summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorPatrick Simianer <p@simianer.de>2014-03-17 13:13:09 +0100
committerPatrick Simianer <p@simianer.de>2014-03-17 13:13:09 +0100
commit3bbbd92fc7b1f9d0d61f8573fee0d0b2f40960d3 (patch)
tree717a12c52c2409ba7d540738bd70beeedad1930a
parent4753aee16960de5b9c3da926dde18b332b34edf4 (diff)
a lot of ... and --- cause moses' compound splitter to hang
-rwxr-xr-xpreprocess2
1 files changed, 1 insertions, 1 deletions
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