diff options
| author | Patrick Simianer <p@simianer.de> | 2014-01-13 12:37:45 +0100 |
|---|---|---|
| committer | Patrick Simianer <p@simianer.de> | 2014-01-13 12:37:45 +0100 |
| commit | a2f803d5b629f31cf41313d77e382150dd007d39 (patch) | |
| tree | c56130c1112e803058b33b98853cff08a8f8450d /corpus/tokenize-anything.sh | |
| parent | 926fb52bfc85dcd58156916ca2536ee32c719954 (diff) | |
| parent | c148f8429c66103a401ba4c3a029e349cd11aa8a (diff) | |
Merge remote-tracking branch 'upstream/master'
Diffstat (limited to 'corpus/tokenize-anything.sh')
| -rwxr-xr-x | corpus/tokenize-anything.sh | 6 |
1 files changed, 4 insertions, 2 deletions
diff --git a/corpus/tokenize-anything.sh b/corpus/tokenize-anything.sh index c7adfa61..5b7933d8 100755 --- a/corpus/tokenize-anything.sh +++ b/corpus/tokenize-anything.sh @@ -5,15 +5,17 @@ SUPPORT=$ROOTDIR/support if [[ $# == 1 && $1 == '-u' ]] ; then NORMARGS="--batchline" + SEDFLAGS="-u" else NORMARGS="" + SEDFLAGS="" fi $SUPPORT/utf8-normalize.sh $NORMARGS | $SUPPORT/quote-norm.pl | $SUPPORT/tokenizer.pl | - sed -e 's/ al - / al-/g' | + sed $SEDFLAGS -e 's/ al - / al-/g' | $SUPPORT/fix-contract.pl | - sed -e 's/^ //' | sed -e 's/ $//' | + sed $SEDFLAGS -e 's/^ //' | sed $SEDFLAGS -e 's/ $//' | perl -e '$|++; while(<>){s/(\d+)(\.+)$/$1 ./; s/(\d+)(\.+) \|\|\|/$1 . |||/; print;}' |