#!/bin/sh

#
# Script to bundle goby and required data into a
# suitable format for submission to a SGE queue
#

# Absolute path to this script.
SCRIPT=$(readlink -f $0)
# Absolute path this script is in.
SCRIPT_DIR=$(dirname $SCRIPT)
# Absolute path to the SGE transcript scripts
TRANSCRIPT_SCRIPT_DIR=${SCRIPT_DIR}/transcript/templates

if [ -z $1 ]; then
    echo "Job name is required"
    exit 1
fi

. sge-env.sh

# Local goby locations
GOBY_DIR=${GOBY_DIR:-$SCRIPT_DIR/../..}

if [ ! -e ${GOBY_DIR}/goby.jar ]; then
    echo "goby.jar not found!"
    exit 2
fi

# The job name identifies the location for the scripts
JOB_TAG=$1
JOB_DIR=$(readlink -f .)/${JOB_TAG}
JOB_RESULTS_DIR=${JOB_RESULTS_DIR:-$(readlink -f .)/${JOB_TAG}-results}

if [ -e ${JOB_DIR} ]; then
    echo "The job output directory already exists"
    exit 3
fi

if [ ! -r ${TRANSCRIPT_DIRECTORY} ]; then
    echo "ERROR: Transcript directory ${TRANSCRIPT_DIRECTORY} cannot be read"
    exit 4
fi

/bin/mkdir -p ${JOB_DIR}

if [ -z ${READS} ] || [ ! -r ${READS} ]; then
    echo "WARNING: Reads ${READS} file cannot be read"
fi

ls -1 ${TRANSCRIPT_DIRECTORY} > ${JOB_DIR}/transcript-list.txt
NUMBER_OF_TRANSCRIPTS=`wc -l < ${JOB_DIR}/transcript-list.txt`
echo "Found ${NUMBER_OF_TRANSCRIPTS} transcript files"

echo "Bundling job submission files"

SGE_ARRAY_DIRECTIVE="#$ -t 1-${NUMBER_OF_TRANSCRIPTS}"
echo "Alignment will run as ${NUMBER_OF_TRANSCRIPTS} jobs"

if [ ! -z ${SGE_STATUS_MAILTO} ]; then
    SGE_MAILTO_DIRECTIVE="#$ -M ${SGE_STATUS_MAILTO}"
fi

# Use a basename that is based on name of the reads file
BASENAME=$(basename $READS .compact-reads)

# Create a somewhat unique tag
# ("stage" of job will be added - index, align, concat)
SGE_JOB_NAME=${JOB_TAG}

# Copy goby and submission scripts to the run directory
/bin/cp ${GOBY_DIR}/goby.jar ${GOBY_DIR}/config/log4j.properties \
        ${SCRIPT_DIR}/sge-env.sh ${JOB_DIR}

# Create job specific scripts from the template files
for FILE in ${TRANSCRIPT_SCRIPT_DIR}/*; do
    FILENAME=$(basename ${FILE})
    sed -e "s|%TRANSCRIPT_DIRECTORY%|${TRANSCRIPT_DIRECTORY}|" \
        -e "s|%NUMBER_OF_TRANSCRIPTS%|${NUMBER_OF_TRANSCRIPTS}|" \
        -e "s|%TRANSCRIPT_INDEX_DIRECTORY%|${TRANSCRIPT_INDEX_DIRECTORY}|" \
        -e "s|%GENE_TRANSCRIPT_MAP_FILE%|${GENE_TRANSCRIPT_MAP_FILE}|" \
        -e "s|%READS%|${READS}|" \
        -e "s|%SGE_QUEUE%|${SGE_QUEUE}|" \
        -e "s|%SGE_MEMORY%|${SGE_MEMORY}|" \
        -e "s|%SGE_JVM_FLAGS%|${SGE_JVM_FLAGS}|" \
        -e "s|%SGE_ARRAY_DIRECTIVE%|${SGE_ARRAY_DIRECTIVE}|" \
        -e "s|%SGE_MAILTO_DIRECTIVE%|${SGE_MAILTO_DIRECTIVE}|" \
        -e "s|%ALIGNER%|${ALIGNER}|" \
        -e "s|%COLORSPACE%|${COLORSPACE}|" \
        -e "s|%CHUNK_SIZE%|${CHUNK_SIZE}|" \
        -e "s|%BASENAME%|${BASENAME}|" \
        -e "s|%BWA_ALIGNER_PATH%|${BWA_ALIGNER_PATH}|" \
        -e "s|%LAST_ALIGNER_PATH%|${LAST_ALIGNER_PATH}|" \
        -e "s|%LASTAG_ALIGNER_PATH%|${LASTAG_ALIGNER_PATH}|" \
        -e "s|%SGE_JOB_NAME%|${SGE_JOB_NAME}|" \
        ${FILE} > ${JOB_DIR}/${FILENAME}
done

chmod +x ${JOB_DIR}/*.sh

echo "Scripts were written to ${JOB_DIR}"