#!/usr/local/bin/perl

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

# 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.  


sub isin {
	local($good,$bad) = @_;
	local($temp,@temp,$count);
	@temp = split(/_/,$bad);
	for ($count=0;$count<=$#temp;++$count) {
		if ($good eq $temp[$count]) {
			return(1); }}
	return(0);
}

open(XX,$ARGV[0]);
open(YY,$ARGV[1]);

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;
		@ffirst = split(/\//,$first[$count]);
		@ssec = split(/\//,$sec[$count]);
		@fresh = split(/_/,$ssec[1]);
		$tottags += $#fresh+1;
		if (&isin($ffirst[1],$ssec[1])) {
			++$goodwd; }
		else { # $badrecord{$ffirst[1]. " " . $ssec[1]}++;
			++$badwd; }
}
}

print "ACCURACY: ", $goodwd/$totwords,"\n";
print "AVG TAGS: ",$tottags/$totwords,"\n";

#while(($key,$val) = each %badrecord) {
#	print "$key $val\n"; }