dmm.rb

require 'rubygems'
require 'open-uri'
require 'nokogiri'

KCODE = 'euc-jp'

index_url = 'http://www.dmm.co.jp/digital/videoa/-/actress/=/keyword=%s/'

%w(a ka sa ta na ha ma ya ra wa).each do |w|
  content = Nokogiri::HTML.parse(open(index_url % w).read)
  links = content.xpath("//div[@class='act-box']/ul/li/a")
  links.each do |link|
    name = link.text
    url = link.attribute('href').to_s
    id = url.gsub(/.*id=([0-9]+).*/, '\1')
    image_path = link.xpath('img').attribute('src').to_s
    puts "#{id},#{name},#{image_path}"
  end
end