fixup-docbook.rb

   1 #!/usr/bin/ruby
   2 # coding: utf-8
   3
   4 require 'nokogiri'
   5
   6 LogLevel=1
   7 raise ArgumentError, 'Source/destination files not specified' if ARGV.size != 2
   8 srcfile = ARGV[0]
   9 dstfile = ARGV[1]
  10
  11 f=File.open(srcfile)
  12 xml = Nokogiri::XML(f)
  13
  14 def log(level,what)
  15   indicators = %w(! • # -)
  16   if level >= LogLevel
  17     print indicators[level]
  18   else
  19     print "\n%s %s\n" % ['*' * (level+1), what]
  20   end
  21 end
  22
  23 def partreplace(xml, partid, tag)
  24   xml.css('part[id=' + partid + ']').each do |part|
  25     part.name = tag
  26   end
  27 end
  28
  29
  30 log 0, 'replace article* with book*'
  31 xml.css('articleinfo').each do |node|
  32   node.name = 'bookinfo'
  33   node.first_element_child.before(<<'XML')
  34 <publisher>
  35     <publishername>Gunnar Wolf</publishername>
  36     <address><city>Mexico City</city></address>
  37 </publisher>
  38 XML
  39   node.first_element_child.before(<<'XML')
  40 <copyright>
  41     <year>2017</year>
  42     <holder>Creative Commons</holder>
  43 </copyright>
  44 XML
  45 end
  46 xml.css('article').each do |node|
  47   node.name = 'book'
  48   node['lang'] = 'en'
  49 end
  50
  51 log 0, 'change parts to colophon, dedication and chapter'
  52 partreplace(xml, 'colophon', 'colophon')
  53 partreplace(xml, 'dedication', 'dedication')
  54 partreplace(xml, 'foreword', 'preface')
  55 partreplace(xml, 'introduction', 'preface')
  56 partreplace(xml, 'bibliography', 'chapter')
  57 partreplace(xml, 'acknowledgments', 'chapter')
  58
  59 log 0, 'place part introduction into <partintro>'
  60 s = xml.xpath("//part/title[text()='The Case Studies']")[0]
  61 if s
  62   s.after('<partintro>')
  63   p = xml.css('part partintro')[0]
  64   s.parent.xpath("//part/para").each do |node|
  65     node.parent = p
  66   end
  67 end
  68
  69 log 0, 'remove title from dedication'
  70 xml.css('dedication title')[0].content = ""
  71
  72 log 0, 'move legal notice to bookinfo'
  73 xml.css('book bookinfo')[0].first_element_child.before('<legalnotice>')
  74 ln = xml.css('book bookinfo legalnotice')[0]
  75 xml.css('para').each do |para|
  76   if para.content =~ /This book is published under a/
  77     log 0, 'found legal'
  78     para.parent = ln
  79     break
  80   end
  81 end
  82
  83 log 0, 'replace colophon page with one for this edition'
  84  xml.xpath('//colophon/para').remove
  85 s = xml.xpath('//colophon')[0]
  86 s.first_element_child.after(<<'XML')
  87 <para>Made With Creative Commons</para>
  88
  89 <para>by Paul Stacey & Sarah Hinchliff Pearson</para>
  90
  91 <para>© 2017 by the Creative Commons Foundation.</para>
  92
  93 <para>Published under a Creative Commons Attribution-ShareAlike
  94 license (CC BY-SA), version 4.0.</para>
  95
  96 <para>ISBN: YET-TO-BE-DECIDED (PDF), YET-TO-BE-DECIDED (ePub),
  97 YET-TO-BE-DECIDED (Paperback) </para>
  98
  99 <para>Illustrations by Bryan Mathers, <ulink url="https://bryanmathers.com/"/></para>
 100
 101 <para>Publisher: Gunnar Wolf.</para>
 102
 103 <!-- space for information about translators -->
 104 <para>&nbsp;</para>
 105
 106 <para>Downloadable e-book available at
 107 <ulink url="https://madewith.cc/"/></para>
 108
 109 <para>This book is published under a CC BY-SA license, which means that you
 110 can copy, redistribute, remix, transform, and build upon the content for
 111 any purpose, even commercially, as long as you give appropriate credit,
 112 provide a link to the license, and indicate if changes were made. If you
 113 remix, transform, or build upon the material, you must distribute your
 114 contributions under the same license as the original. License details:
 115 <ulink url="http://creativecommons.org/licenses/by-sa/4.0/"/></para>
 116
 117 <para>Made With Creative Commons is published with the kind support of
 118 Creative Commons and backers of our crowdfunding-campaign on the
 119 Kickstarter.com platform.</para>
 120
 121 XML
 122
 123 log 0, 'change CC logo images to informalfigure'
 124 xml.css('figure mediaobject imageobject imagedata[width="40.0%"]').each do |id|
 125   f = id.parent.parent.parent
 126   f.name = 'informalfigure'
 127 end
 128
 129 log 0, 'Writing processed file'
 130 # Unable to figure out API way to replace DOCTYPE
 131 data = xml.to_xml().gsub!(/DOCTYPE article/, 'DOCTYPE book')
 132 File.open(dstfile, 'w') {|f| f.write(data)}