# David Tate
# Voter targeting and matching.

# Result should be: (tab-delimited)
# Part 1 (voter history compression)
#		$uniq_voter_id	\t$election_dt (most-recent)\t$num_times_voted
# Part 2 (matching with voter file information)
#		(above plus large amount of extra information)
#		(these records should be tab-delimited and should conform to the
#			specs in NCVoter_readme)  
# Keep tab-delimited format and other constraints in place when 
# creating the new database.  See NCVoter_readme.txt for a 
# full description of the full format.

use Time::Local;
$input_county_id = $ARGV[0];
$result_file = $ARGV[1] || "results" . $input_county_id . ".txt";

%county_voters;

$start = (times)[0];

MAIN: {
	# define which records are which so the rest can be generic code
	%record_format = ( 
				county_id => 0,
				county_desc => 1,
				voter_reg_num => 2,
				status_cd => 3,
				last_name => 9,
				first_name => 10,
				midl_name => 11,
				name_sufx_cd => 12,
				house_num => 15,
				half_code => 16,
				street_dir => 17,
				street_name => 18,
				street_type_cd => 19,
				street_sufx_cd => 20,
				unit_designator => 21,
				unit_num => 22,
				res_city_desc => 23,
				state_cd => 24,
				zip_code => 25,
				mail_addr1 => 28,
				mail_addr2 => 29,
				mail_addr3 => 30,
				mail_addr4 => 31,
				mail_city => 32,
				mail_state => 33,
				mail_zipcode => 34,
				area_cd => 36,
				phone_num => 37,
				full_phone_number => 38,
				race_code => 40,
				party_cd => 42,
				sex_code => 44,
				birth_dt => 46,
				birth_place => 47,
				registr_dt => 48,
				precinct_abbrv => 49,
				municipality_abbrv => 51,
				ward_abbrv => 53,
				cong_dist_abbrv => 55,
				super_court_abbrv => 57,
				judic_court_abbrv => 59,
				nc_senate_abbrv => 61,
				nc_house_abbrv => 63,
				county_commiss_abbrv => 65,
				township_abbrv => 67,
				school_dist_abbrv => 69,
				fire_dist_abbrv => 71,
				water_dist_abbrv => 73,
				sewer_dist_abbrv => 75,
				sanit_dist_abbrv => 77,
				rescue_dist_abbrv => 79,
				munic_dist_abbrv => 81,	
				dist_1_abbrv => 83,
				dist_2_abbrv => 85,	
				);

	%voter_format = reverse %record_format;

	condense_hist($input_county_id);
	write_new_db($input_county_id, $result_file);
}
$end = (times)[0];
print "Took: ", $end - $start, "\n";

sub condense_hist {
	my $county = shift;
	
	my $history_file = "ncvhistory/" . $county . "vht.dat";
	open HIST_FILE, "$history_file" or die "Can't open file: $history_file, count: $input_county_id: $!";

	while (<HIST_FILE>) {
		my @record = split ('\t');
	
		my $county_id        = $record[0];
		my $voter_reg_num    = $record[2];
		my $election_dt      = $record[4];
	
		# Create unique id
		my $uniq_voter_id = $county_id . $voter_reg_num;
	
		my %voter;
	
		# Grab current voter info to update it.
		if (exists($county_voters{$uniq_voter_id})) {
			%voter = %{$county_voters{$uniq_voter_id}};
		}
		
		$voter{'voted_count'}++;
	
		# Update most recent vote if needed.
		if (to_epoch($election_dt) > to_epoch($voter{'most_recent_vote'})) {
			$voter{'most_recent_vote'} = $election_dt;
		}
	
		$county_voters{$uniq_voter_id} = \%voter;
	}
}

sub write_new_db {
	my $county = shift;
	my $result_file = shift;
	open RESULT_FILE, ">$result_file" or die "Can't open $result_file: $!";

	# Create voter file name from county_id
	$voter_file = "ncvoter/" . $county . "voter.dat";
	open VOTER_FILE, "$voter_file" or die "Can't open $voter_file, in county: $county : $!";
	while (<VOTER_FILE>) {
		my @record = split('\t');
		my $updated_record = "";
	
		my $uniq_voter_id = $county_id . $voter_reg_num;
		$updated_record .= $uniq_voter_id . "\t"; 

		# Output the new database one updated record at a time ...	
		foreach my $key (sort values(%voter_format)) {
			$updated_record .= $record[$key] . "\t";	
		}

		$updated_record .= "\t" . $county_voters{$uniq_voter_id}{'most_recent_vote'} . "\t" . $county_voters{$uniq_voter_id}{'voted_count'} . "\n";

		print RESULT_FILE $updated_record;
	}
}

# Convert date to epoch for easy comparisions.
# Format of the datetime is:
# 1994-11-08 00:00:00.000	11/08/1994
sub to_epoch {
	my $date = shift;
	if ($date == undef) { return 0; }
	my @sections = split(' ', $date);
	my @parts = split('-', $sections[0]);
	
	# timelocal($sec, $min, $hours, $mday, $mon, $year)
	my $seconds = timelocal(0, 0, 0, $parts[2], $parts[1] - 1, $parts[0] - 1900); 
	return $seconds; 
}