#!/bin/zsh

# pairs.sh LENGTH FILE
# count glyph pair occurance in a document

# version	date		author
# 0		22 Apr 2012	Brian M Zick [21326.info]

length="$1"		# length of pair group
input="$2"		# input file

# Rundown of script:
# 1. Groups pairs without whitespace
# 2. Counts pairs
# 3. Calculates percentages


group() {
	 grep -Po '[^\s\b]'"{$length}" 
}

count() {
	sort | uniq -c | sort -rn
}

percent() {
	file=`mktemp`
	tee $file | while read num pair; do (( sum += num )); done
	cat "$file" | while read num pair; do
		(( percent = ($num.0 / $sum.0) * 100 ))
		print -- "$pair\t$num\t$percent"
	done
}


cat $input | group | count | percent