#!/usr/bin/perl

#
# Codez (c) March 1, 2003
# Written by Peter Beckman <beckman@purplecow.com>
# Distributed under the GNU Public License
# Some code adapted from ddclient Perl program without permission.
# http://purplecow.com/
#

require 5.004;
use strict;
use Getopt::Long;
use Sys::Hostname;
use IO::Socket;

my $version  = "0.1";
my $result = "";
my $programd  = $0;
$programd =~ s%^.*/%%;
my $program   = $programd;
$program  =~ s/d$//;

sub _msg {
    my $log    = shift;
    my $prefix = shift;
    my $format = shift;
    my $buffer = sprintf $format, @_;
    chomp($buffer);

    $prefix = sprintf "%-9s ", $prefix if $prefix;
    $buffer .= "\n";
    print $prefix.$buffer;

    my $msgs .= $buffer  if $log;

}

sub msg     { _msg(0, '',         @_);                                  }
sub verbose { _msg(1, @_)             if opt('verbose');                }
sub debug   { _msg(0, 'DEBUG:',   @_) if opt('debug');                  }
sub warning { _msg(1, 'WARNING:', @_);                                  }

sub opt {
    my $v = shift;
    if ($v eq "verbose") { return 0; }
    if ($v eq "debug") { return 1; }
    if ($v eq "exec") { return 1; }
    #return $opt{$v}     if defined $opt{$v};
    #return $globals{$v} if defined $globals{$v};
    #return default($v)  if defined default($v);
    #return undef;
    return 0;
}

######################################################################
## geturl
######################################################################
sub geturl {
    my $proxy    = shift || '';
    my $url      = shift || '';
    my $method   = shift || 'GET';
	my $vars 	 = shift || '';
    my $login    = shift || '';
    my $password = shift || '';
    my ($peer, $server, $port);
    my ($sd, $rq, $request, $reply);

    ## canonify proxy and url
    $proxy  =~ s%^http://%%i;
    $url    =~ s%^http://%%i;
    $server = $url;
    $server =~ s%/.*%%;
    $url    = "/" unless $url =~ m%/%;
    $url    =~ s%^[^/]*/%%;

    ## determine peer and port to use.
    $peer   = $proxy || $server;
    $peer   =~ s%/.*%%;
    $port   = $peer;
    $port   =~ s%^.*:%%;
    $port   = 80 unless $port =~ /^\d+$/;
    $peer   =~ s%:.*$%%;

    my $to =  sprintf "%s%s", $server, $proxy ? " via proxy $peer:$port" : "";
    verbose("CONNECT:", "%s", $to);


    $request  = $method." ";
    $request .= "http://$server" if $proxy;
    $request .= "/$url HTTP/1.0\n";
    $request .= "Host: $server\n";

    #my $auth = encode_base64("${login}:${password}");
    #$request .= "Authorization: Basic $auth" if $login || $password;

    $vars =~ s/ /%20/g;  # should expand this for all values that need massaging
    my $varsLength = length($vars);

    $request .= "User-Agent: ${program}/${version}\n";
    $request .= "Connection: close\n";
    if ($method =~ /^post$/i) {
	  $request .= "Content-type: application/x-www-form-urlencoded\nContent-length: $varsLength\n";
	  $request .= "\n".$vars;
	}
    $request .= "\n";

    ## make sure newlines are <cr><lf> for some pedantic proxy servers
    ($rq = $request) =~ s/\n/\r\n/g;
    # local $^W = 0;
    if (! opt('exec')) {
        debug("skipped network connection");
        verbose("SENDING:", $request);

    } elsif (! defined($sd = IO::Socket::INET->new(PeerAddr => $peer, PeerPort => $port, Proto => 'tcp'))) {
        warning("cannot connect to $peer:$port socket: $@");

    } else {
        ## send the requet to the http server
        verbose("CONNECTED:", "");
        verbose("SENDING:", $request);

        my $result = send $sd, $rq, 0;
        if ($result != length($rq)) {
            warning("cannot send to $peer:$port ($!).");

        } else {
            my $timeout = 0;
            local $SIG{'ALRM'} = sub { $timeout = 1; };

            alarm(opt('timeout')) if opt('timeout') > 0;
            while (!$timeout &&  ($_ = <$sd>)) {
                verbose("RECEIVE:", "%s", $_);
                $reply .= $_ if defined $_;
            }
            if (opt('timeout') > 0) {
                alarm(0);
            }
            close($sd);
            if ($timeout) {
                warning("TIMEOUT: %s after %s seconds", $to, opt('timeout'));
                $reply = '';
            }
            $reply = '' if !defined $reply;
        }
    }

    ## during testing simulate reading the URL

#    if (opt('test')) {
#        my $filename = "$server/$url";
#        $filename =~ s|/|%2F|g;
#        if (opt('exec')) {
#            $reply = save_file("${savedir}$filename", $reply, 'unique');
#        } else {
#            $reply = load_file("${savedir}$filename");
#        }
#    }

    $reply =~ s/\r//g if defined $reply;
    return $reply;
}

sub getzip {
    my $zip  = shift || '';
    my $url = "http://www.usps.com/zip4/zip_response.jsp";
    my $reply = geturl('',$url,'POST','zipcode='.$zip);
	my @lines = split(/\n/,$reply);
	my $start = 0;
	my $line; 
	my $row = "";
	my $in = 0;
	my $num = 0;
	my @out = ();
	
	print "Cities that match Zip Code $zip\n";
	print "-----------------------------------------------------------------------\n";
	foreach $line (@lines) {
	  if (opt('verbose')) { print $line."\n"; }
	  if ($line =~ /ZIP Code Type/) {
		 $start = 1;
	  }
	  if ($start) {
		 if ($line =~ /^[A-Z]/) {
			$line =~ s/<.*>//g;
			$out[$num][$in] = $line;
			$in++;
		 }
		 if ($in == 4) {
			$in = 0;
			$num++;
		 }
	  }
	}
	$in = 0;
	foreach $row (@out) {
	  printf("%-40s %-21s %-15s\n",$row->[0].", ".$row->[1], $row->[2], $row->[3]);
	}
	print "\nData Scraped from USPS.com without permission\nVisit http://www.usps.com/ and buy stuff\nfrom them so they don't mind.\n";

}

sub getcitystate {
    my $zip  = shift || '';
    my $url = "http://56.0.134.24/zip4/citystate_response.jsp";

	# process citystate stuff
	my @split = split(/,/,$zip);
	my $state = uc($split[1]);
	$state =~ s/\s+//g;
	my $city = ucfirst($split[0]);
	$city =~ s/^\s+//g;
	$city =~ s/\s+$//g;
	$zip = $city." ".$state;
    my $reply = geturl('',$url,'POST','citystate='.$zip.'&state='.$state.'&city='.$city.'&Selection=2&Submit_x=29&Submit_y=10');
	my @lines = split(/\n/,$reply);
	my $start = 0;
	my $line; 
	my $row = "";
	my $in = 0;
	my $num = 0;
	my @out = ();
	
	print "Zip Codes that match $zip\n";
	print "-----------------------------------------------------------------------\n";
	foreach $line (@lines) {
	  if ($line =~ /ZIP Codes/) {
		 $start = 1;
	  }
	  if ($start) {
		 if ($line =~ /^\s+\<center\>[\w \(\)]{5,}\<\/center\>/) {
			#print $in.": ".$line."\n";
			$line =~ s/<[^>]*>|\s{2,}//g; 
			$out[$num][$in] = $line;
			$in++;
		 }
		 if ($in == 3) {
			$in = 0;
			$num++;
		 }
	  }
	}
	$in = 0;
	foreach $row (@out) {
	  printf("%-10s %-30s %-20s\n",$row->[0], $row->[1], $row->[2]);
	}
	print "\nData Scraped from USPS.com without permission\nVisit http://www.usps.com/ and buy stuff\nfrom them so they don't mind.\n";

}


sub getarea {
    my $ac  = shift || '';
    my $url = "http://www.melissadata.com/lookups/phonelocation.asp?number=".$ac."&submit1=Process";
    my $reply = geturl('',$url,'GET');
	my @lines = split(/\n/,$reply);
	my $start = 0;
	my $line; 
	my @row = "";
	my $row = "";
	my $in = 0;
	my $num = 0;
	my @out = ();
     my $hdrcnt; my $x;
	 my @hdr;
	
	print "Info about $ac\n";
	print "---------------------------------------------------------------------------\n";
	foreach $line (@lines) {
	  if ($line =~ /value="Process"/) {
		 $start = 1;
	  }
	  if ($start) {
		 if ($line =~ /\<tr\>\<td/) {
			my @foo = split(/<\/td><td[^>]*>/,$line);
			for ($x=0;$x<=$#foo;$x++) {
			   $foo[$x] =~ s/<[^>]*>|\s{2,}//g; 
			   $foo[$x] =~ s/&nbsp;/ /g; 
			   $out[$num][$x] = $foo[$x];
			}
			$num++;
		 } elsif ($line =~ /\<tr[^>]*><t[h|d][^>]*>/) {
			if ($hdrcnt == 0) { $hdrcnt++; }
			elsif ($hdrcnt == 1) {
			   @hdr = split(/<\/th><th[^>]*>/i,$line);
			   for ($x=0;$x<=$#hdr;$x++) {
				  $hdr[$x] =~ s/<[^>]*>|\s{2,}//g; 
			   }
			}
		 }
		 #if ($line =~ /^<\/table>$/) { last; }
	  }
	}
	$in = 0;
    printf("%-30s %-25s %-15s %-15s\n", $hdr[0], $hdr[1], $hdr[2], $hdr[3]);
	foreach $row (@out) {
	  printf("%-30s %-25s %-15s %-15s\n",$row->[0],$row->[1], $row->[2], $row->[3]);
	  #print $in++.": ".$row.":".$row->[0]."\n";
	}
	print "\nData Scraped from MelissaData without permission\nVisit http://www.melissadata.com/ and buy stuff\nfrom them so they don't mind.\n";
}

sub parseinput {
   my $input = shift || '';
   print "Codez v".$version." (C)2003 by Peter Beckman\n";
   if ($input eq "") {
	  print "No input.\nExiting...\n";
	  exit(0);
   }
   if ($input =~ /^\d{5}(-?[\d]{4})?$/) {
	  if ($input !~ /-/ and length($input)>5) { $input = substr($input,0,5)."-".substr($input,5,4); }
	  getzip($input);
      debug("getzip 1");
   } elsif ($input =~ /^[0-9]{3,10}$/) {
	  getarea($input);
      debug("getarea 2");
   } elsif ($input =~ /^[0-9]{3}-[0-9]{3,4}$/) {
	  getarea($input);
      debug("getarea 3");
   } elsif ($input =~ /^[0-9]{3}-[0-9]{3}-[0-9]{4}$/) {
	  getarea($input);
      debug("getarea 4");
   } elsif ($input =~ /^[a-zA-Z0-9 ]+,?\s*[a-zA-Z]{2}$/) {
	  getcitystate($input);
      debug("getcitystate 5");
   } else {
	  print "Not a valid area code, zip code or city/state.\nExiting...\n";
	  exit(0);
   }
}

my $myinput = $ARGV[0]." ".$ARGV[1]." ".$ARGV[2]." ".$ARGV[3];
$myinput =~ s/^\s*//g;
$myinput =~ s/\s*$//g;
parseinput($myinput);

__END__

=head1 NAME

codez - return information about area codes or zip codes

=head1 SYNOPSIS

codez B<code>

=head1 DESCRIPTION

codez gets all relevant information about a US Zip Code or telephone
area code, exchanges and phone numbers themselves.  The data is scraped
from USPS.com (zip codes) and MelissaData.com (phone info).

"B<code>" is one of the following:
   - B<a 5 digit US zip code>
   - B<a 3 digit area code>
   - B<a 6 digit area code + exchange>
   - B<a 7 digit phone number (not including the area code)>
   - B<a 10 digit area code + phone number>
   - B<a city, state combination (i.e. "New York, NY")>

=head1 EXAMPLES

 codez Falls Church, VA
 codez Falls Church VA
 codez 22042
 codez 22042-1917
 codez 703
 codez 703-536
 codez 536-3600
 codez 703-536-3600
 codez 7035363600

=head1 ACTUAL REGEXPS

C< /^\d{5}(-?[\d]{4})?$/ >             -> C< getzip($input); >
C< /^[0-9]{3,10}$/ >                   -> C< getarea($input); >
C< /^[0-9]{3}-[0-9]{3,4}$/ >           -> C< getarea($input); >
C< /^[0-9]{3}-[0-9]{3}-[0-9]{4}$/ >    -> C< getarea($input); >
C< /^[a-zA-Z0-9 ]+,?\s*[a-zA-Z]{2}$/ > -> C< getcitystate($input); >


=head1 ACKNOWLEDGMENTS

Peter Beckman wrote it.  Yeah.  Visit his website at I<http://purplecow.com/>.

=cut