diff options
| author | mjdenkowski <michael.j.denkowski@gmail.com> | 2013-12-12 15:12:25 -0500 |
|---|---|---|
| committer | mjdenkowski <michael.j.denkowski@gmail.com> | 2013-12-12 15:12:25 -0500 |
| commit | 09b71909b0ecc2d69dfdfb81f535db2519cd5642 (patch) | |
| tree | fd5c57b5fa7f7bea8ce1809026d12b6a18bf5db6 /corpus | |
| parent | 6fb6d64e8191912ed80a652175ccf4e7ce9c0b00 (diff) | |
Restore unbuffered functionality as option
Diffstat (limited to 'corpus')
| -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;}' |