#! /usr/local/bin/perl

#####
#
# Example web robot to make an automated query on google.com
# Before using this see: http://www.google.com/terms_of_service.html
#
# See http://www.google.com/apis/ for an authorized way to automate
# queries on google. The Net::Google or WWW::Search::Google modules 
# available from CPAN provide an interface.
#
# Other options include: WWW::Search::Scraper::Google, also on CPAN.
#
#####

#
# Required modules
use LWP::UserAgent;
use HTTP::Request;
use HTTP::Response;
use HTML::Parse;


#
# Constants
$searchterm = "porsche+911";
$root = "http://www.google.com/search?q=$searchterm";


#
# Choose what kind of browser we look like to the web site
$browserid = 'Mozilla/4.0 (compatible; MSIE 5.5; Windows NT 5.0)';
#$browserid = 'Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.0)';
#$browserid = 'Mozilla/4.0 (compatible; MSIE 5.0; Win32)';
#$browserid = 'Mozilla/4.78 [en] (Windows NT 5.0; U)';
#$browserid = 'Mozilla/4.5 [en] (X11; U; SunOS 5.6 sun4u)';


# Create a new UserAgent object
$ua = new LWP::UserAgent;

# Set the product token to fake a regular browser
# If you don't do this LWP identifies itself and google blocks the request
$ua->agent($browserid);  

# Create a request object
$request = new HTTP::Request('GET', $root);

# Get the page
$response = $ua->request($request);

# Parse the HTML in the returned page
$parsed = HTML::Parse::parse_html($response->content);

# Loop though links in the parsed HTML
for (@{ $parsed->extract_links() }) {
  $link = $_->[0];
  
  # Skip links that point back to google,
  # print the rest
  if ($link =~ /(?=^(?:(?!google\.com|search\?q=cache).)*$)^http/) {
    print "$link \n";
  }
  
}

exit 0;