summaryrefslogtreecommitdiff
path: root/word-aligner/classify.pl
diff options
context:
space:
mode:
authorChris Dyer <redpony@gmail.com>2010-02-01 17:38:39 -0500
committerChris Dyer <redpony@gmail.com>2010-02-01 17:38:39 -0500
commitc97b8a8b58f7385fb48b74e2cf1ea9610cd1202f (patch)
tree3bc1b02c39927a810862136534d5a0e35d7ed4fc /word-aligner/classify.pl
parentda222df300e4f87ad185a7decbf119ad56aa34e0 (diff)
word aligner cleanup, new features
Diffstat (limited to 'word-aligner/classify.pl')
-rwxr-xr-xword-aligner/classify.pl27
1 files changed, 0 insertions, 27 deletions
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 $cc = 0;
-while(<C>) {
- chomp;
- my ($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(<T>) {
- chomp;
- my @cats = map { $dict{$_} or die "Undefined class for $_"; } split /\s+/;
- print "@cats\n";
-}
-