#!/bin/bash

#SBATCH --job-name=toe
#SBATCH --account=nn9106k
#SBATCH --time=72:00:00
#SBATCH --mail-type=FAIL
#SBATCH --nodes=1
#SBATCH --ntasks-per-node=6
#SBATCH --mem-per-cpu=2048M

if [ -n "${SLURM_JOB_ID}" ]; then
  . /cluster/bin/jobsetup;
  if [ -n "${SCRATCH}" ]; then
    export LAPTMP=${SCRATCH}/tmp;
    mkdir -p ${LAPTMP} > /dev/null 2>&1;
    export TMPDIR=${LAPTMP};
  fi
fi

if [ -z "${1}" ]; then
  echo "toe: missing input file argument; exit."
  exit 1;
fi

if [ -z "${LAPTREE}" ]; then
  LAPTREE=/projects/lap/development/trunk/tree;
fi
. ${LAPTREE}/etc/library.bash;

base=$(dirname ${1});

for name in $(cat ${1}); do
  input=${base}/txt/${name}.txt;
  output=${base}/rdf/${name}.ttl;
  if [ ! -f ${input} ]; then
    echo "toe: skipping invalid file ‘${input}’; continue."
    continue;
  fi

  echo "toe: processing ‘${input}’ ...";
  ${LAPTREE}/etc/driver ==tool import ==process text \
    ${input} ${LAPTMP}/media.${USER}.${$}.rpt;
  ${LAPTREE}/etc/driver ==tool tokenizer \
    ${LAPTMP}/media.${USER}.${$}.rpt ${LAPTMP}/tokenizer.${USER}.${$}.rpt;
  ${LAPTREE}/etc/driver ==tool repp \
    --segmenter tokenizer --style repp \
    ${LAPTMP}/tokenizer.${USER}.${$}.rpt ${LAPTMP}/repp.${USER}.${$}.rpt
  ${LAPTREE}/etc/driver ==tool bn \
    --segmenter tokenizer --tokenizer repp \
    ${LAPTMP}/repp.${USER}.${$}.rpt ${LAPTMP}/bn.${USER}.${$}.rpt
  ${LAPTREE}/etc/driver ==tool export ==process rdf \
    ${LAPTMP}/bn.${USER}.${$}.rpt ${output};
  #
  # in case an intermediate processing step failed, try deleting all receipts
  #
  ${LAPLIBRARY}/bin/store --purge ${LAPTMP}/bn.${USER}.${$}.rpt;
  ${LAPLIBRARY}/bin/store --purge ${LAPTMP}/repp.${USER}.${$}.rpt;
  ${LAPLIBRARY}/bin/store --purge ${LAPTMP}/tokenizer.${USER}.${$}.rpt;
  ${LAPLIBRARY}/bin/store --purge ${LAPTMP}/media.${USER}.${$}.rpt;
  echo -n "output: "
  /bin/ls -l "${output}";

done