source: trunk/libIGCM/AA_pack_output @ 1516

#-Q- curie ######################
#-Q- curie ## CURIE   TGCC/CEA ##
#-Q- curie ######################
#-Q- curie #MSUB -r PACKOUTPUT     # Nom du job
#-Q- curie #MSUB -o PACKOUTPUT.out_%I
#-Q- curie #MSUB -e PACKOUTPUT.out_%I
#-Q- curie #MSUB -n 1              # Reservation du processus
#-Q- curie #MSUB -T 36000          # Limite de temps elapsed du job
#-Q- curie #MSUB -q ::default_node::
#-Q- curie #MSUB -c ::default_core::
#-Q- curie #MSUB -Q normal
#-Q- curie #MSUB -A ::default_project::
#-Q- curie set +x
#-Q- irene ######################
#-Q- irene ## IRENE   TGCC/CEA ##
#-Q- irene ######################
#-Q- irene #MSUB -r PACKOUTPUT     # Job name
#-Q- irene #MSUB -o PACKOUTPUT.out_%I
#-Q- irene #MSUB -e PACKOUTPUT.out_%I
#-Q- irene #MSUB -n 1              # Number of cores
#-Q- irene #MSUB -T 36000          # Maximum elapsed time
#-Q- irene #MSUB -q ::default_node::
#-Q- irene #MSUB -c ::default_core::
#-Q- irene #MSUB -Q normal
#-Q- irene #MSUB -A ::default_post_project::
#-Q- irene #MSUB -m store,work,scratch
#-Q- irene set +x
#-Q- jeanzay #!/bin/ksh 
#-Q- jeanzay ######################
#-Q- jeanzay ## JEANZAY    IDRIS ##
#-Q- jeanzay ######################
#-Q- jeanzay #SBATCH --job-name=PACKOUTPUT         # Job Name
#-Q- jeanzay #SBATCH --output=PACKOUTPUT.out_%J    # standard output
#-Q- jeanzay #SBATCH --error=PACKOUTPUT.out_%J     # error output
#-Q- jeanzay #SBATCH -N  1                        # Number of core
#-Q- jeanzay #SBATCH --partition=prepost          # Post-processing partition
#-Q- jeanzay #SBATCH --time=10:00:00               # Wall clock limit (seconds)
#-Q- jeanzay #SBATCH --account ::default_project::@cpu
#-Q- jeanzay set +x
#-Q- ada #!/bin/ksh
#-Q- ada #######################
#-Q- ada ## ADA         IDRIS ##
#-Q- ada #######################
#-Q- ada # @ job_type = mpich
#-Q- ada # @ requirements = (Feature == "prepost")
#-Q- ada # Temps Elapsed max. d'une requete hh:mm:ss
#-Q- ada # @ wall_clock_limit = 10:00:00
#-Q- ada # Memory required for ncrcat
#-Q- ada # @ as_limit = 30Gb
#-Q- ada # Nom du travail LoadLeveler
#-Q- ada # @ job_name   = PACKOUTPUT
#-Q- ada # Fichier de sortie standard du travail
#-Q- ada # @ output     = $(job_name).$(jobid)
#-Q- ada # Fichier de sortie d'erreur du travail
#-Q- ada # @ error      =  $(job_name).$(jobid)
#-Q- ada # pour recevoir un mail en cas de depassement du temps Elapsed (ou autre pb.)
#-Q- ada # @ notification = error
#-Q- ada # @ environment  = $DEBUG_debug ; $BigBrother ; $postProcessingStopLevel ; $MODIPSL ; $libIGCM ; $libIGCM_SX ; $POST_DIR ; $Script_Post_Output ; $SUBMIT_DIR ; $DateBegin ; $DateEnd ; $PeriodPack ; $StandAlone ; $MASTER ; wall_clock_limit=$(wall_clock_limit)
#-Q- ada # @ queue
#-Q- lxiv8 ######################
#-Q- lxiv8 ## OBELIX      LSCE ##
#-Q- lxiv8 ######################
#-Q- lxiv8 #PBS -N PACKOUTPUT
#-Q- lxiv8 #PBS -m a
#-Q- lxiv8 #PBS -j oe
#-Q- lxiv8 #PBS -q medium
#-Q- lxiv8 #PBS -o PACKOUTPUT.$$
#-Q- lxiv8 #PBS -S /bin/ksh
#-Q- ifort_CICLAD ######################
#-Q- ifort_CICLAD ##   CICLAD    IPSL ##
#-Q- ifort_CICLAD ######################
#-Q- ifort_CICLAD #PBS -N PACKOUTPUT
#-Q- ifort_CICLAD #PBS -m a
#-Q- ifort_CICLAD #PBS -j oe
#-Q- ifort_CICLAD #PBS -q std
#-Q- ifort_CICLAD #PBS -S /bin/ksh
#-Q- default #!/bin/ksh
#-Q- default ##################
#-Q- default ## DEFAULT HOST ##
#-Q- default ##################

#**************************************************************
# Author: Sebastien Denvil
# Contact: Sebastien.Denvil__at__ipsl.jussieu.fr
# $Revision::                                          $ Revision of last commit
# $Author::                                            $ Author of last commit
# $Date::                                              $ Date of last commit
# IPSL (2006)
#  This software is governed by the CeCILL licence see libIGCM/libIGCM_CeCILL.LIC
#
#**************************************************************

#set -eu
#set -vx

date

#D- Task type DO NOT CHANGE (computing, post-processing or checking)
TaskType=post-processing

########################################################################

#D- Flag to determine if this job in a standalone mode
#D- Default : value from AA_job if any
StandAlone=${StandAlone:=true}

#D- Path to libIGCM
#D- Default : value from AA_job if any
# WARNING For StandAlone use : To run this script on some machine (ulam and cesium)
# WARNING you must check MirrorlibIGCM variable in sys library.
# WARNING If this variable is true, you must use libIGCM_POST path instead
# WARNING of your running libIGCM directory.
libIGCM=${libIGCM:=::modipsl::/libIGCM}

#-D- $hostname of the MASTER job when SUBMIT_DIR is not visible on postprocessing computer.
MASTER=${MASTER:=ada|curie}

#D- Flag to determine begin date for restart pack
#D- Default : value from AA_job if any
DateBegin=${DateBegin:=20000101}

#D- Flag to determine end date for restart pack
#D- Default : value from AA_job if any
DateEnd=${DateEnd:=20691231}

#D- Flag to determine pack period
#D- Default : value from AA_job if any
PeriodPack=${PeriodPack:=10Y}

#D- Uncomment to run interactively
#D- For testing purpose, will be remove
#SUBMIT_DIR=${PWD}
#RUN_DIR_PATH=${SCRATCHDIR}/Pack_Test

#D- Increased verbosity (1, 2, 3)
#D- Default : value from AA_job if any
Verbosity=${Verbosity:=3}

#D- Low level debug : to bypass lib test checks and stack construction
#D- Default : value from AA_job if any
DEBUG_debug=${DEBUG_debug:=false}

########################################################################

. ${libIGCM}/libIGCM_debug/libIGCM_debug.ksh
. ${libIGCM}/libIGCM_card/libIGCM_card.ksh
. ${libIGCM}/libIGCM_date/libIGCM_date.ksh
#-------
. ${libIGCM}/libIGCM_sys/libIGCM_sys.ksh
. ${libIGCM}/libIGCM_config/libIGCM_config.ksh
. ${libIGCM}/libIGCM_post/libIGCM_post.ksh
#-------
RUN_DIR=${RUN_DIR_PATH}
IGCM_sys_MkdirWork ${RUN_DIR}
IGCM_sys_Cd ${RUN_DIR}
#-------
( ${DEBUG_debug} ) && IGCM_debug_Check
( ${DEBUG_debug} ) && IGCM_card_Check
( ${DEBUG_debug} ) && IGCM_date_Check

########################################################################

#set -vx

# ------------------------------------------------------------------
# Test if all was right before proceeding further
# ------------------------------------------------------------------
IGCM_debug_Verif_Exit

if [ ${StandAlone} = true ] ; then
    CARD_DIR=${SUBMIT_DIR}
else
    CARD_DIR=${RUN_DIR_PATH}
    IGCM_sys_Get_Master ${SUBMIT_DIR}/config.card ${RUN_DIR_PATH}
    IGCM_sys_Get_Master ${SUBMIT_DIR}/run.card    ${RUN_DIR_PATH}
    IGCM_sys_Get_Master ${SUBMIT_DIR}/COMP        ${RUN_DIR_PATH}
    IGCM_sys_Get_Master ${SUBMIT_DIR}/POST        ${RUN_DIR_PATH}
fi

#==================================
# First of all
#
# Read libIGCM compatibility version in config.card
# Read UserChoices section
# Read Ensemble section
# Read Post section
# Define all netcdf output directories
#==================================
IGCM_config_CommonConfiguration ${CARD_DIR}/config.card

# ------------------------------------------------------------------
# Activate BigBrother so as to supervise this job
# ------------------------------------------------------------------
IGCM_debug_BigBro_Initialize

#==================================
# Read ListOfComponents section
# to drive the loop over find
IGCM_card_DefineArrayFromSection ${CARD_DIR}/config.card ListOfComponents

#==================================
# Test and set up directories
#==================================
IGCM_sys_TestDirArchive ${R_SAVE}
[ $? != 0 ] && IGCM_debug_Exit "IGCM_sys_TestDirArchive"

# Where to store used file list /!\ TEMPORARY /!\
STORE_DEBUG=${R_SAVE}/DEBUG

# Switch to script variables meaning (try to be compatible with ipsl_pack TGCC moving procedure)
JobName=${config_UserChoices_JobName}
echo $JobName $DateBegin $DateEnd

# ------------------------------------------------------------------
# Test if all was right before proceeding further
# ------------------------------------------------------------------
IGCM_debug_Verif_Exit

IGCM_debug_Print 1 "Check coherence between PackFrequency and PeriodLength"
IGCM_post_CheckModuloFrequency PeriodPack config_UserChoices_PeriodLength NbPeriodPerFrequency
# ------------------------------------------------------------------
# Test if all was right before proceeding further
# ------------------------------------------------------------------
IGCM_debug_Verif_Exit

IGCM_debug_Print 1 "We must process ${NbPeriodPerFrequency} files for each pack"

# Init loop
date_begin_pack=${DateBegin}
date_end_simulation=${DateEnd}
number_pack=1

IGCM_debug_PrintVariables 3 date_begin_pack
IGCM_debug_PrintVariables 3 date_end_simulation

while [ ${date_begin_pack} -le ${date_end_simulation} ] ; do

  IGCM_debug_PrintVariables 3 number_pack
  DaysTemp=$( IGCM_date_DaysInCurrentPeriod ${date_begin_pack} ${PeriodPack} )
  date_end_pack=$( IGCM_date_AddDaysToGregorianDate ${date_begin_pack} $(( ${DaysTemp} - 1 )) )

  for comp in ${config_ListOfComponents[*]} ; do
    dirList=$( find ${R_BUFR}/${comp}/Output -maxdepth 1 -mindepth 1 -type d )
    for dir in ${dirList} ; do
      # dirID is like ATM.Output.MO
      dirID=$( echo $dir | sed "s:${R_BUFR}/::" | sed "s:/:.:g" )
      # Sort what's in the directory
      find ${dir} -type f -name "${JobName}*.nc" -ls | sort -k 11 > liste_files.${dirID}.txt
      # How much file type. Example : 1M_histmthCOSP.nc, 1M_histmth.nc, 1M_histmthNMC.nc, 1M_paramLMDZ_phy.nc
      # /!\ fileType include the .nc extension /!\
      fileType=$( gawk '{print $11}' liste_files.${dirID}.txt | gawk -F$dir/ '{print $2}' | sed "s:${JobName}_[0-9]\{8,9\}_[0-9]\{8,9\}_::g" | sort | uniq )
      # Loop over the file type and pack them when in between date_begin_pack and date_end_pack
      for myType in ${fileType} ; do
        grep ${myType} liste_files.${dirID}.txt > liste_files.${dirID}.${myType}.txt
        nbfile=0
        for file in $( gawk '{print $11}' liste_files.${dirID}.${myType}.txt ); do
          extract_date_file=$( echo ${file}  | sed -e "s/.*${JobName}_[0-9]*_//" )
          date_file=$( echo ${extract_date_file} | sed 's/\([0-9]\{8\}\)_.*$/\1/g' )
          # echo pack n°${number_pack}  ${date_file} ${date_begin_pack} ${date_end_pack}
          if [ ${date_file} -le ${date_end_pack} ] && [ ${date_file} -ge ${date_begin_pack} ] ; then
            echo ${file} >> liste_pack_${myType}_${date_begin_pack}_${date_end_pack}.txt
            ncdump -h ${file} | grep -E 'float|double' | cut -f 1 -d '(' | cut -f 2 -d ' ' >> liste_variables_${myType}_${date_begin_pack}_${date_end_pack}.txt
            (( nbfile = nbfile + 1 ))
          fi
        done

        if [ ${nbfile} = 0 ] ; then
          IGCM_debug_Print 1 "We found no file to process"
          IGCM_debug_Print 1 "We should have found ${NbPeriodPerFrequency} files"
          IGCM_debug_Print 1 "As some files can be produced only for some selected period we consider we can move to the next file type"
          continue
        fi

        # Select list of variables to work with
        list_var=$( cat liste_variables_${myType}_${date_begin_pack}_${date_end_pack}.txt | sort | uniq -c | awk -v nbfile=$nbfile '{if ($1 != nbfile) {print $2}}' | paste -s -d ',' )
        liste_file_tmp=$( for i in $( cat liste_pack_${myType}_${date_begin_pack}_${date_end_pack}.txt ) ; do basename $i ; done )
        # Create packed files
        IGCM_debug_Print 1 "Ncrcat ongoing for ${dir} and ${myType}"
        if [ ! ${nbfile} = ${NbPeriodPerFrequency} ] ; then
          IGCM_debug_Print 1 "Number of files to process is not equal to what it should be"
          IGCM_debug_Print 1 "We found ${nbfile} files and it should have been ${NbPeriodPerFrequency} files"
          IGCM_debug_Exit "ERROR in number of files to process. STOP HERE INCLUDING THE COMPUTING JOB"
          IGCM_debug_Verif_Exit
        fi
        output=${JobName}_${date_begin_pack}_${date_end_pack}_${myType}
        #cat liste_pack_${myType}_${date_begin_pack}_${date_end_pack}.txt | xargs ncrcat -v ${list_var} -o ${output}
        if [ X${list_var} = X ] ; then
          IGCM_sys_ncrcat -p ${dir} ${liste_file_tmp} --output ${output}
        else
          IGCM_sys_ncrcat -x -v ${list_var} -p ${dir} ${liste_file_tmp} --output ${output}
        fi
        # ------------------------------------------------------------------
        # Test if all was right before proceeding further
        # ------------------------------------------------------------------
        IGCM_debug_Verif_Exit
        # Save it
        IGCM_sys_Put_Out ${output} ${R_SAVE}/$( echo $dir | sed "s:${R_BUFR}/::" )/${output}
        # Clean file produced by ncrcat
        IGCM_sys_Rm ${output}
        # ------------------------------------------------------------------
        # Test if all was right before proceeding further
        # ------------------------------------------------------------------
        IGCM_debug_Verif_Exit
        # Clean files used by ncrcat
        cat liste_pack_${myType}_${date_begin_pack}_${date_end_pack}.txt | xargs rm
        # Save the list of files that has been pack (ncrcat)
        #mv liste_pack_${myType}_${date_begin_pack}_${date_end_pack}.txt ${STORE_DEBUG}
        IGCM_debug_Print 1 "Ncrcat and cleaning done for ${dir} and ${myType}"
        echo
      done
    done
  done
  (( number_pack = number_pack + 1 ))
  # Add 1 day to date_end_pack to have the new date_begin_pack
  date_begin_pack=$( IGCM_date_AddDaysToGregorianDate ${date_end_pack} 1 )
done

# Flush post-processing submission
if [ -f ${R_BUFR}/FlushPost_${DateEnd}.ksh ] ; then
  . ${R_BUFR}/FlushPost_${DateEnd}.ksh
  IGCM_FlushPost
  #IGCM_sys_Rm -f ${R_BUFR}/FlushPost_${DateEnd}.ksh
fi

# Clean RUN_DIR_PATH (necessary for cesium and titane only)
IGCM_sys_RmRunDir -Rf ${RUN_DIR_PATH}

# ------------------------------------------------------------------
# Finalize BigBrother to inform that the jobs end
# ------------------------------------------------------------------
IGCM_debug_BigBro_Finalize

date