#!/usr/bin/perl -w

## tagger - a baseline tagger


# Read training data 

open TRAIN, "$ARGV[0]" || die "Can't open training file $ARGV[0]!\n";

print STDERR "Reading $ARGV[0]\n";

while (<TRAIN>) {
  chomp;
  my @line = split /\s+/;
  foreach my $word (@line) {
    my ($w,$t) = split /_/, $word;
    $word_counts{$w}{$t}++;
    $tag_counts{$t}++;
  }
}

close TRAIN;

# Find the most common tag per word

print STDERR "Finding most common tags\n";

foreach my $w (keys %word_counts) {
  my $c = 0;
  my $tag = "";
  foreach my $t (keys %{$word_counts{$w}}) {
    if ($word_counts{$w}{$t} > $c) {
      $c = $word_counts{$w}{$t};
      $tag = $t;
    }
  }
  $tags{$w} = $tag;
}

# Find the most common tag overall

$default_tag = "";
$c = 0;

foreach my $t (keys %tag_counts) {
  if ($tag_counts{$t} > $c) {
    $c = $tag_counts{$t};
    $default_tag = $t;
  }
}

# Add tags to the test data

open TEST, "$ARGV[1]" || die "Can't open test file $ARGV[1]!\n";

print STDERR "Reading $ARGV[1]\n";

while (<TEST>) {
  chomp;
  my @line = split /\s+/;
  foreach my $w (@line) {
    if (exists $tags{$w}) {
      print $w, "_", $tags{$w}, " ";
    } else {
      print $w, "_", $default_tag, " ";
    }
  }
  print "\n";
}

close TEST;