Hiroyuki Komatsu
komat****@users*****
2005年 3月 7日 (月) 19:01:57 JST
Index: prime/lib/makedict/dictformat-wordnet.rb diff -u /dev/null prime/lib/makedict/dictformat-wordnet.rb:1.2 --- /dev/null Mon Mar 7 19:01:57 2005 +++ prime/lib/makedict/dictformat-wordnet.rb Mon Mar 7 19:01:57 2005 @@ -0,0 +1,44 @@ +# dictformat-wordnet.rb: Converter module for cntlist by WordNet. +# See also: http://www.itri.bton.ac.uk/~Adam.Kilgarriff/bnc-readme.html +# $Id: dictformat-wordnet.rb,v 1.2 2005/03/07 10:01:57 komatsu Exp $ +# +# Copyright (C) 2003 Hiroyuki Komatsu <komat****@taiya*****> +# All rights reserved. +# This is free software with ABSOLUTELY NO WARRANTY. +# +# You can redistribute it and/or modify it under the terms of +# the GNU General Public License version 2. + +## This module is design for the file wordnet-2.0g/dict/lexfiles/cntlist. + +require 'prime/prime-config' +require 'prime/makedict/dictformat' + +class DictFormatWordNet < DictFormat + + def initialize (is_interactive = true) + super(is_interactive) + @pos_table = { + "1" => "noun", + "2" => "verb", + "3" => "adjective", + "4" => "adverb", + "5" => "adjective satellite", + } + end + + def parse (line) + line.chomp! + words = [] + (frequency, sense_key, sense_number) = line.split(/ /) + (literal, pos_id, lex_filenum, lex_id, head_word, head_id) = + sense_key.split(/[%:]/) + pos = @pos_table[pos_id] + if pos != "adjective satellite" then + literal.gsub!("_", " ") + score = frequency.to_i / 5 + words << [literal, pos, literal, score] + end + return words + end +end
