#!/bin/sh
# copyright (c) jose nazario (jose at monkey dot org), all rights
# reserved. this is NOT GPLd, it is available under a BSD style license.

# mk_pdf_index: utility to build the search index for PS, PDF files
# requires: xpdf package (which contains pdftotext) and ghostscript 5.5
# or later, which contains ps2pdf.

export UNDER=${HOME}

# back it up
if [ -d ${HOME}/.index ]; then
	mv ${HOME}/.index ${HOME}/.index-old
fi

# if it doesn't exist make it
mkdir ${HOME}/.index

# convert all PS files to PDF and txt index
for i in `find ${UNDER} -name "*.ps"`
        do
        ps2pdf $i /tmp/tmpfile.pdf
        pdftotext /tmp/tmpfile.pdf
        echo $i > `basename $i ps`idx
        cat /tmp/tmpfile.txt >> `basename $i ps`idx
        mv `basename $i ps`idx ${HOME}/.index
        rm /tmp/tmpfile*
done

# handle the PDFs ...
for i in `find ${UNDER} -name "*.pdf"`
        do 
        cp $i /tmp/tmpfile.pdf
        pdftotext /tmp/tmpfile.pdf
        echo $i > `basename $i pdf`idx
        cat /tmp/tmpfile.txt >> `basename $i pdf`idx
        mv `basename $i pdf`idx ${HOME}/.index
        # XXX   will remove real PDFs stored in /tmp ...
        rm /tmp/tmpfile* /tmp/*.pdf
done

# clean up the backup
rm -rf ${HOME}/.index-old
