From db6a6ecfa350cae29739c59df1210d8f76a479c9 Mon Sep 17 00:00:00 2001
From: Patrick Simianer <p@simianer.de>
Date: Thu, 5 Dec 2013 07:56:38 +0100
Subject: init

---
 preprocess | 5 +++++
 1 file changed, 5 insertions(+)
 create mode 100755 preprocess

(limited to 'preprocess')

diff --git a/preprocess b/preprocess
new file mode 100755
index 0000000..716255d
--- /dev/null
+++ b/preprocess
@@ -0,0 +1,5 @@
+#!/bin/zsh
+
+LANG=$1
+~/scripts/htmlentities 2>htmlentities.$LANG.err | ~/scripts/normalize-punctuation 2>normalize-punctuation.$LANG.err | ~/moses/scripts/tokenizer/tokenizer.perl -a -b -threads 1 -l $LANG 2>tokenizer.$LANG.err | ~/moses/scripts/tokenizer/lowercase.perl 2>lowercase.$LANG.err
+
-- 
cgit v1.2.3