Parent:
[a8e6ce]
(diff)
Child:
[2bb14c]
(diff)
Download this file
rcltex
107 lines (88 with data), 2.4 kB
#!/bin/sh
# @(#$Id: rcltex,v 1.2 2007-11-09 15:56:14 dockes Exp $ (C) 2004 J.F.Dockes
#================================================================
# Translate TeX files for recoll. Uses either untex or detex to translate to html
#================================================================
# set variables
LANG=C ; export LANG
LC_ALL=C ; export LC_ALL
progname="rcltex"
filetype=TeX
#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
if iscmd detex ; then
checkcmds iconv
CMD="detex -n -e ''"
else
checkcmds untex iconv
CMD="untex -giso -a"
fi
# output the result
echo '<html><head>'
#echo '<title>' "$title" '</title>'
echo '<meta http-equiv="Content-Type" content="text/html;charset=UTF-8">'
echo '</head><body>'
echo '<pre>'
#untex -giso -a "$infile" | \
$CMD "$infile" | \
iconv -c -f iso-8859-1 -t utf-8 | \
sed \
-e 's/</</g' -e 's/&/&/g'
echo '</pre>'
echo '</body></html>'
# exit normally
exit 0