From c97b8a8b58f7385fb48b74e2cf1ea9610cd1202f Mon Sep 17 00:00:00 2001
From: Chris Dyer <redpony@gmail.com>
Date: Mon, 1 Feb 2010 17:38:39 -0500
Subject: word aligner cleanup, new features

---
 word-aligner/extract_vocab.pl | 20 --------------------
 1 file changed, 20 deletions(-)
 delete mode 100755 word-aligner/extract_vocab.pl

(limited to 'word-aligner/extract_vocab.pl')

diff --git a/word-aligner/extract_vocab.pl b/word-aligner/extract_vocab.pl
deleted file mode 100755
index 070d4202..00000000
--- a/word-aligner/extract_vocab.pl
+++ /dev/null
@@ -1,20 +0,0 @@
-#!/usr/bin/perl -w
-use strict;
-
-print STDERR "Extracting vocabulary...\n";
-my %dict = ();
-my $wc = 0;
-while(<>) {
-  chomp;
-  my @words = split /\s+/;
-  for my $word (@words) { $wc++; $dict{$word}++; }
-}
-
-my $tc = 0;
-for my $word (sort {$dict{$b} <=> $dict{$a}} keys %dict) {
-  print "$word\n";
-  $tc++;
-}
-
-print STDERR "$tc types / $wc tokens\n";
-
-- 
cgit v1.2.3