#!/usr/local/bin/perl

# Copyright @ 1994 MIT 
# Written by Eric Brill 
#THIS SOFTWARE IS PROVIDED "AS IS", AND M.I.T. MAKES NO REPRESENTATIONS 
#OR WARRANTIES, EXPRESS OR IMPLIED.  By way of example, but not 
#limitation, M.I.T. MAKES NO REPRESENTATIONS OR WARRANTIES OF 
#MERCHANTABILITY OR FITNESS FOR ANY PARTICULAR PURPOSE OR THAT THE USE OF 
#THE LICENSED SOFTWARE OR DOCUMENTATION WILL NOT INFRINGE ANY THIRD PARTY 
#PATENTS, COPYRIGHTS, TRADEMARKS OR OTHER RIGHTS.  


# given 2 files, prints percentage of matches (% correct if one is
# gold standard.

# print("XX - $ARGV[0]\n");
# print("YY - $ARGV[1]\n");
# print("ZZ - $ARGV[2]\n");
open(XX,$ARGV[0]);
open(YY,$ARGV[1]);
open(ZZ,$ARGV[2]);
while(<ZZ>) {
	@temp = split;
	$seen{$temp[0]}++;
}
while(<XX>) {
	$first = $_;
	$sec = <YY>;
	@first = split(/\s+/,$first);
	@sec = split(/\s+/,$sec);
	if ($#first != $#sec) {
		print "SCREW UP!!\n";
		print "FIRST: @first\n";
		print "SEC: @sec\n";
		$unknscore = $unseengood/($unseengood+$unseenbad);
		print "unk score is: $unknscore  with $unseengood good\n";
		print "and $unseenbad bad\n";
		print "and $totwords words and $goodwords correct\n";
		exit;
	}
	for ($count=0;$count<=$#first;++$count) {
		++$totwords;
		if ($first[$count] eq $sec[$count]) {
			@split = split(/\//,$first[$count]);
			@split2 =  split(/\//,$sec[$count]); 
			if ($seen{$split[0]}) {
				$seengood++;
			}
			else {
				$unseengood++;
			}
			++$goodwords;
			$goodtag{$split[0] . " " . $split[1] . " " . $split2[1]}++; 
		}
		else {
			@split1 = split(/\//,$first[$count]);
			@split2 = split(/\//,$sec[$count]); 
			if ($seen{$split1[0]}) {
				$seenbad++;
			}
			else {
				$unseenbad++;   $unkbadtag{$split1[0] . " " . $split1[1] . " " . $split2[1]}++
			}
				$badtag{$split1[0] . " " . $split1[1] . " " . $split2[1]}++;
		}
	}
}

$score = $goodwords/$totwords;
print "score is: $score\n";
$unknscore = $unseengood/($unseengood+$unseenbad);
print "unk score is: $unknscore\n";
print "unk score is: $unknscore  with $unseengood good\n";
print "and $unseenbad bad\n";
print "SEEN GD: $seengood SEEN BAD: $seenbad\n";
print "and $totwords words and $goodwords correct\n";
if ($ARGV[3] eq "-verbose") {
	print "BAD TAGS\n";
	while(($key,$val) = each %badtag) {
		print "$key $val\n";
	}
	print "GOOD TAGS\n";
	while(($key,$val) = each %goodtag) {
		print "$key $val\n";
	}
}