#!/usr/bin/perl -w -s

my ($f1,$f2) = @ARGV;

our ($batch);

my $RATIO = 0.19;

die "Need two aligned files in NATools format\n" unless -f $f1 && -f $f2;

open F1, $f1 or die "Cannot open file $f1\n";
open F2, $f2 or die "Cannot open file $f1\n";

if ($batch) {
  open O1, ">$f1.wrong" or die "Cannot open file $f1.wrong\n";
  open O2, ">$f2.wrong" or die "Cannot open file $f2.wrong\n";

  open OK1, ">$f1.ok" or die "Cannot open file $f1.ok\n";
  open OK2, ">$f2.ok" or die "Cannot open file $f2.ok\n";
} else {
  open N1, ">$f1.new" or die "Cannot open file $f1.new\n";
  open N2, ">$f2.new" or die "Cannot open file $f2.new\n";
}


$/ = "\n\$\n";

my ($l,$r);

my $min = 1;
my $max = 0;
my $bad = 0;

my $quit = 0;

while (defined($l = <F1>) && defined($r = <F2>))  {

  chomp($l);
  chomp($r);

  my $cl =()= $l =~ /\s+/g; $cl++;
  my $cr =()= $r =~ /\s+/g; $cr++;

  $max = $cl if $cl > $max;
  $max = $cr if $cr > $max;

  my $ratio = min($cl,$cr)/max($cl,$cr);

  if ($batch) {
    if ($ratio < $RATIO) {
      $min = $ratio if $min > $ratio;
      $bad++;
      print "($cl,$cr) $ratio\n" if $ratio < $RATIO;
      print O1 "$l\n\$\n";
      print O2 "$r\n\$\n";
    } else {
      print OK1 "$l\n\$\n";
      print OK2 "$r\n\$\n";
    }
  } else {
    print STDERR "$ratio, $cl, $cr\n" if $ratio < $RATIO;
    if ($ratio < $RATIO) {
      $min = $ratio if $min > $ratio;
      $bad++;
      if ($quit) {
	print N1 "$l\n\$\n";
	print N2 "$r\n\$\n";
      } else {
	print "\n\n-[$.]----------- ($cl,$cr) $ratio\n" if $ratio < $RATIO;
	print "\n$l\n---\n$r";

	print "\n *** (A)ccept (D)elete (E)dit (Q)uit *** \n";

	my $ans;
	{ local $/ = "\n"; chomp($ans = <STDIN>) }

	if ($ans =~ m!^a!i) {
	  print "Accepting pair...\n";
	  print N1 "$l\n\$\n";
	  print N2 "$r\n\$\n";
	} elsif ($ans =~ m!^d!i) {
	  print "Deleting pair...\n";
	} elsif ($ans =~ m!^q!i) {
	  print "Accepting all remaining pairs...\n";
	  print N1 "$l\n\$\n";
	  print N2 "$r\n\$\n";
	  $quit = 1;
	} elsif ($ans =~ m!^e!i) {
	  my ($nl,$nr);
	  print "New left: ";
	  { local $/ = "\n"; chomp($nl = <STDIN>) }
	  print "New right: ";
	  { local $/ = "\n"; chomp($nr = <STDIN>) }
	  print N1 "$nl\n\$\n";
	  print N2 "$nr\n\$\n";
	} else {
	  print "What? $ans? Bah.. accepting...\n";
	  print N1 "$l\n\$\n";
	  print N2 "$r\n\$\n";
	}
      }

    } else {
      print N1 "$l\n\$\n";
      print N2 "$r\n\$\n";
    }


  }

}

print "Min: $min\n";
print "Max Sentence size in words: $max\n";
print "Number of bad pairs: $bad\n";


sub max {
  $_[0]>$_[1]?$_[0]:$_[1]
}

sub min {
  $_[0]<$_[1]?$_[0]:$_[1]
}