use strict;
use LWP::Simple;
use DB_File;
use Digest::MD5 qw(md5_hex);

my $alert_amount = 0.1; #  difference to alert
my @sites = (
	     'http://www.herdofsheep.com',
	     'http://www.theonion.com',
	     'http://news.bbc.co.uk',
	     );
my $db_file = 'sites.db';

my (%sites_db);
tie %sites_db, "DB_File", "$db_file", O_RDWR|O_CREAT or die "Cannot open database '$db_file': $!";

foreach (@sites) {
  my ($lines, $matches, $new_entry);
  my %new = map { md5_hex($_) => 1 } split(/<.*?>/, get($_));
  unless (%new) {print "Could not get $_\n"; next}
  my %old = map { $_ => 1 } split(',', $sites_db{$_}) if exists($sites_db{$_});

  foreach (keys %new) {
    $lines++;
    $new_entry .= $_ . ',';
    $matches++ if $old{$_};
  }
  if ((($lines - $matches) / $lines) > $alert_amount ) {
    print "$_ seems to be different\n";
    $sites_db{$_} = $new_entry;
  }
}