source: trunk/MagicSoft/Mars/scripts/preparemc@ 8637

#!/bin/sh
# ========================================================================
#
# *
# * This file is part of MARS, the MAGIC Analysis and Reconstruction
# * Software. It is distributed to you in the hope that it can be a useful
# * and timesaving tool in analysing Data of imaging Cerenkov telescopes.
# * It is distributed WITHOUT ANY WARRANTY.
# *
# * Permission to use, copy, modify and distribute this software and its
# * documentation for any purpose is hereby granted without fee,
# * provided that the above copyright notice appear in all copies and
# * that both that copyright notice and this permission notice appear
# * in supporting documentation. It is provided "as is" without express
# * or implied warranty.
# *
#
#
#   Author(s): Daniela Dorner  05/2006 <mailto:dorner@astro.uni-wuerzburg.de>
#
#   Copyright: MAGIC Software Development, 2000-2007
#
#
# ========================================================================
#
#
#
##############################################################################
#
# This script creates mc sequences and datasets for a psf, observation mode 
# and zenith range chosen by the user.
#
# variables, that have to be set by the user:
# - path     directory, where the mc sequence and dataset files are stored 
#              be careful: don't move the sequence files afterwards, as the 
#              paths are stored in the datasetfiles
# - zdmin    minimum zenith distance
# - zdmax    maximum zenith distance
# - psf      psf (at the moment mc with psf 14 and 20 is available)
# - modes    observation mode
#              the explanation for the modes can be found in 
#              /magic/montecarlo/rawfiles/README.txt
# - numruns  number  of runs, that are in the sequence file, which are used 
#            for training (SequencesOn in $mcdataset)
# 
# Remark: For the training of the RF for the energy estimation you need the 
#         macro trainengery.C in the path $path/macros in a modified version, 
#         i.e. the inputfile (mcdataset) and the outputfile (rf-root file) 
#         are given as options: 
#         $macrospath/trainenergy.C+\("\"$mcdataset\""\,"\"$rffile\""\) 
#
# Remark: You have to run the script in your Mars directory.
# 
##############################################################################

# to be set by the user
path=/home/dorner/final_analysis
# path where your mc sequence and dataset files are stored
mcoutpath=$path/mc
# path where you have your modified version of trainenergy.C
macrospath=$path/macros

# zenith range of your data
zdmin=17
zdmax=36
# point spread function of your data
psf=14
# observation mode of your data
#  for explanations see /magic/montecarlo/rawfiles/README.txt
modes=( "02" ) 
# number of runs which are in the sequence files used for training
# please adjust this number such, that you use about 30000 Events for training
numruns=1




# some checks
# checking if the given files and paths are existing
if ! [ -e $mcoutpath ]
then
   echo "Your outpath for the mc $mcoutpathdoes not exist."
   exit
fi
if ! [ -e $macrospath ]
then
   echo "Your macros path $macrospath does not exist."
   exit
fi
if ! [ -e $macrospath/trainenergy.C ]
then
   echo "Your trainenergy.C does not exist in path $macrospath."
   exit
fi

mcdatasettrain=$mcoutpath/mcdataset-for-training.txt
mcdatasetsponde=$mcoutpath/mcdataset-for-sponde.txt

# be careful with $date, when path changes
mcpath=/magic/montecarlo 

zbinmin=`echo "scale=2 ; 100*(1 - c($zdmin*3.14/180))+1" | bc -l`
zbinmax=`echo "scale=2 ; 100*(1 - c($zdmax*3.14/180))+1" | bc -l`

zbinmin=`echo $zbinmin | cut -d. -f1`
zbinmax=`echo $zbinmax | cut -d. -f1`

echo "zd:   min: $zdmin max: $zdmax"
echo "zbin: min: $zbinmin max: $zbinmax"
echo "$numruns runs are classified as test the rest as train"

j=0
for mode in $modes
do 
   for (( i=$zbinmin ; i < $zbinmax ; i++ ))
   do 
      zbin=`printf %02d $i`
      path=$mcpath/rawfiles/19$zbin/$mode/$psf
      runs=(`ls $path | grep Gamma | cut -d_ -f2 | sed -e 's/^0//g' -e 's/^0//g' -e 's/^0//g' -e 's/^0//g' -e 's/^0//g' -e 's/^0//g' | tr "\n" " "`)
      if [ "$runs" = "" ]
      then
         echo "  No runs for zbin $i found. Please try to find some MC with zbin $i!"
         continue
      fi
      echo "found ${#runs[@]} Gamma MC files in path "$path

      runsforfirst=""
      firstrun=${runs[0]}
      secondrun=${runs[${numruns}]}
      firstrunno=`printf %08d $firstrun`
      secondrunno=`printf %08d $secondrun`
      unset runs[0]
      for (( k=1 ; k < $numruns ; k++ ))
      do 
         runsforfirst=$runsforfirst" ${runs[$k]}"
         unset runs[$k]
      done
      date=`echo $path | cut -d/ -f5-7 | sed -e 's/\//-/g'`
      
      trainsequfile=$mcoutpath/sequence$firstrunno.txt
      trainsequences[$j]=$firstrunno
#      echo "  writing train-sequfile "$trainsequfile 
      echo "Sequence:        $firstrun" > $trainsequfile
      echo "Night:           $date" >> $trainsequfile
      echo "" >> $trainsequfile
      echo "CalRuns:         1" >> $trainsequfile
      echo "PedRuns:         2" >> $trainsequfile
      echo "DatRuns:         $firstrun$runsforfirst" >> $trainsequfile
      echo "" >> $trainsequfile
      echo "MonteCarlo: Yes" >> $trainsequfile
      echo "" >> $trainsequfile

      testsequfile=$mcoutpath/sequence$secondrunno.txt
      testsequences[$j]=$secondrunno
#      echo "writing test-sequfile "$testsequfile 
      echo "Sequence:        $secondrun" > $testsequfile
      echo "Night:           $date" >> $testsequfile
      echo "" >> $testsequfile
      echo "CalRuns:         1" >> $testsequfile
      echo "PedRuns:         2" >> $testsequfile
      echo "DatRuns:         ${runs[@]}" >> $testsequfile
      echo "" >> $testsequfile
      echo "MonteCarlo: Yes" >> $testsequfile
      echo "" >> $testsequfile
      
      j=$j+1
   done
done

echo "# test sequences: ${#testsequences[@]}"
echo "# train sequences: ${#trainsequences[@]}"

echo "AnalysisNumber: 1 " > $mcdatasettrain
echo "" >> $mcdatasettrain
echo "SequencesOn: ${trainsequences[@]}" >> $mcdatasettrain
echo "" >> $mcdatasettrain
echo "SequencesOff: ${testsequences[@]}" >> $mcdatasettrain
echo "" >> $mcdatasettrain
echo "" >> $mcdatasettrain

echo "AnalysisNumber: 1 " > $mcdatasetsponde
echo "" >> $mcdatasetsponde
echo "SequencesOn: ${testsequences[@]}" >> $mcdatasetsponde
echo "" >> $mcdatasetsponde
echo "" >> $mcdatasetsponde

for (( i=0 ; i < ${#testsequences[@]} ; i++ ))
do 
   numtrain=${trainsequences[$i]}
   notrain=`echo $numtrain | cut -c 0-4`
   echo "Sequence$numtrain.File: $mcoutpath/sequence$numtrain.txt" >> $mcdatasettrain
   echo "Sequence$numtrain.Dir:  $mcpath/star/$notrain/$numtrain" >> $mcdatasettrain

   numtest=${testsequences[$i]}
   notest=`echo $numtest | cut -c 0-4`
   echo "Sequence$numtest.File: $mcoutpath/sequence$numtest.txt" >> $mcdatasettrain
   echo "Sequence$numtest.Dir:  $mcpath/star/$notrain/$numtrain" >> $mcdatasettrain
   echo "Sequence$numtest.File: $mcoutpath/sequence$numtest.txt" >> $mcdatasetsponde
   echo "Sequence$numtest.Dir:  $mcpath/star/$notrain/$numtrain" >> $mcdatasetsponde
done


# train the rf for energy estimation
logfile=$mcoutpath/trainenergy.log
rffile=$mcoutpath/rf-energy.root

echo "Your mcdataset for training: $mcdatasettrain"
echo "Your rffile: $rffile"
echo ""
echo "Training the RF..."
root -q -b $macrospath/trainenergy.C+\("\"$mcdatasettrain\""\,"\"$rffile\""\) | tee $logfile

echo "" 
echo "Please use rf-file $rffile in your sponde.rc, in case you want to use the RF energy estimator there. "