#! /bin/bash

Strawman="LC"
Candidates="ExhaustiveFSS_mmre ExhaustiveFSS_pred25 ExhaustiveFSS_pred30 ExhaustiveFSS_pred40 ExhaustiveFSS_pred50 ExhaustiveFSS_corr ExhaustiveFSS_median_mre ExhaustiveFSS_sd(mre) ExhaustiveFSS_mmre10to90 ExhaustiveFSS_mmre20to80 ExhaustiveFSS_mmre25to75 ExhaustiveFSS_mmre33to67"
Subsets="coc81_all nasa93_all coc81_kind_min coc81_lang_ftn coc81_mode_e coc81_kind_max coc81_mode_org coc81_lang_mol nasa93_project_gro nasa93_cat2_missionplanning nasa93_cat2_avionicsmonitoring nasa93_mode_semidetached nasa93_project_sts nasa93_forg_g nasa93_center_5 nasa93_year_1975 nasa93_year_1980 nasa93_mode_embedded nasa93_center_2"
Datasets="../new_results.csv"
#FILENAME,I,LEARNER,ESTIMATE,ACTUAL,MRE,COST_DRIVERS

for Data in $Datasets
do
  DataName=$(basename $Data .csv)
  if [ -d $DataName ]; then
	echo -n ""
  else
	mkdir $DataName
  fi
  
  for Candidate in $Candidates
  do
  	Path=$DataName"/"$Candidate
	if [ -d $Path ]; then
		echo -n ""
	else
		mkdir $Path
	fi
	
	for Subset in $Subsets
	do
		RawFile=$Path"/"$Subset"_Raw.txt"
		RanksFile=$Path"/"$Subset"_Ranks.txt"
	
		grep $Subset $Data | 
		gawk -v A=$Strawman -v B=$Candidate 'BEGIN{FS=OFS=","}{if (($3==A)||($3==B)) {print $3,$4,$5}}' | 
		../../nonparametric/nonParametricAnalysisWMW RawResultsFileName=$RawFile RanksFileName=$RanksFile
	done	# end of subsets
  
	# Put the first aggregate function in this script too (and then another to aggregate results against strawman
	
	cd $Path
	../../../../nonparametric/aggregate
	cd ../../
  
  done	# end of candidates
  
  # Now compile the results of each learner's performance against the strawman into one list
  {
	echo "key,ties,wins,losses,wins-losses"
	for Candidate in $Candidates
	do
		Path=$DataName"/"$Candidate
		grep -v $Strawman $Path"/all_aggregate.txt" | grep -v "wins-losses"
	done
  } > $DataName"/StrawmanVsCandidatesSummary.txt"
  
done	# end of datasets