summaryrefslogtreecommitdiff
path: root/preprocess
diff options
context:
space:
mode:
authorPatrick Simianer <p@simianer.de>2014-03-16 19:17:24 +0100
committerPatrick Simianer <p@simianer.de>2014-03-16 19:17:24 +0100
commit224be546596aae5fe3bcd9c46d3eac3f84b386fd (patch)
treeb314906965b93b7ec55691d6a22f73b4afad9c60 /preprocess
parent369c3495975ebb21a15aa08063987416e22f6c0c (diff)
no non printables in preproc
Diffstat (limited to 'preprocess')
-rwxr-xr-xpreprocess2
1 files changed, 1 insertions, 1 deletions
diff --git a/preprocess b/preprocess
index 4bf782a..f3c3d10 100755
--- a/preprocess
+++ b/preprocess
@@ -1,5 +1,5 @@
#!/bin/bash
LANG=$1
-/toolbox/scripts/htmlentities 2>htmlentities.$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.$LANG.err
+/toolbox/scripts/no_non_printables | /toolbox/scripts/htmlentities 2>htmlentities.$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