#!/bin/sh

help() { echo "vartak-results-data - convert vartak results pdf into CSV data files

USAGE:
	vartak-results-data [OPTION]... <FILE>

OPTIONS:
	-d DEST   produce data in the specified DESTination path/directory
	-t        top header
	-h        show this help message"; }

warn() { printf "WARNING: %s\n" "$@" >&2; }
err() { printf "vartak-result-data: %b\n" "$@" >&2; exit 1; }

while getopts 'td:h' o; do case "$o" in
	d) export DEST="$OPTARG" ;;
	t) tflag=1 ;;
	h) help; exit ;;
	*) err "invalid option -- '$OPTARG'" ;;
esac done
shift $((OPTIND - 1))

[ "$#" -lt 1 ] && help >&2 && exit 1

[ "${1#*ATKT}" != "$1" ] && err "ATKT files are not supported"

filetype="$(file --mime-type --brief "$1")"
case "$filetype" in
	application/pdf) pdftotext -layout "$1"; file="${1%.pdf}.txt"; tmp=1 ;;
	text/plain) file="$1" ;;
	*) err "only PDF and text files are supported\n$1 :-\n\t$filetype" ;;
esac

file="$(realpath "$file")"
printf "\n:: Preparing data in : %s\n" "${DEST:=$PWD}"
mkdir -pv "$DEST"
cd "$DEST" || exit

if [ "$tflag" = 1 ]; then
	header="$(sed -nE "0,/^Seat .*Name \s{2,}(.*Total).*$/ s//\1/p" "$file" |
		sed -E 's/\s?\[ [0-9]+(\s\])?\s*/,/g; s/,$//')"
else
	header="$(sed -nE '0,/^[0-9]{5} [/A-Z ]*\s{2,}(.*)/ s//\1/p' "$file" |
		sed -E 's/\s?\[ [0-9]+(\s\])?\s*/,/g; s/,$//')"
fi

if [ -f names.csv ]; then
	echo "names.csv already exists, skipping..."
else
	echo "Seat No,Name" >> names.csv
	awk -f - "$file" >> names.csv << EOF
		/^[0-9]{5}/ {
			name = ""
			for (i = 2; \$i !~ /^(${header%%[ ,]*}|INTER|TW)/; i++) { name = name " " \$i }
			sub(/^\s*/, "", name)
			print \$1 "," name
		}
EOF
fi

header="Seat No,$header"

if [ -f marks.csv ]; then
	echo "marks.csv already exists, skipping..."
else
	echo "$header" > marks.csv
	awk '
		/^[0-9]{5}/ { printf("%d", $1) }
		!/^[0-9]/ && !/Seat/ && !/Total\s*\[/ && /Total/ {
			start = 1
			while ($start != "Total") start++
			start++
			for (i = start; i <= NF; i++) {
				if ($i ~ /^AB$/) {
					printf(",%s", "0")
				} else if ($i ~ /^[0-9]+$/) {
					printf(",%d", $i)
				}
			}
			printf("\n")
		}
	' "$file" >> marks.csv
fi

if [ -f GP.csv ]; then
	echo "GP.csv already exists, skipping..."
else
	echo "${header%,Total}" > GP.csv
	awk '
		/^[0-9]{5}/ { printf("%d", $1) }
		/ GP / {
			start = 1
			while ($start != "GP") start++
			start++
			for (i = start; i <= NF; i++) {
				if ($i ~ /F/) {
					printf(",%s", "0")
				} else if ($i ~ /^[0-9]+$/) {
					printf(",%d", $i)
				}
			}
			printf("\n")
		}
	' "$file" >> GP.csv
fi

if [ -f CGP.csv ]; then
	echo "CGP.csv already exists, skipping..."
else
	echo "$header,CGPA" > CGP.csv
	awk '
		/^[0-9]{5}/ { printf("%d", $1) }
		/ GPA / { gpa = $NF }
		/ CG / {
			start = 1
			while ($start != "CG") start++
			start++
			for (i = start; i <= NF; i++) {
				if ($i ~ /F/) {
					printf(",%s", "0")
				} else if ($i ~ /^[0-9]+$/) {
					printf(",%d", $i)
				}
			}
			printf(",%.2f\n", gpa)
		}
	' "$file" >> CGP.csv
fi

rows() {
	[ -z "$(sed -n '1!d; /^Seat No,/p' "$1")" ] &&
		warn "missing header in file: $1";
	grep -cv 'Seat No' "$1"
}

# no. of rows for each files
nr_names="$(rows names.csv)"
nr_marks="$(rows marks.csv)"
nr_GP="$(rows GP.csv)"
nr_CGP="$(rows CGP.csv)"

if [ "$nr_names" != "$nr_marks" ] ||
		[ "$nr_names" != "$nr_GP" ] ||
		[ "$nr_names" != "$nr_CGP" ]; then
	warn "inconsitent number of rows"
fi

cols() { awk -F, '
		NR == 1 { min = NF }
		{ if (NF > max) max = NF; if (NF < min) min = NF }
		END {
			if (max != min)
				print "WARNING: inconsitent columns in file: " FILENAME > "/dev/stderr"
			print max
		}
	' "$1"; }

# no. of columns for each file
nc_names="$(cols names.csv)"
nc_marks="$(cols marks.csv)"
nc_GP="$(cols GP.csv)"
nc_CGP="$(cols CGP.csv)"

[ "$nc_names" != 2 ] &&
	warn "names data doesn't have exactly 2 columns"
[ "$nc_marks" != "$(( nc_GP + 1))" ] &&
	warn "marks data doesn't have an additional column GP data"
[ "$nc_CGP" != "$(( nc_marks + 1 ))" ] &&
	warn "CGP data doesn't have an additional column to marks data"

printf ":: Finished\n"

# clean up
[ "$tmp" = 1 ] && rm -f "$file"