#! /bin/bash

Path="/tmp/$USER"
Filename=$1
Learner="attrLC"
RankingFile="rank.tmp"
RankerLogFile=$Path"/cocomin_ranking.log"
TargetBase="-"
EMstart=1
FieldCount=$(gawk 'BEGIN{FS=","}NR==1{print NF}' $Filename)
((EMend=FieldCount-2))

for ((i=$EMstart;i<=$EMend;i++))
do
	TargetEMs=$TargetBase$i-
	LogFile=$RankerLogFile".ranking."$i
	gawk -f lc.awk Pass=1 TargetEMs=$TargetEMs $Filename Pass=2 TargetEMs=$TargetEMs $Filename > $LogFile
	Stats=$(./calc $LogFile)
	
	#Calculate Standard Deviation of Attribute
	AttrDev=$(gawk -f columnstdev.awk Columns=$TargetEMs $Filename)

	#Calculate Entropy of Attribute (using discrete formula because the effort multipliers are discrete)
	AttrEntropy=$(gawk -f columnentropy.awk Columns=$TargetEMs $Filename)

	AvgCorr=$(echo $Stats | gawk 'BEGIN{FS=","} NR==1{print $4}')

	echo $TargetEMs","$AvgCorr","$AttrDev","$AttrEntropy
	
done > $RankingFile

sort -t, -nrk2 $RankingFile | gawk -F, '{print $1}' | gawk 'BEGIN{FS="-"}{print $2-1}' > "cor_hi"
sort -t, -nrk3 $RankingFile | gawk -F, '{print $1}' | gawk 'BEGIN{FS="-"}{print $2-1}' > "var_hi"
sort -t, -nrk4 $RankingFile | gawk -F, '{print $1}' | gawk 'BEGIN{FS="-"}{print $2-1}' > "ent_hi"
sort -t, -nk2 $RankingFile | gawk -F, '{print $1}' | gawk 'BEGIN{FS="-"}{print $2-1}' > "cor_lo"
sort -t, -nk3 $RankingFile | gawk -F, '{print $1}' | gawk 'BEGIN{FS="-"}{print $2-1}' > "var_lo"
sort -t, -nk4 $RankingFile | gawk -F, '{print $1}' | gawk 'BEGIN{FS="-"}{print $2-1}' > "ent_lo"