From 3bbbd92fc7b1f9d0d61f8573fee0d0b2f40960d3 Mon Sep 17 00:00:00 2001
From: Patrick Simianer <p@simianer.de>
Date: Mon, 17 Mar 2014 13:13:09 +0100
Subject: a lot of ... and --- cause moses' compound splitter to hang

---
 preprocess | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'preprocess')

diff --git a/preprocess b/preprocess
index f3c3d10..69eaa54 100755
--- a/preprocess
+++ b/preprocess
@@ -1,5 +1,5 @@
 #!/bin/bash
 
 LANG=$1
-/toolbox/scripts/no_non_printables | /toolbox/scripts/htmlentities 2>htmlentities.$LANG.err | /toolbox/scripts/normalize_punctuation 2>normalize-punctuation.$LANG.err | /toolbox/moses/scripts/tokenizer/tokenizer.no-escape.perl -a -b -threads 1 -l $LANG 2>tokenizer.$LANG.err | /toolbox/moses/scripts/tokenizer/lowercase.perl 2>lowercase.$LANG.err
+/toolbox/scripts/no_non_printables | sed "s|[-,\.]\{4,\}|...|" | /toolbox/scripts/htmlentities 2>htmlentities.$LANG.err | /toolbox/scripts/normalize_punctuation 2>normalize-punctuation.$LANG.err | /toolbox/moses/scripts/tokenizer/tokenizer.no-escape.perl -a -b -threads 1 -l $LANG 2>tokenizer.$LANG.err | /toolbox/moses/scripts/tokenizer/lowercase.perl 2>lowercase.$LANG.err
 
-- 
cgit v1.2.3