#!/bin/sh
# @(#$Id: rclsoff,v 1.12 2008-10-08 08:27:34 dockes Exp $ (C) 2004 J.F.Dockes
# Parts taken from Estraier:
#================================================================
# Estraier: a personal full-text search system
# Copyright (C) 2003-2004 Mikio Hirabayashi
#================================================================
#================================================================
# Extract text from an openoffice/soffice file
#
#================================================================
# set variables
LANG=C ; export LANG
LC_ALL=C ; export LC_ALL
progname="rclsoff"
filetype=openoffice
#RECFILTCOMMONCODE
##############################################################################
# !! Leave the previous line unmodified!! Code imported from the
# recfiltcommon file
# Utility code common to all shell filters. This could be sourced at run
# time, but it's slightly more efficient to include the code in the
# filters at build time (with a sed script).
# Describe error in a way that can be interpreted by our caller
senderror()
{
echo RECFILTERROR $*
# Also alert on stderr just in case
echo ":2:$progname::: $*" 1>&2
exit 1
}
iscmd()
{
cmd=$1
case $cmd in
*/*)
if test -x $cmd -a ! -d $cmd ; then return 0; else return 1; fi ;;
*)
oldifs=$IFS; IFS=":"; set -- $PATH; IFS=$oldifs
for d in $*;do test -x $d/$cmd -a ! -d $d/$cmd && return 0;done
return 1 ;;
esac
}
checkcmds()
{
for cmd in $*;do
if iscmd $cmd
then
a=1
else
senderror HELPERNOTFOUND $cmd
fi
done
}
# show help message
if test $# -ne 1 -o "$1" = "--help"
then
echo "Convert a $filetype file to HTML text for Recoll indexing."
echo "Usage: $progname [infile]"
exit 1
fi
infile="$1"
# check the input file existence (may be '-' for stdin)
if test "X$infile" != X- -a ! -f "$infile"
then
senderror INPUTNOSUCHFILE "$infile"
fi
# protect access to our temp files and directories
umask 77
##############################################################################
# !! Leave the following line unmodified !
#ENDRECFILTCOMMONCODE
checkcmds xsltproc unzip
# We need a temporary directory
if test z"$RECOLL_TMPDIR" != z; then
ttdir=$RECOLL_TMPDIR
elif test z"$TMPDIR" != z ; then
ttdir=$TMPDIR
else
ttdir=/tmp
fi
tmpdir=$ttdir/rclsoff_tmp$$
mkdir $tmpdir || exit 1
mkdir $tmpdir/rclsofftmp || exit 1
cleanup()
{
# Note that we're using a constant part (rclsofftmp), that hopefully
# guarantees that we can't do big mistakes here.
rm -rf $tmpdir/rclsofftmp
rmdir $tmpdir
}
trap cleanup EXIT HUP QUIT INT TERM
# Unzip the input file and change to the unzipped directory
unzip -q -d $tmpdir/rclsofftmp "$infile"
cd $tmpdir/rclsofftmp
echo '
'
xsltproc --novalid --nonet - meta.xml <
<xsl:value-of select="."/>
abstract
keywords
author
keywords
EOF
echo ''
xsltproc --novalid --nonet - content.xml <
EOF
echo ''
cd /
exit 0