tab-to-sfst-now
929 Bytes
#!/bin/bash
which fst-compiler-utf8 > /dev/null
if [ $? -ne 0 ]
then
echo "Can't find SFST compiler executable (fst-compiler-utf8)."
exit 2
fi
if [ $# -ne 1 ]
then
echo "Usage: `basename $0` tabfile"
echo
echo "tabfile should consist of lines in the following format:"
echo "orth lemma tag"
exit 65
fi
mydir=`dirname $0`
stem=${1%.txt}
echo "Converting data to SFST format. This may take a while..."
$mydir/tab-to-sfst.py $1 $stem.input
echo
echo "Running FST compiler"
# prepare a helper file, containing just one line with quoted filename of the
# generated input file (the generated input is in a simplified format, we just
# include it; the helper file is what SFST wants to read)
echo \"${1%.txt}.input\" > $stem.helper
# have it compiled into a transducer
fst-compiler-utf8 -c $stem.helper $stem.fst
# clean up
rm $stem.input
rm $stem.helper
stat --printf "\nAll done\nOutput: %n (%s bytes)\n" $stem.fst