Gihyo Digital Publishingの本のソースコードをシンタックスハイライトする

rougify-gdp-book.rb

# coding: utf-8
gem 'epub-parser', '>= 0.2.4'
gem 'epub-maker', '0.0.3'
require 'English'
require 'epub/maker'
require 'rouge'
require 'rouge/lexers/docker'

ROUGE_THEME = 'github'
ROUGE_CSS_SCOPE = 'code'
CSS_PATH = 'syntax.css'
$formatter = Rouge::Formatters::HTML.new(wrap: false)
$default_lexers = {
  '.rb' => :Ruby,
  'Vagrantfile' => :Ruby,
  '.yaml' => :YAML,
  '.yml' => :YAML,
  '.toml' => :TOML,
  '.sh' => :Shell,
  '.go' => :Go,
  '.json' => :JSON,
  'nginx.conf' => :Nginx,
  'Dockerfile' => :Docker,
  '.ini' => :INI
}
$lexers = {
  'urn:uuid:bffeb831-ee98-40a9-9eae-b5693185ce74' => {
    'リスト2　パッケージからインストールした場合の設定ファイル' => :TOML,
    'リスト6　Private Variableの参照' => :YAML,
    'リスト11　envの設定' => :YAML,
    'リスト12　scriptの記述例' => :YAML,
    'リスト13　notifyの設定例' => :YAML,
    'リスト14　publishの設定例' => :YAML,
    'リスト15　gitの設定例' => :YAML,
    'リスト16　sshの設定例' => :YAML,
    'リスト17　bashの設定例' => :YAML,
    'supervisord.conf' => :INI
  },
  'urn:uuid:14eadcd9-214a-4f45-9623-d8ac54f22af7' => {
    'Ruby' => :Ruby,
    'C言語' => :C,
    'Perl' => :Perl,
    'Python' => :Python,
    'ノンブロッキングI/Oはfdのフラグ' => :Ruby,
    'システムコール' => :C,
    'fdがオープンされたまま残る' => :Python,
    'timegm関数と閏秒' => :C,
    'Rubyにおける時刻の扱いの問題' => :PlainText,
  }
}

def main(argv)
  book_path = argv.shift
  $book = parse_book(book_path)
  add_css_files
  highlight_contents
  update_or_insert_modified_date Time.now
rescue => error
  $stderr.puts error
  $stderr.puts error.backtrace
  $stderr.puts "Usage: ruby #{$PROGRAM_NAME} EPUB_PATH"
  abort
end

def parse_book(path)
  EPUB::Parser.parse(path)
end

def add_css_files
  dummy_origin = Addressable::URI.parse('file:///')
  rootfile_path = dummy_origin + $book.ocf.container.rootfile.full_path
  theme = ROUGE_THEME
  scope = ROUGE_CSS_SCOPE
  if $book.unique_identifier.content == 'urn:uuid:479892c1-6cd6-409e-915e-647aca7e25a2'
    scope = 'code[data-rouge-language="nginx"]'
  end
  style = Rouge::Theme.find(theme).new(scope: scope).render
  $book.package.edit do |package|
    package.manifest.make_item do |item|
      item.href = Addressable::URI.parse("../#{CSS_PATH}") # IMPROVEMENT: Less need to call Addressable::URI.parse explicitly
      # IMPROVEMENT: Want to call item.entry_name = $css_path
      item.media_type = 'text/css'
      item.id = CSS_PATH
      item.content = style
      item.save # IMPROVEMENT: Less need to call save explicitly
    end
  end
end

def highlight_contents
  $book.resources.select(&:xhtml?).each do |xhtml|
    xhtml.edit_with_nokogiri do |doc|
      add_links_to_css_files(doc, xhtml)
      markup(doc)
    end
  end
end

def add_links_to_css_files(doc, xhtml)
  dummy_root_iri = Addressable::URI.parse('http://example.net/')
  xhtml_entry_name = dummy_root_iri + xhtml.entry_name
  head = (doc/'head').first
  entry_name = dummy_root_iri + CSS_PATH
  href = entry_name.route_from(xhtml_entry_name)
  link = Nokogiri::XML::Node.new('link', doc)
  link['href'] = href
  link['type'] = 'text/css'
  link['rel'] = 'stylesheet'
  head << link
end

def markup(doc)
  lexers = $default_lexers.merge($lexers[$book.unique_identifier.content] || {})
  lexers_regexp = Regexp.new('(?<extension>' + lexers.keys.map {|k| k.gsub('.', '\.')}.join('|') + ')')
  (doc/'code').each do |code|
    list_name = ''
    case $book.unique_identifier.content
    when 'urn:uuid:bffeb831-ee98-40a9-9eae-b5693185ce74'
      list_name = code.parent.previous_element.content if code.parent.previous_element
    when 'urn:uuid:14eadcd9-214a-4f45-9623-d8ac54f22af7'
      classes = code['class'] || ''
      classes = classes.split(/\s+/)
      next if classes.include? 'bw'
      prev  = code.parent.previous_element
      while prev
        if prev.name =~ /h\d/
          list_name = prev.content
          break
        end
        prev = prev.previous_element
      end
      if list_name == 'クローズして問題を解決する'
        list_name = if code.content.start_with? 'from '
                      'Python'
                    elsif code.content.start_with? 'use '
                      'Perl'
                    end
      end
      unless list_name.match lexers_regexp
        list_name = case list_name
                    when ''
                      'C言語'
                    else
                      'Ruby'
                    end
      end
    when 'urn:uuid:479892c1-6cd6-409e-915e-647aca7e25a2'
      code_title_element = code.parent.previous_element
      classes = []
      if code_title_element && code_title_element['class']
        classes = code_title_element['class'].split(/\s+/)
      end
      if classes.include? 'code_title'
        list_name = code_title_element.content
      end
      if list_name.match 'nginxの公式リポジトリ情報（Debian GNU/Linux）'
        list_name = ''
      end
      if list_name.match 'nginxの公式リポジトリ情報（Debian GNU/Linux）'
        list_name = '.ini'
      end
      content = code.content
      unless content.lines.first.chomp == '# pkg install nginx' or content[0] == '$'
        list_name = 'nginx.conf'
      end
    end
    match = list_name.match lexers_regexp
    list_marker = '◆リスト'
    if match and
      ($book.unique_identifier.content == 'urn:uuid:bffeb831-ee98-40a9-9eae-b5693185ce74' ? list_name.start_with?(list_marker) : true)
      lexer_name = lexers[match[:extension]]
      lexer = Rouge::Lexers.const_get(lexer_name).new
      if $book.unique_identifier.content == 'urn:uuid:479892c1-6cd6-409e-915e-647aca7e25a2'
        code['data-rouge-language'] = 'nginx'
      end
    else
      lexer = Rouge::Lexer.guess(source: code.content)
    end
    code.inner_html = $formatter.format(lexer.lex(code.content))
  end
end

def update_or_insert_modified_date(time)
  $book.package.edit do |package|
    modified = package.metadata.metas.find {|meta|
      meta.property == 'dcterms:modified'
    }
    unless modified
      modified = EPUB::Publication::Package::Metadata::Meta.new
      modified.property = 'dcterms:modified'
    end
    modified.content = time.utc.iso8601
  end
end

main(ARGV)