From bae5fe99037ae7e101953ad0df118127191c711c Mon Sep 17 00:00:00 2001
From: Chris Dyer <cdyer@allegro.clab.cs.cmu.edu>
Date: Tue, 15 Jan 2013 01:20:00 -0500
Subject: corpus files

---
 corpus/filter-length.pl | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

(limited to 'corpus/filter-length.pl')

diff --git a/corpus/filter-length.pl b/corpus/filter-length.pl
index 70032ca7..3cfa40cc 100755
--- a/corpus/filter-length.pl
+++ b/corpus/filter-length.pl
@@ -3,8 +3,8 @@ use strict;
 use utf8;
 
 ##### EDIT THESE SETTINGS ####################################################
-my $MAX_LENGTH = 99;  # discard a sentence if it is longer than this
-my $AUTOMATIC_INCLUDE_IF_SHORTER_THAN = 6; # if both are shorter, include
+my $MAX_LENGTH = 150;  # discard a sentence if it is longer than this
+my $AUTOMATIC_INCLUDE_IF_SHORTER_THAN = 7; # if both are shorter, include
 my $MAX_ZSCORE = 1.8; # how far from the mean can the (log)ratio be?
 ##############################################################################
 
@@ -128,6 +128,8 @@ while(<F>) {
       next;
     }
     print;
+  } else {
+    print;
   }
   $to++;
 }
-- 
cgit v1.2.3