module Nanoc
  class PygmentizerFilter < Nanoc::Filter
    identifier :pygmentizer
    type :text

    def run(content, params = {})
      doc = Nokogiri::HTML::fragment(content, 'UTF-8')
      nodes = doc.search("pre>code")

      nodes.each do |node|
        code = node.inner_html || "[++where is the code?++]"
        node.parent.swap(pygmentize(code))
      end

      doc.to_html
    end

    private
    def pygmentize(string, pattern = /\A:::(\w+)\s*(\n|&#x000A;)/i)
      refs = pattern.match(string) # extract language name
      if refs # found a language identifier!
        lang = refs[1]
        str = unescape_html(string.sub(pattern, ""))
        code = ::Pygments.highlight(str, options: {
          encoding: 'utf-8',
          linenos: 'table',
          lexer: lexers.include?(lang) ? lang : nil
        })
        "<div class='code'>#{code}</div>"
      else # No language identifer,
        "<pre class='plaincode'>#{string}</pre>"
      end
    end

    private
    def lexers
      @lexers ||= Pygments.lexers.map { |lex| lex[1][:aliases] }.flatten
    end

    private
    def unescape_html(string)
      string.to_s.gsub(/&#x000A;/i, "\n").gsub("&lt;", '<').gsub("&gt;", '>').gsub("&amp;", '&')
    end  end
end