMathParaPr>#{ooxml}</m:oMathPara>"
+          "m:val='#{dir}'/></m:oMathParaPr>#{ooxml}</m:oMathPara>"
       end
     end
     ooxml

data/lib/html2doc/mime.rb

@@ -7,19 +7,20 @@ require "fileutils"
 module Html2Doc
   def self.mime_preamble(boundary, filename, result)
     <<~"PREAMBLE"
-    MIME-Version: 1.0
-    Content-Type: multipart/related; boundary="#{boundary}"
+      MIME-Version: 1.0
+      Content-Type: multipart/related; boundary="#{boundary}"
 
-    --#{boundary}
-    Content-Location: file:///C:/Doc/#{File.basename(filename)}.htm
-    Content-Type: text/html; charset="utf-8"
+      --#{boundary}
+      Content-ID: <#{File.basename(filename)}>
+      Content-Disposition: inline; filename="#{File.basename(filename)}"
+      Content-Type: text/html; charset="utf-8"
 
-    #{result}
+      #{result}
 
     PREAMBLE
   end
 
-  def self.mime_attachment(boundary, filename, item, dir)
+  def self.mime_attachment(boundary, _filename, item, dir)
     content_type = mime_type(item)
     text_mode = %w[text application].any? { |p| content_type.start_with? p }
 
@@ -28,12 +29,13 @@ module Html2Doc
 
     encoded_file = Base64.strict_encode64(content).gsub(/(.{76})/, "\\1\n")
     <<~"FILE"
-    --#{boundary}
-    Content-Location: file:///C:/Doc/#{File.basename(filename)}_files/#{item}
-    Content-Transfer-Encoding: base64
-    Content-Type: #{content_type}
+      --#{boundary}
+      Content-ID: <#{File.basename(item)}>
+      Content-Disposition: inline; filename="#{File.basename(item)}"
+      Content-Transfer-Encoding: base64
+      Content-Type: #{content_type}
 
-    #{encoded_file}
+      #{encoded_file}
 
     FILE
   end
@@ -41,7 +43,7 @@ module Html2Doc
   def self.mime_type(item)
     types = MIME::Types.type_for(item)
     type = types ? types.first.to_s : 'text/plain; charset="utf-8"'
-    type = type + ' charset="utf-8"' if /^text/.match(type) && types
+    type = %(#{type} charset="utf-8") if /^text/.match(type) && types
     type
   end
 
@@ -52,25 +54,37 @@ module Html2Doc
 
   def self.mime_package(result, filename, dir)
     boundary = mime_boundary
-    mhtml = mime_preamble(boundary, filename, result)
-    mhtml += mime_attachment(boundary, filename, "filelist.xml", dir)
+    mhtml = mime_preamble(boundary, "#{filename}.htm", result)
+    mhtml += mime_attachment(boundary, "#{filename}.htm", "filelist.xml", dir)
     Dir.foreach(dir) do |item|
       next if item == "." || item == ".." || /^\./.match(item) ||
         item == "filelist.xml"
-      mhtml += mime_attachment(boundary, filename, item, dir)
+
+      mhtml += mime_attachment(boundary, "#{filename}.htm", item, dir)
     end
     mhtml += "--#{boundary}--"
-    File.open("#{filename}.doc", "w:UTF-8") { |f| f.write mhtml }
+    File.open("#{filename}.doc", "w:UTF-8") { |f| f.write contentid(mhtml) }
+  end
+
+  def self.contentid(mhtml)
+    mhtml.gsub %r{(<img[^>]*?src=")([^\"']+)(['"])}m do |m|
+      repl = "#{$1}cid:#{File.basename($2)}#{$3}"
+      /^data:|^https?:/.match($2) ? m : repl
+    end.gsub %r{(<v:imagedata[^>]*?src=")([^\"']+)(['"])}m do |m|
+      repl = "#{$1}cid:#{File.basename($2)}#{$3}"
+      /^data:|^https?:/.match($2) ? m : repl
+    end
   end
 
   # max width for Word document is 400, max height is 680
-  def self.image_resize(i, path, maxheight, maxwidth)
-    realSize = ImageSize.path(path).size
-    s = [i["width"].to_i, i["height"].to_i]
-    s = realSize if s[0].zero? && s[1].zero?
-    return [nil, nil] if realSize.nil? || realSize[0].nil? || realSize[1].nil?
-    s[1] = s[0] * realSize[1] / realSize[0] if s[1].zero? && !s[0].zero?
-    s[0] = s[1] * realSize[0] / realSize[1] if s[0].zero? && !s[1].zero?
+  def self.image_resize(img, path, maxheight, maxwidth)
+    realsize = ImageSize.path(path).size
+    s = [img["width"].to_i, img["height"].to_i]
+    s = realsize if s[0].zero? && s[1].zero?
+    return [nil, nil] if realsize.nil? || realsize[0].nil? || realsize[1].nil?
+
+    s[1] = s[0] * realsize[1] / realsize[0] if s[1].zero? && !s[0].zero?
+    s[0] = s[1] * realsize[0] / realsize[1] if s[0].zero? && !s[1].zero?
     s = [(s[0] * maxheight / s[1]).ceil, maxheight] if s[1] > maxheight
     s = [maxwidth, (s[1] * maxwidth / s[0]).ceil] if s[0] > maxwidth
     s
@@ -83,19 +97,22 @@ module Html2Doc
   end
 
   def self.warnsvg(src)
-    warn "#{src}: SVG not supported" if /\.svg$/i.match(src)
+    warn "#{src}: SVG not supported" if /\.svg$/i.match?(src)
+  end
+
+  def self.localname(src, localdir)
+    %r{^([A-Z]:)?/}.match?(src) ? src : File.join(localdir, src)
   end
 
   # only processes locally stored images
   def self.image_cleanup(docxml, dir, localdir)
     docxml.traverse do |i|
       next unless i.element? && %w(img v:imagedata).include?(i.name)
-      #warnsvg(i["src"])
-      next if /^http/.match i["src"]
-      next if %r{^data:(image|application)/[^;]+;base64}.match i["src"]
-      local_filename = %r{^([A-Z]:)?/}.match(i["src"]) ? i["src"] :
-        File.join(localdir, i["src"])
-      new_filename = "#{mkuuid}#{File.extname(i["src"])}"
+      next if /^http/.match? i["src"]
+      next if %r{^data:(image|application)/[^;]+;base64}.match? i["src"]
+
+      local_filename = localname(i["src"], localdir)
+      new_filename = "#{mkuuid}#{File.extname(i['src'])}"
       FileUtils.cp local_filename, File.join(dir, new_filename)
       i["width"], i["height"] = image_resize(i, local_filename, 680, 400)
       i["src"] = File.join(File.basename(dir), new_filename)
@@ -103,7 +120,7 @@ module Html2Doc
     docxml
   end
 
-  # do not parse the header through Nokogiri, since it will contain 
+  # do not parse the header through Nokogiri, since it will contain
   # non-XML like <![if !supportFootnotes]>
   def self.header_image_cleanup(doc, dir, filename, localdir)
     doc.split(%r{(<img [^>]*>|<v:imagedata [^>]*>)}).each_slice(2).map do |a|
@@ -111,16 +128,14 @@ module Html2Doc
     end.join
   end
 
-  def self.header_image_cleanup1(a, dir, filename, localdir)
+  def self.header_image_cleanup1(a, dir, _filename, localdir)
     if a.size == 2 && !(/ src="https?:/.match a[1]) &&
         !(%r{ src="data:(image|application)/[^;]+;base64}.match a[1])
       m = / src=['"](?<src>[^"']+)['"]/.match a[1]
-      #warnsvg(m[:src])
       m2 = /\.(?<suffix>[a-zA-Z_0-9]+)$/.match m[:src]
       new_filename = "#{mkuuid}.#{m2[:suffix]}"
-      old_filename = %r{^([A-Z]:)?/}.match(m[:src]) ? m[:src] : File.join(localdir, m[:src])
-      FileUtils.cp old_filename, File.join(dir, new_filename)
-      a[1].sub!(%r{ src=['"](?<src>[^"']+)['"]}, " src='file:///C:/Doc/#{filename}_files/#{new_filename}'")
+      FileUtils.cp localname(m[:src], localdir), File.join(dir, new_filename)
+      a[1].sub!(%r{ src=['"](?<src>[^"']+)['"]}, " src='cid:#{new_filename}'")
     end
     a.join
   end
@@ -131,6 +146,7 @@ module Html2Doc
         <o:MainFile HRef="../#{filename}.htm"/>}
       Dir.entries(dir).sort.each do |item|
         next if item == "." || item == ".." || /^\./.match(item)
+
         f.write %{  <o:File HRef="#{item}"/>\n}
       end
       f.write("</xml>\n")

data/lib/html2doc/notes.rb

@@ -6,6 +6,7 @@ module Html2Doc
     fn = []
     docxml.xpath("//a").each do |a|
       next unless process_footnote_link(docxml, a, i, fn)
+
       i += 1
     end
     process_footnote_texts(docxml, fn)
@@ -22,13 +23,13 @@ module Html2Doc
     footnote_cleanup(docxml)
   end
 
-  def self.footnote_div_to_p(f)
-    if %w{div aside}.include? f.name
-      if f.at(".//p")
-        f.replace(f.children)
+  def self.footnote_div_to_p(elem)
+    if %w{div aside}.include? elem.name
+      if elem.at(".//p")
+        elem.replace(elem.children)
       else
-        f.name = "p"
-        f["class"] = "MsoFootnoteText"
+        elem.name = "p"
+        elem["class"] = "MsoFootnoteText"
       end
     end
   end
@@ -36,34 +37,39 @@ module Html2Doc
   FN = "<span class='MsoFootnoteReference'>"\
     "<span style='mso-special-character:footnote'/></span>".freeze
 
-  def self.footnote_container(docxml, i)
-    ref = docxml&.at("//a[@href='#_ftn#{i}']")&.children&.to_xml(indent: 0).
-      gsub(/>\n</, "><") || FN
+  def self.footnote_container(docxml, idx)
+    ref = docxml&.at("//a[@href='#_ftn#{idx}']")&.children&.to_xml(indent: 0)
+      &.gsub(/>\n</, "><") || FN
     <<~DIV
-      <div style='mso-element:footnote' id='ftn#{i}'>
-        <a style='mso-footnote-id:ftn#{i}' href='#_ftn#{i}'
-           name='_ftnref#{i}' title='' id='_ftnref#{i}'>#{ref.strip}</a></div>
+      <div style='mso-element:footnote' id='ftn#{idx}'>
+        <a style='mso-footnote-id:ftn#{idx}' href='#_ftn#{idx}'
+           name='_ftnref#{idx}' title='' id='_ftnref#{idx}'>#{ref.strip}</a></div>
     DIV
   end
 
-  def self.process_footnote_link(docxml, a, i, fn)
-    return false unless footnote?(a)
-    href = a["href"].gsub(/^#/, "")
+  def self.process_footnote_link(docxml, elem, idx, footnote)
+    return false unless footnote?(elem)
+
+    href = elem["href"].gsub(/^#/, "")
     note = docxml.at("//*[@name = '#{href}' or @id = '#{href}']")
     return false if note.nil?
-    set_footnote_link_attrs(a, i)
-    if a.at("./span[@class = 'MsoFootnoteReference']")
-      a.children.each do |c|
-        if c.name == "span" and c["class"] == "MsoFootnoteReference"
-          c.replace(FN)
-        else
-          c.wrap("<span class='MsoFootnoteReference'></span>")
-        end
+
+    set_footnote_link_attrs(elem, idx)
+    if elem.at("./span[@class = 'MsoFootnoteReference']")
+      process_footnote_link1(elem)
+    else elem.children = FN
+    end
+    footnote << transform_footnote_text(note)
+  end
+
+  def self.process_footnote_link1(elem)
+    elem.children.each do |c|
+      if c.name == "span" && c["class"] == "MsoFootnoteReference"
+        c.replace(FN)
+      else
+        c.wrap("<span class='MsoFootnoteReference'></span>")
       end
-    else
-      a.children = FN
     end
-    fn << transform_footnote_text(note)
   end
 
   def self.transform_footnote_text(note)
@@ -76,16 +82,16 @@ module Html2Doc
     note.remove
   end
 
-  def self.footnote?(a)
-    a["epub:type"]&.casecmp("footnote")&.zero? ||
-      a["class"]&.casecmp("footnote")&.zero?
+  def self.footnote?(elem)
+    elem["epub:type"]&.casecmp("footnote")&.zero? ||
+      elem["class"]&.casecmp("footnote")&.zero?
   end
 
-  def self.set_footnote_link_attrs(a, i)
-    a["style"] = "mso-footnote-id:ftn#{i}"
-    a["href"] = "#_ftn#{i}"
-    a["name"] = "_ftnref#{i}"
-    a["title"] = ""
+  def self.set_footnote_link_attrs(elem, idx)
+    elem["style"] = "mso-footnote-id:ftn#{idx}"
+    elem["href"] = "#_ftn#{idx}"
+    elem["name"] = "_ftnref#{idx}"
+    elem["title"] = ""
   end
 
   # We expect that the content of the footnote text received is one or
@@ -94,8 +100,8 @@ module Html2Doc
   # are present in the HTML, they need to have been cleaned out before
   # passing to this gem
   def self.footnote_cleanup(docxml)
-    docxml.xpath('//div[@style="mso-element:footnote"]/a').
-      each do |x|
+    docxml.xpath('//div[@style="mso-element:footnote"]/a')
+      .each do |x|
       n = x.next_element
       n&.children&.first&.add_previous_sibling(x.remove)
     end