#! /bin/bash

# TO DO

# move cocomost folder to sandbox and make a cleaner version for the main

# make it work for a set of learners
# move gnuplot stuff to a function or a script so it can be used in a loop
# fix gnuplot use of some variables such as $Learner
# include code for cocomost and include in graphs
# include mmre as a label or something on the graphs for each learner

# duplicate code for nasa93 and make sure nothing is hardcoded (use more variables and make sure to set them)

rm -r log/recordGraphs
mkdir log/recordGraphs

#grep LC nfoldCocomostExperimentSep28.csv | gawk 'BEGIN{FS=OFS=","}{print $1,$2,$NF}'

MakeSubsets=1
DataPath="data/subsets"
if ((MakeSubsets==1))
then
 rm -r $DataPath
 mkdir -p $DataPath
 cp data/coc81modetypelangtype.csv $DataPath"/coc81_all.csv"
 cp data/nasa93.csv $DataPath"/nasa93_all.csv"

 #Make the subsets of coc81 data
 gawk -f csvsubset.awk Col="kind" Want="min" data/coc81modetypelangtype.csv > $DataPath"/coc81_kind_min.csv"
 gawk -f csvsubset.awk Col="lang" Want="ftn" data/coc81modetypelangtype.csv > $DataPath"/coc81_lang_ftn.csv"
 gawk -f csvsubset.awk Col="mode" Want="e" data/coc81modetypelangtype.csv > $DataPath"/coc81_mode_e.csv"
 gawk -f csvsubset.awk Col="kind" Want="max" data/coc81modetypelangtype.csv > $DataPath"/coc81_kind_max.csv"
 gawk -f csvsubset.awk Col="mode" Want="org" data/coc81modetypelangtype.csv > $DataPath"/coc81_mode_org.csv"
 gawk -f csvsubset.awk Col="lang" Want="mol" data/coc81modetypelangtype.csv > $DataPath"/coc81_lang_mol.csv"
 #Make the subsets of nasa93 data
 gawk -f csvsubset.awk Col="projectname" Want="gro" data/nasa93.csv > $DataPath"/nasa93_project_gro.csv"
 gawk -f csvsubset.awk Col="cat2" Want="missionplanning" data/nasa93.csv > $DataPath"/nasa93_cat2_missionplanning.csv"
 gawk -f csvsubset.awk Col="cat2" Want="avionicsmonitoring" data/nasa93.csv > $DataPath"/nasa93_cat2_avionicsmonitoring.csv"
 gawk -f csvsubset.awk Col="mode" Want="semidetached" data/nasa93.csv > $DataPath"/nasa93_mode_semidetached.csv"
 gawk -f csvsubset.awk Col="projectname" Want="sts" data/nasa93.csv > $DataPath"/nasa93_project_sts.csv"
 gawk -f csvsubset.awk Col="forg" Want="g" data/nasa93.csv > $DataPath"/nasa93_forg_g.csv"
 gawk -f csvsubset.awk Col="center" Want="5" data/nasa93.csv > $DataPath"/nasa93_center_5.csv"
 gawk -f csvsubset.awk Col="year" Want="1975" data/nasa93.csv > $DataPath"/nasa93_year_1975.csv"
 gawk -f csvsubset.awk Col="year" Want="1980" data/nasa93.csv > $DataPath"/nasa93_year_1980.csv"
 gawk -f csvsubset.awk Col="mode" Want="embedded" data/nasa93.csv > $DataPath"/nasa93_mode_embedded.csv"
 gawk -f csvsubset.awk Col="center" Want="2" data/nasa93.csv > $DataPath"/nasa93_center_2.csv"
fi

Learner="LC"
Learners="LC cocomost"

Experiment="log/nfoldCocomostExperimentSep28.csv"

Coc81Subsets="coc81_all.csv coc81_kind_min.csv coc81_lang_ftn.csv coc81_mode_e.csv coc81_kind_max.csv coc81_mode_org.csv coc81_lang_mol.csv" 
Nasa93Subsets="nasa93_all.csv nasa93_project_gro.csv nasa93_cat2_missionplanning.csv nasa93_cat2_avionicsmonitoring.csv nasa93_mode_semidetached.csv nasa93_project_sts.csv nasa93_forg_g.csv nasa93_center_5.csv nasa93_year_1975.csv nasa93_year_1980.csv nasa93_mode_embedded.csv nasa93_center_2.csv"
AllSubsets=$Coc81Subsets" "$Nasa93Subsets

# Isolate the mre's of learner on the subset
for Learner in $Learners
do
for Subset in $AllSubsets
do
  Converter="-"$(gawk 'BEGIN{FS=","}NR<=2{next}{printf("%d-",$1);}' $DataPath"/"$Subset)
  grep $Learner $Experiment | 
  grep $Subset | 
  gawk 'BEGIN{FS=OFS=","}{print $NF}' |
  gawk -F, -v Converter=$Converter '
   {Data[NR]=$1}
   END{
    Rec=1
    for(X=1;X<=63;X++)
    {
      targ="-"X"-";
      if (index(Converter, targ)!=0) {
        print Data[Rec]
        Rec++;                    
      }
      else
      {
        print "NA"
      }
    }
   }
  ' > "/tmp/"$Learner"_"$Subset
done
done

TmpFile1="/tmp/mre_list.1"
TmpFile2="/tmp/mre_list.2"

# First handle COC81
# Now make a graph of the mre's of each record
for ((i=1;i<=63;i++))
do

  OutFile="log/recordGraphs/coc81_"$Learner"_"$i".pdf"
  Title="coc81_record_#"$i
  
  # Get a list of the mre's for the record
  echo -n "" > $TmpFile1
  echo -n "" > $TmpFile2

  # Generate the graph input for LC
  count=1;
  for Subset in $Coc81Subsets
  do
    gawk -v I=$i -v Count=$count 'NR==I{if($0!="NA"){printf("%d %f\n",Count,$0);}}' "/tmp/LC_"$Subset >> $TmpFile1
    ((count++))
  done
  # Generate the graph input for cocomost
  count=1;
  for Subset in $Coc81Subsets
  do
    gawk -v I=$i -v Count=$count 'NR==I{if($0!="NA"){printf("%d %f\n",Count,$0);}}' "/tmp/cocomost_"$Subset >> $TmpFile2
    ((count++))
  done

  ./plotMRE $Title $OutFile $TmpFile1 $TmpFile2
done