From c97b8a8b58f7385fb48b74e2cf1ea9610cd1202f Mon Sep 17 00:00:00 2001 From: Chris Dyer Date: Mon, 1 Feb 2010 17:38:39 -0500 Subject: word aligner cleanup, new features --- word-aligner/classify.pl | 27 --------------------------- 1 file changed, 27 deletions(-) delete mode 100755 word-aligner/classify.pl (limited to 'word-aligner/classify.pl') diff --git a/word-aligner/classify.pl b/word-aligner/classify.pl deleted file mode 100755 index 893c7b22..00000000 --- a/word-aligner/classify.pl +++ /dev/null @@ -1,27 +0,0 @@ -#!/usr/bin/perl -w -use strict; - -die "Usage: $0 classes.txt corpus.txt" unless scalar @ARGV == 2; - -my ($class, $text) = @ARGV; -open C, "<$class" or die "Can't read $class: $!"; -open T, "<$text" or die "Can't read $text: $!"; - -my %dict = (); -my $cc = 0; -while() { - chomp; - my ($word, $cat) = split /\s+/; - die "'$word' '$cat'" unless (defined $word && defined $cat); - $dict{$word} = $cat; - $cc++; -} -close C; -print STDERR "Loaded classes for $cc words\n"; - -while() { - chomp; - my @cats = map { $dict{$_} or die "Undefined class for $_"; } split /\s+/; - print "@cats\n"; -} - -- cgit v1.2.3