Haskell installed packages DocSet for Dash

process.rb

require 'cgi'
require 'nokogiri'

xml = <<EOF
<?xml version="1.0" encoding="UTF-8"?>
<Tokens version="1.0">
EOF

doc = nil

open 'Documents/doc/index.html' do |f|
  doc = Nokogiri::HTML(f)
end
doc.css('li span.module a').each do |t|
  href = t['href'].scan(/doc.*/)[0]
  xml << "<File path=\"#{href}\"><Token><TokenIdentifier>//apple_ref/java/cl/#{t.content}</TokenIdentifier></Token></File>\n"
end

open 'Documents/doc/doc-index-All.html' do |f|
  doc = Nokogiri::HTML f
  src = ''
  doc.css('tr').each do |tr|
    if tr.css('td.src').first != nil
      src = tr.css('td.src').first.content
    elsif tr.css('td.module a').first != nil
      t = tr.css('td.module a').first
      href = t['href'].scan(/doc.*/)[0]
      content = CGI.escapeHTML("#{t.content}.#{src}")
      xml << "<File path=\"#{href}\"><Token><TokenIdentifier>//apple_ref/java/clm/#{content}</TokenIdentifier></Token></File>\n"
    end
  end
end

xml << '</Tokens>'

open 'Tokens.xml', 'w' do |f|
  f.write xml
end