#!/bin/bash

DIX=/Users/sagiemaoz/dev/gsoc/apertium-mt-he/apertium-mt-he.mt.dix
BIN=/Users/sagiemaoz/dev/gsoc/apertium-mt-he/mt-he.automorf.bin
CORPUS=$1
XLANG=mt

if [ -z $CORPUS ]; then
	CORPUS="wikipedia";
fi

cat $CORPUS/$XLANG.crp.txt | python strip_html.py | apertium-destxt | lt-proc -w $BIN | apertium-retxt | sed 's/\$\W*\^/$\n^/g' > /tmp/$XLANG.coverage.txt

EDICT=`cat $DIX | grep '<e lm' | wc -l`;
EPAR=`cat $DIX | grep '<pardef n' | wc -l`;
TOTAL=`cat /tmp/$XLANG.coverage.txt | wc -l`
KNOWN=`cat /tmp/$XLANG.coverage.txt | grep -v '*' | wc -l`
COV=`calc $KNOWN / $TOTAL`;
DATE=`date`;

echo -e $DATE"\t"$EPAR":"$EDICT"\t"$KNOWN"/"$TOTAL"\t"$COV >> logs/history.$CORPUS.log
tail -1 logs/history.$CORPUS.log