#!/usr/bin/perl

use strict;
use Getopt::Long;

#my $outdir = 'bylsa';
my $tolerance = 5;
my $trim = '';
my $justone = '';
#GetOptions("outdir=s", \$outdir, "iterlimit=i", \$iterlimit);
GetOptions("tolerance=i", \$tolerance, "trim!", \$trim, "justone!", \$justone);
print $justone;

#my %boundaries;
my %boundaryloc;

foreach my $file (@ARGV) {
	open READ, $file;
	$file =~ s#/.*## if $trim;
	my $curline = 1;
	#my $prevtime = "00:00";
	while (<READ>) {
		#if (/^(\d\d:\d\d)/) {
			#$prevtime = $1;
		#}
		#elsif (/^#boundary/) {
			#	my $nexttime = <READ>;
			#chomp $nexttime;
			#$nexttime =~ s/^(\d\d:\d\d).*/$1/;
			#push @{$boundaries{$file}}, "$prevtime/$nexttime";
		#}
		#
		if (/^#boundary/) {
			$boundaryloc{$file}{$curline} = $curline;
		} else {
			$curline++;
		}
	}
}

my %correlation;
my %numboundaries;
my %matches;
foreach my $file (keys %boundaryloc) {
	$numboundaries{$file} = scalar keys(%{$boundaryloc{$file}});
	print "$file has $numboundaries{$file} boundaries\n";
	foreach my $offset (keys %{$boundaryloc{$file}}) {
		OTHERFILE: foreach my $other (keys %boundaryloc) {
			next if $other eq $file;
			foreach my $otheroff (($offset-$tolerance)..($offset+$tolerance)) {
				if ($boundaryloc{$other}{$otheroff}) {
					#print "$file$offset/$other$otheroff\n";
					#$correlation{$file}{$other}++ unless ($justone && scalar @{$matches{$file}{$other}{$offset}});
					$correlation{$file}{$other}++ unless ($justone && scalar $matches{$file}{$other}{$otheroff});
					$matches{$file}{$other}{$otheroff} = $offset;
					next OTHERFILE if $justone;
					#finding multiple boundaries from
					#somebody else is good -- they
					#think it's redundant, and you
					#correlate with all of them
				}
			}
		}
	}
}




#my %index;
#my %correlation;
#foreach (keys %boundaries) {
	##print $boundaries{$_}[0] . "\n";
	#print "$_ has " . scalar @{$boundaries{$_}} . " boundaries.\n";
	#$index{$_} = 0;
	#foreach my $other (keys %boundaries) {
		#next if $other eq $_;
		#if (&boundariesmatch($boundaries{$_}[0], $boundaries{$other}[0])) {
			#$correlation{$_}{$other}++;
			#}
		#}
#}


#while (!allIndeciesAtBottom(\%boundaries, \%index)) {
	##find an earliest one
	#my $earliesttime = "24:00/24:00";
	#my $earliest;
	#foreach (keys %boundaries) {
		#if (earlier($boundaries{$_}[$index{$_}], $earliesttime)) {
			#$earliest = $_;
			#$earliesttime= $boundaries{$_}[$index{$_}];
			#}
		#}
	#last if $earliesttime eq "24:00/24:00";  # hack
	##advance it
	##print "$earliest^".$index{$earliest}++ . " ";
	#$index{$earliest}++;
	#check to see if it matches any others
	#foreach (keys %boundaries) {
		#next if $_ eq $earliest;
		#if (boundariesmatch($boundaries{$earliest}[$index{$earliest}], $boundaries{$_}[$index{$_}])) {
			##stick the match in both their cells
			#$correlation{$earliest}{$_}++;
			#$correlation{$_}{$earliest}++;
			#}
		#}
#}

foreach my $file (sort keys %correlation) {
	print "*** Correlation for $file:\n";
	foreach my $other (keys %{$correlation{$file}}) {
		#print "$correlation{$file}{$other}\t$other\n";
		printf "$correlation{$file}{$other}:%.3f\t$other\n", $correlation{$file}{$other} / $numboundaries{$other};
	}
}


sub allIndeciesAtBottom {
	my $boundhash = shift;
	my $indexhash = shift;
	my $atbottom = 1;
	foreach (keys %$boundhash) {
		$atbottom = 0 if ($indexhash->{$_} < $#{$boundhash->{$_}});
		#print $indexhash->{$_} ."<". $#{$boundhash->{$_}} . "?";
		#print $indexhash->{$_} < $#{$boundhash->{$_}};
		#print "\n";
	}
	return $atbottom;
}

sub boundariesmatch {
	my @b1 = split m#/#, shift;
	my @b2 = split m#/#, shift;

	if ($b1[0] eq $b2[0] || $b1[0] eq $b2[1] || $b1[1] eq $b2[1] || $b1[1] eq $b2[0]) {
		return 1;
	}
	else { return 0; }
}

sub earlier {
	return 0 if $_[0] eq '';
	#print "$_[0]*$_[1]_" . ($_[0] lt $_[1]);
	return $_[0] lt $_[1];
	#my @time1 = split m#/#, shift;
	#my @time2 = split m#/#, shift;
	#
	#if ($time1[0] eq $time2[0]) {
		#if ($time1[1] eq $time2[1]) {
			#return 0;
		#}
		#if ($time1[1] le 
}