diff options
| author | Patrick Simianer <p@simianer.de> | 2013-12-12 19:41:27 +0100 |
|---|---|---|
| committer | Patrick Simianer <p@simianer.de> | 2013-12-12 19:41:27 +0100 |
| commit | b34ba08c04e4c10aca1879c484ed27119551aa09 (patch) | |
| tree | da861e54d5e5e33e3e6468075125161edce30a06 | |
| parent | 820c06de59307ad06557a603ffee08679fcf3e43 (diff) | |
| parent | 85952929c7f4451a65597b589deade7b6c63403a (diff) | |
Merge branch 'master' of github.com:pks/scripts
| -rwxr-xr-x | preprocess | 2 | ||||
| -rwxr-xr-x | preprocess_nolow | 5 |
2 files changed, 1 insertions, 6 deletions
@@ -1,5 +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.$LANG.err +/toolbox/scripts/htmlentities 2>htmlentities.$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 diff --git a/preprocess_nolow b/preprocess_nolow deleted file mode 100755 index fc466b6..0000000 --- a/preprocess_nolow +++ /dev/null @@ -1,5 +0,0 @@ -#!/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 - |