classif

#!/bin/bash

#shopt -s extglob

###################################################################
# Lexicon-based detection of depression language
#   - - variable DEPR_DIR sets the PATH for the main folder.
#   - - The system contains a tokenizer but it requires Linguakit to be installed. You can dowload Linguakit from:
#       https://github.com/citiususc/Linguakit/archive/master.zip


#
# ProLNat Group - CiTIUS - USC
###################################################################
###################################################################

############################
# Config
############################

DEPR_DIR="../depression_classification"
PROGS=$DEPR_DIR"/scripts"
LEX=$DEPR_DIR"/lexicons"

LING="en"

############################
# Functions
############################

help()
{
  echo "Syntax: cat input_file |./classif  <module> [PoS_tags]
      
      module= tok or tag | you can just use a tokenizer (tok) or a PoS tagger (tag)
      PoS_tags= N, V, J | if you use module tag, it is possible to classify using a specific tag: noun (N), verb (V), adj (J), or a combination of them: e.g. NJ, VNJ, JV, ... (by default: NVJ).

      example of use: 
      echo "I hate my life" |./classif tag VJ 
       "
  exit
}


# Parámetros obrigatorios
[ $# -lt 1 ] && help
MODULE=$1

# Parámetros opcionais
TAG="NVJ" #by default
[ "$2" != "" ] && TAG=$2

SENT=$PROGS"/sentences_exe.perl"; 
TOK=$PROGS"/tokens_exe.perl";


if [ "$MODULE" == "" ]  ; then
  help;
  echo "help" ;
fi

if [ "$MODULE" == "tok" ] ; then

    tr -d '\015' | $SENT  | $TOK  | $PROGS/select_lex.perl $LEX/lexicon.txt | $PROGS/baseline_types.perl
fi
if [ "$MODULE" == "tag" ] && [ "$TAG" != "" ] ; then
    cat $LEX/lexicon_lem.txt |grep -P "[$TAG]" > $LEX/temp.txt

   tr -d '\015' | $DEPR_DIR/Linguakit-master/linguakit tagger $LING   | $PROGS/select_lex_tagged.perl  $LEX/temp.txt $TAG |  $PROGS/baseline_tagged.perl $TAG
fi