Below is the full bag of tricks:
#!/bin/bash
# set the language id (name of hunspell dic without extension)
if [ ! $1 ] ; then
echo "ENTER THE NAME OF THE DICTIONARY FILE WITHOUT .DIC AS A PARAMTER"
else
if [ -f $1.dic ] ; then
if [ -f $1.aff ] ; then
LANG=$1
# try to unmunch
echo "UNMUNCHING"
unmunch $LANG.dic $LANG.aff | sed "s/\/.*$//g" > $LANG.txt
# use dictionary as input anyway, removing flags
echo "MAKING DICTIONARY WITHOUT FLAGS"
cat $LANG.dic | sed "s/\/.*$//g" >> $LANG.tmp
# add a small modifyer to the dic words to generate more alternatives
echo "ADDING MODIFIED DICTIONARY ITEMS FOR MORE SUGGESTIONS"
cat $LANG.tmp >> $LANG.txt
cat $LANG.tmp | sed "s/.$/x/g" >> $LANG.txt
rm $LANG.tmp
# add random words (languages that looks a bit like it or just noise)
echo "ADDING OTHER WORDS IF AVAILABLE"
if [ -f other.txt ] ; then
cat other.txt >> $LANG.txt
else
echo " no other.txt present; add one (from similar language) for better
results"
fi
# sorting and getting unique
echo "MAKING WORDS IN LIST UNIQUE"
cat $LANG.txt | sort | uniq > $LANG.in
rm $LANG.txt
# use the input to generate suggestions
echo "USING HUNSPELL TO GET SUGGESTIONS (TERRIBLY SLOW!)"
hunspell -i utf-8 -d $LANG -a $LANG.in | grep "&" > $LANG.suggestions
# edit the suggestions into words
echo "ADD SUGGESTIONS TO WORDS LIST"
cat $LANG.suggestions | sed "s/&.*: //g" | sed "s/, /\n/g" | sort |
uniq >> $LANG.in
rm $LANG.suggestions
# get all correct words
echo "SPELLCHECK ALL WORDS TO GET CORRECT WORDS"
hunspell -i utf-8 -G -d $LANG $LANG.in > $LANG.okay
else
echo "$1.aff IS MISSING"
fi
else
echo "$1.dic IS MISSING"
fi
fi
------------------------------------------------------------------------------
_______________________________________________
Languagetool-devel mailing list
Languagetool-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/languagetool-devel
