#!/bin/sh

# Set these paths appropriately

BIN=./bin
CMD=./cmd
LIB=./lib

OPTIONS="-token -lemma -sgml"

TOKENIZER=${CMD}/utf8-tokenize.perl
TAGGER=${BIN}/tree-tagger
ABBR_LIST=${LIB}/english-abbreviations
PARFILE=${LIB}/english-utf8.par

PARFILE2=${LIB}/english-chunker-utf8.par
FILTER=${CMD}/filter-chunker-output.perl

$TOKENIZER -e -a $ABBR_LIST $* |
# remove empty lines
grep -v '^$' |
# tagging
$TAGGER $OPTIONS $PARFILE | 
perl -nae 'if ($#F==0){print}else{print "$F[0]-$F[1]\n"}' |
$TAGGER $PARFILE2 -token -sgml -eps 0.00000001 -hyphen-heuristics -quiet |
$FILTER |
$TAGGER $OPTIONS $PARFILE |
perl -pe 's/\tIN\/that/\tIN/;s/\tV[BDHV]/\tVB/'
