#! /bin/bash

File=$1
Method=$2

# find the avgestimate and avgactual
AvgEstimate=$(gawk -v Method=$Method 'BEGIN{FS=OFS=","}{if ($2==Method){print $0}}' $File | gawk 'BEGIN{FS=OFS=","}{sum+=$3}END{print sum/NR}')
AvgActual=$(gawk -v Method=$Method 'BEGIN{FS=OFS=","}{if ($2==Method){print $0}}' $File | gawk 'BEGIN{FS=OFS=","}{sum+=$4}END{print sum/NR}')

gawk -v Method=$Method 'BEGIN{FS=OFS=","}{if ($2==Method){print $0}}' $File |
gawk -v Method=$Method -v AvgEstimate=$AvgEstimate -v AvgActual=$AvgActual 'BEGIN{FS=OFS=","}
{	Re = ($3-$4)/$4
	mre[NR] = Re < 0 ? -1* Re : Re;
	EstimatedEffort=$3
	ActualEffort=$4
	Corr_Se+=((EstimatedEffort-AvgEstimate)*(EstimatedEffort-AvgEstimate));
	Corr_Sa+=((ActualEffort-AvgActual)*(ActualEffort-AvgActual));
	Corr_Sea+=((EstimatedEffort-AvgEstimate)*(ActualEffort-AvgActual));
}
END {
	sum=0;
	for (i=1;i<=NR;i++) {sum+=mre[i]}
	
	mmre=100*(sum/NR)
	
	pred30=0
	for (i=1;i<=NR;i++) {if (mre[i]<=0.3){pred30++;}}
	pred30=100*(pred30/NR)
	
	Corr_Se=Corr_Se/(NR-1);
	Corr_Sa=Corr_Sa/(NR-1);
	Corr_Sea=Corr_Sea/(NR-1);
	Corr=(Corr_Sea/sqrt(Corr_Se*Corr_Sa));
	
	n = asort(mre)
	i = int (n/2)
	med_mre = 100*mre[i]
	
	printf("%s,%.2f,%.2f,%.1f,%.3f\n",Method, mmre, med_mre, pred30, Corr);
	}'