#!/bin/bash
# 2017-02-02

# Various stats from UoB style exam results.
#
# Use as: uobestat file [no of A questions] [no of B questions] <options>
# options are listed if no correct number of them given
# File is of the format Cand.No;desk;A1;A2;A3;A4;A5;B1;B2;comment
# only lines starting with a number are considered.
#

if [ "$4" = "a" ]; then
 cat "$1" |
 awk -F\; -v ano="$2" 'BEGIN {
  printf("C.no	desk	")
  for (i=1; i<(1+ano); i+=1) printf("A"i"	");
  print("Sum A")
 }
 {
  if ($1 + 0 == $1) {
   asum=0
   printf($1"	"$2"	")
   for (i=3; i<(3+ano); i+=1) {
    asum=asum+$i
    printf($i"	")
   }
   print(asum)
  };
 }'

elif [ "$4" = "b" ]; then
 cat "$1" |
 awk -F\; -v ano="$2" -v bno="$3" 'BEGIN {
  printf("C.no	desk	")
  for (i=1; i<(1+bno); i+=1) printf("B"i"	");
  print("Sum B")
 }
 {
  if ($1 + 0 == $1) {
   bsum=0
   printf($1"	"$2"	")
   for (i=(3+ano); i<(3+ano+bno); i+=1) {
    bsum=bsum+$i
    printf($i"	")
   }
   print(bsum)
  };
 }'

elif [ "$4" = "abs" ]; then
 cat "$1" |
 awk -F\; -v ano="$2" -v bno="$3" 'BEGIN {
  printf("C.no	desk		")
  for (i=1; i<(1+ano); i+=1) printf("A"i"	");
  printf("	")
  for (i=1; i<(1+bno); i+=1) printf("B"i"	");
  printf("	")
  print("Sum A	Sum B		Sum")
 }
 {
  if ($1 + 0 == $1) {
   asum=0
   bsum=0
   printf($1"	"$2"		")
   for (i=3; i<(3+ano); i+=1) {
    asum=asum+$i
    printf($i"	")
   }
   printf("	")
   for (i=(3+ano); i<(3+ano+bno); i+=1) {
    bsum=bsum+$i
    printf($i"	")
   }
   printf("	")
   print(asum"	"bsum"		"asum+bsum)
  };
 }'

elif [ "$4" = "s" ]; then
 cat "$1" |
 awk -F\; 'BEGIN {
  print("C.no	desk	Sum")
 }
 {
  if ($1 + 0 == $1) {
   sum=0
   for (i=3; i<(3+ano+bno); i+=1) sum=sum+$i;
   printf($1"	"$2"	")
   print(sum)
  };
 }' ano="$2" bno="$3"

elif [ "$4" = "v" ]; then
 cat "$1" |
 awk -F\; -v ano="$2" -v bno="$3" 'BEGIN {
  for (i=3; i<(3+ano+bno); i+=1) su[i]=0;
  nos=0
  for (i=1; i<=ano; i+=1) printf("A"i"	");
  for (i=1; i<=bno; i+=1) printf("B"i"	");
  print("Sum A	Sum B	Sum	No of students")
 }
 {
  if ($1 + 0 == $1) {
   for (i=3; i<(3+ano+bno); i+=1) su[i]=su[i]+$i;
   nos=nos+1
  };
 }
 END {
  for (i=3; i<(3+ano+bno); i+=1) av[i]=su[i]/nos;
  suma=0
  sumb=0
  for (i=3; i<(3+ano+bno); i+=1) {
   printf("%3.2f %s",av[i],"	")
  }
  for (i=3; i<(3+ano); i+=1) {
   suma=suma+av[i]
  }
  for (i=(3+ano); i<(3+ano+bno); i+=1) {
   sumb=sumb+av[i]
  }
  printf("%3.2f%s%3.2f%s%3.2f%s%3.0f\n",suma,"	",sumb,"	",suma+sumb,"	",nos)
 }'

elif [ "$4" = "h" ]; then

 elej=${1%.*}
 #k=v3$elej`date +%N`
 k=v3$elej`date +%y%m%d_%H%M%S`_`strings /dev/urandom | tr -dc a-z-0-9 | tr -d - | head -c5`
 mkdir /tmp/$k
 chmod 700 /tmp/$k/
 gnufile=/tmp/$k/"$elej"_gnu.txt
 rawfile=/tmp/$k/"$elej"_raw.csv
 scaledfile=/tmp/$k/"$elej"_raw_scaled.csv

 cat "$1" |
 awk -F\; ' {
  if ($1 + 0 == $1) {
   sum=0
   for (i=3; i<(3+ano+bno); i+=1) sum=sum+$i;
   print(sum)
  };
 }' ano="$2" bno="$3" > "$rawfile"

 msarg=`echo "markscaler \"$5\" \"$6\" $rawfile"`
 eval "$msarg"

 nol=`wc -l "$rawfile" | awk '{ print($1) }'`

 paste -d \	 "$rawfile" "$scaledfile" |
 
 sort -g |

 awk -F $'\t' -v nol="$nol" 'BEGIN {
 print("Raw	   %	Scaled		"nol" students")
 raws=0
 scs=0
 fl=0
 th=0
 sii=0
 si=0
 f=0
 hf=0
 }
 {
 printf("%3.0f%s%3.0f%s%3.2f\n",$1,"	",100*NR/nol,"%	",$2)
 raws=raws+$1
 scs=scs+$2
 if ($2 >= 84.5 ) hf=hf+1;
 if ($2 >= 69.5 && $2 < 84.5) f=f+1;
 if ($2 >= 59.5 && $2 < 69.5) si=si+1;
 if ($2 >= 49.5 && $2 < 59.5) sii=sii+1;
 if ($2 >= 39.5 && $2 < 49.5) th=th+1;
 if ($2 < 39.5) fl=fl+1;
 }
 END {
 print("Averages (of "NR" students):")
 printf("%3.2f%s%3.2f\n",raws/NR,"		",scs/NR)
 print("")
 print("Classes:")
 printf("%s%3.0f%s%3.0f%s\n","85+:	",hf,"	",hf/NR*100,"%")
 printf("%s%3.0f%s%3.0f%s\n","1:	",f,"	",f/NR*100,"%")
 printf("%s%3.0f%s%3.0f%s\n","2i:	",si,"	",si/NR*100,"%")
 printf("%s%3.0f%s%3.0f%s\n","2ii:	",sii,"	",sii/NR*100,"%")
 printf("%s%3.0f%s%3.0f%s\n","3:	",th,"	",th/NR*100,"%")
 printf("%s%3.0f%s%3.0f%s\n","fail:	",fl,"	",fl/NR*100,"%")
 }'

 cat "$scaledfile" |

 awk -F\; 'BEGIN {
  for (i=0; i<=100; i+=1) h[i]=0;
 }
 {
  for (i=0; i<=100; i+=1) {
   if ($1 >= i-0.5 && $1 < i+0.5)
    h[i]=h[i]+1 ;
  }
 }
 END {
  for (i=0; i<=100; i+=1) print(i" "h[i]" ");
 }' > "$gnufile"

gnuplot -persist <<EOF
set xrange [-1:101]
set xtics 5
set style fill solid 0.3
set xtics rotate out
plot "$gnufile" using 1:2 with boxes notitle
EOF

 l=1
 while [ -n "$l" ]; do
  sleep 1
  l=`lsof 2>/dev/null +d /tmp/$k`
 done
 rm -f -R /tmp/$k

else

 echo "Use as: uobestat <file> [no of A questions] [no of B questions] <options>"
 echo "File must be of the format Cand.No;desk;A1;A2;A3;A4;A5;B1;B2;comment"
 echo "Only lines starting with a number are considered. Field separator must be ; ."
 echo "Options are (no - please):"
 echo "a	print sum of section A marks for individual students"
 echo "b	print sum of section B marks for individual students"
 echo "s	print total sum for individual students"
 echo "abs	print all three above"
 echo "v	print averages"
 echo "h	show scaled histogram"
 echo "		needs \"f_1 f_2 ... f_k\" \"t_1 t_2 .. t_k\""
 echo "		where k is any integer at least 2, f_1 is the smallest possible raw mark, f_1...f_k is a strictly increasing sequence, f_k is the highest possible raw mark,"
 echo "		t_1...t_k is a non-decreasing sequence of scaled marks that correspond in order to the raw marks f_1...f_k."
 echo "		Only integers are allowed in scaling scheme."
fi