autostart_gbs_qc

#!/bin/sh

POLL_INTERVAL=900      # seconds - i.e. 15 minutes 

function read_answer_with_default() {
   read answer
   if [ -z "$answer" ]; then
      answer=$@
   fi
}

function get_opts() {
   help_text="
This is used to auto-start a run. It will poll for a landmark file indicating upload is complete then start the processing\n
\n
usage example (currently run in a screen session) :\n
\n
autostart_gbs_qc 200511_D00390_0545_BHFMY3BCX3 2>&1 | tee /dataset/2024_illumina_sequencing_e/scratch/postprocessing/autostart_logs/200511_D00390_0545_BHFMY3BCX3.log \n
"
   while getopts ":h" opt; do
   case $opt in
       h)
         echo -e $help_text
         exit 0
         ;;
       \?)
         echo "Invalid option: -$OPTARG" >&2
         exit 1
         ;;
       :)
         echo "Option -$OPTARG requires an argument." >&2
         exit 1
         ;;
     esac
   done

   shift $((OPTIND-1))
   
   if [ -z "$1" ]; then
      echo "must supply a run name "
      exit 1
   fi

   ARGRUN=$1
}


function get_pipeline() {
   gbs_version=$1
   RUN=$2 

   export SEQ_PRISMS_BIN=/dataset/gseq_processing/active/bin/gbs_prism/seq_prisms
   export GBS_PRISM_BIN=/dataset/gseq_processing/active/bin/gbs_prism
   cd $GBS_PRISM_BIN

   # (to move to another KGD version, git pull  in the KGD folder, and then check out the version you want)
   kgd_version=`$GBS_PRISM_BIN/get_kgd_version.sh`
   echo "running gbs qc pipeline version $gbs_version (KGD version $kgd_version)"
   is_alpha=`echo $gbs_version | grep alpha`
   if [ ! $is_alpha ]; then 
      git checkout -q $gbs_version 
      if [ $? != 0 ]; then
         echo "unable to checkout pipeline version $gbs_version"
         exit 1
      fi
      echo "running exec ./_run_gbs_qc -r $RUN $gbs_version $kgd_version"
      send_mail
      exec ./_run_gbs_qc -r $RUN $gbs_version $kgd_version
   else
      echo "(no checkout for alpha versions, just running in current branch)"
      echo "running exec ./_run_gbs_qc -r $RUN $gbs_version $kgd_version"
      send_mail
      exec ./_run_gbs_qc -r $RUN $gbs_version $kgd_version
   fi
}

function send_mail() {
   echo "sending mail to vanstijnt , mccullocha, bairdh, perrybe, andersonr, andrewsa, henryh, frenchm, hicklandm "
   echo "" | mutt -s "FYI - looks like upload of $RUN is complete so auto-starting processing" vanstijnt , mccullocha, bairdh, perrybe, andersonr, andrewsa, henryh, frenchm, hicklandm
}

function get_landmark() {
   RUN=$1
   landmark=""
   if [ -f /dataset/2024_illumina_sequencing_e/active/$RUN/RTAComplete.txt ]; then
      landmark=/dataset/2024_illumina_sequencing_e/active/$RUN/RTAComplete.txt
   fi
}

function get_digest() {
   RUN=$1
   digest=`ls -lR /dataset/2024_illumina_sequencing_e/active/$RUN/ | md5sum -b `
   echo $digest
}


function poll_for() {
   RUN=$1 

   # polls for the run landmark 
   poll_count=0
   while [ 1 ]; do
      get_landmark $RUN
      get_digest $RUN
      if [ ! -z "$landmark" ]; then
         if [ $poll_count == 0 ]; then
            echo "*** warning - landmark $landmark is already there ! *** "
            echo "(final digest : "
            get_digest $RUN
            echo ")"
            get_pipeline 1.0.4alpha $RUN
         else
            echo "found landmark ($landmark) - starting processing"
            echo "(final digest : "
            get_digest $RUN
            echo ")"
            get_pipeline 1.0.4alpha $RUN
         fi
      else
         let poll_count=${poll_count}+1
         echo "sleeping for $POLL_INTERVAL (poll count = $poll_count )"
         sleep $POLL_INTERVAL
      fi
   done
}

get_opts "$@"
poll_for $ARGRUN