#!/bin/bash

wc -l $1 > result.log

cut -f 2 $1 > hypothesis.txt
cut -f 3 $1 > reference.txt

awk 'BEGIN{FS=OFS="\t"}{print $0,"(sp1_"NR")"}' < hypothesis.txt > hypothesis.trn
awk 'BEGIN{FS=OFS="\t"}{print $0,"(sp1_"NR")"}' < reference.txt > reference.trn

sclite -f 0 -r reference.trn trn -h hypothesis.trn trn -e utf-8 -i rm -o all stdout | grep Scores > scores 
awk 'BEGIN{OFS="\t"}{print $6, $7, $8, $9}' < scores | awk '{ print ($2 + $3 + $4)/($2 + $3 + $1) }' > wer.tsv

awk 'BEGIN{FS=OFS="\t"}{ bad = bad + $1; all++ }END{ print bad/all }' < wer.tsv >> wer.txt 
awk 'BEGIN{FS=OFS="\t"}{ if ( $1 == 0 ) good++; all++ }END{ print good/all }' < wer.tsv >> srr.txt

tail -n 500 wer.txt > wer.txt
tail -n 500 srr.txt > srr.txt