#!/usr/bin/env ruby
#
# -*- ruby -*-
#
# Searches the google cache
#
# See file for options
#
# Created: Fri May 11 06:20:36 2007
#

require 'net/http'
require 'cgi'
require 'getoptlong'

def log(msg)
  STDERR.puts "*** " + msg
end

def scan(page, re, host)
  log(page + "...")
  h = Net::HTTP.new(host, 80)
  resp, data = h.get(page, nil)
  if resp.message == "OK"
    lines = data.split /<div/
    lines.each do |x|
      x.scan(re) { |y| yield y}
    end
  else
    puts "ERROR: #{resp}"
  end
end

def no_tags(s)
  res = s
  res = res.gsub(/<[^>]+>/, "")
  res = res.gsub(/&[^;]+;/, "")
  return res
end

def parse_args
  opts = GetoptLong.new(
                        [ "--html",           "-h",   GetoptLong::NO_ARGUMENT ],
                        [ "--links",          "-l",   GetoptLong::NO_ARGUMENT ],
                        [ "--titles",         "-t",   GetoptLong::NO_ARGUMENT ]
                        )
  
  opts.each do |opt, arg|
    $html   = true if opt == "--html"   || opt == "-h"
    $links  = true if opt == "--links"  || opt == "-l"
    $titles = true if opt == "--titles" || opt == "-t"
  end
end



########## Main ##########

$html   = false
$links  = false
$titles = false

parse_args()

$links = $titles = true if !$links and !$titles

ARGV.each do |arg|
  start = 0
  while true do
    host = "www.google.com"
    page = "/search?q=site:" + arg
    page += "&start=" + start.to_s if start != 0
    count = 0
    scan(page, /<a href="[^"]+"\s*class=l[^>]*>(.*)<\/a><\/h2>.*<span class=a>([^\s]+)\s+-[^<]*<\/span>/mi, host) do |x|
      title = no_tags x[0]
      link  = no_tags x[1]
      link = "http://" + link if !/^http.*/.match link
      if $html then
        puts "<a href='" + link + "'>" + title + "</a><br/>"
      elsif $links && $titles then
        puts title + ":" + link
      elsif $links
        puts link
      elsif $titles
        puts title
      end
      count += 1
    end
    break if count == 0
    start += 10
  end
end