#!/bin/bash -l
# tisha Melia
# June 29, 2015
# All lines starting with "#$" are SGE qsub commands
# qsub -N mergeCuffAll -P waxmanlab -pe single_node 2-8 cuffmerge.qsub prefix input_file output_folder
#-N refer to name of output folder, prefix- name of your gtf file result, input:samples.txt, output_folder: /rest../../varsha/cufflinks/cuffmerg
# Specify which shell to use
#$ -S /bin/bash
# Run on the current working directory
#$ -cwd

# Join standard output and error to a single file
#$ -j y

# Send an email when the job begins and when it ends running


# Whom to send the email to



if [ $# -lt 3 ] ; then
      echo "Need at least 3 line arguments"
      echo "qsub -N name -P waxmanlab -pe single_node 2-8 cuffmerge.qsub [prefix] [input_file] [output_folder]"
      exit 0
fi

# Now let's keep track of some information just in case anything goes wrong
echo "=========================================================="
echo "Starting on : $(date)"
echo "Running on node : $(hostname)"
echo "Current directory : $(pwd)"
echo "Current job ID : $JOB_ID"
echo "Current job name : $JOB_NAME"
echo "Task index number : $TASK_ID"
echo "=========================================================="


# Go to local scratch directory
cd ${TMPDIR}
ls 
echo $TMPDIR

# print out some diagnostic stuff
echo Running on host `hostname`
echo
echo Directory is `pwd`
echo
echo Start time is `date`
echo


module load boost/1.69.0
module load samtools/0.1.19
module load bowtie2/2.3.4.1
module load cufflinks/2.2.1
module load python2/2.7.16




SAMPLE_PREFIX=$1
INPUT_FILE_LIST=$2
RESULT_CUFFMERGE_DIR=$3 


IFS=$'\t'

while read SAMPLE_NAME INPUT_FILE; do
	if [[ "$SAMPLE_NAME" =~ \#.* ]];then
		echo "skipped comment line"
	else
		
		echo $SAMPLE_NAME
		echo $INPUT_FILE
		echo "cp $INPUT_FILE ${SAMPLE_NAME}.gtf"
		cp $INPUT_FILE ${SAMPLE_NAME}.gtf
		echo "${SAMPLE_NAME}.gtf" >> assemblies.txt
       fi
done < $INPUT_FILE_LIST





#printing assemblies.txt
cat assemblies.txt
BASE_AUX_DIR="/net/waxman-server/mnt/data/waxmanlabvm_home/kkarri/varsha/cufflinks/cuffmerge_withkk/new"
cp ${BASE_AUX_DIR}/genomeWithRandom.fa . #exactly the same as the previous one
cp ${BASE_AUX_DIR}/genes.gtf .


echo "Finished copying..."
echo "Current dir:"
pwd

# run my commands
echo
echo "running commands..."
echo
ls 



OUTPUT_DIR="${SAMPLE_PREFIX}_cuffmerge"
cuffmerge -o ${OUTPUT_DIR} -p $NSLOTS -g genes.gtf -s genomeWithRandom.fa assemblies.txt
echo "cuffmerge -o ${OUTPUT_DIR} -p $NSLOTS -g genes.gtf -s genomeWithRandom.fa assemblies.txt"

#Usage:
#    cuffmerge [Options] <assembly_GTF_list.txt>
#
#Options:
#    -h/--help                               Prints the help message and exits
#    -o                     <output_dir>     Directory where merged assembly will be written  [ default: ./merged_asm  ]
#    -g/--ref-gtf                            An optional "reference" annotation GTF.
#    -s/--ref-sequence      <seq_dir>/<seq_fasta> Genomic DNA sequences for the reference.
#    --min-isoform-fraction <0-1.0>          Discard isoforms with abundance below this       [ default:           0.05 ]
#    -p/--num-threads       <int>            Use this many threads to merge assemblies.       [ default:             1  ]
#    --keep-tmp                              Keep all intermediate files during merge


# copy the output files to users storage dir
cp -r ${OUTPUT_DIR} ${RESULT_CUFFMERGE_DIR}/.

echo "done running. copying..."
#Just list everything in scratch dir (Don't know what output files I need)
echo "List of files in scratch"
ls *
#Also want the file size of everything
echo "Size of files in scratch"
du -ch *

# print out some diagnostic stuff
echo Stop time is `date`