#!/usr/bin/perl

# Copyright (c) 1998, 1999 Bruce Martin <bmartin@cs.ucf.edu>

# Usage:
# 1.  Set $old_path and $new_path appropriately.
# 2.  Run sense-changes < map.out > sense-changes.out

  require 5.000;	# For hard references.
  use Search::Dict;	# For binary search.

# Define the paths to the wordnet files.
  $pos       = "verb";
  $old_path  = "/home/bruce/proj/wn/wordnet/dict";
  $old_index = "$old_path/index.$pos";
  $new_path  = "/home/bruce/proj/wn/wordnet-1.6/dict";
  $new_index = "$new_path/index.$pos";

# Open the index files.
  open( OLD_INDEX, $old_index ) or die "Can't open $old_index";
  open( NEW_INDEX, $new_index ) or die "Can't open $new_index";

# Read the mapping into a hash.
  while( <> )
  {
    next unless ( $old_synset_id, $new_synset_id ) = split /\s+/;
    $map{$old_synset_id} = $new_synset_id;
  }

# Loop through the old words sequentially.
  while( <OLD_INDEX> )
  {
    my( $word, $old_sense, $new_sense );

    next unless ( $word ) = /^(\S+)/;
    foreach $old_synset_id ( / (\d{8})/g )
    {
      $new_sense = &new_word_sense( $word, $map{ $old_synset_id } );
      next if ++$old_sense == $new_sense;
      print       $word, $word=~/\d$/ && "~", $old_sense;
      print "\t", $word, $word=~/\d$/ && "~", $new_sense if $new_sense;
      print "\n";
    }
  }

###############################################################################

##### Find the new word sense for the given synset id.
  sub new_word_sense
  {
    my( $word, $new_synset_id ) = @_;
    my( $s, $sense );

    foreach $id ( &synset_ids( "NEW_INDEX", $word ) )
    {
      $s++;
      $sense = $s, last if $id == $new_synset_id;
    }
    $sense;
  }

##### List all the synset ids for a word.
  sub synset_ids
  {
    my( $index, $word ) = @_;
    my( @ids );

    look *$index, $word, 0, 1;
    $_ = <$index>;
    / (\d{8})/g if /^$word /i;
  }