Parent:
[62da0f]
(diff)
Child:
[7b8031]
(diff)
Download this file
rcllyx
224 lines (193 with data), 6.3 kB
#!/bin/sh
# @(#$Id: rcllyx,v 1.6 2007-06-08 13:51:09 dockes Exp $ (C) 2004 J.F.Dockes
# There may still be code from Estraier in here:
#================================================================
# Estraier: a personal full-text search system
# Copyright (C) 2003-2004 Mikio Hirabayashi
#================================================================
#================================================================
# Convert a lyx file to recoll HTML.
#
# We use lyx --export. It was suggested to use untex, but it doesn't give
# good results on raw lyx (of course, this is not TeX), and exporting to
# LaTex then using untex doesn't look nice when we can use the native lyx
# text export.
# The character encoding of the exported text is defined by the
# \inputencoding directive in the lyx file header and, in quite an obscure
# way, by the \language parameter. We use a heuristic to divine the output
# text encoding and it is guaranteed not to work in all cases. Trials using
# an intermediary dvi, pdf or ps file gave worse results. This needs
# improvement. It doesn't even take into account the fact that the language
# can change inside the doc (does this change the encoding or not ?). To be
# frank, this is not entirely my fault, the lyx format is a joke.
#
# As there is unfortunately no way to define the output file name, we have
# to use a temporary directory and link the input file in there.
# set variables
#LANG=C ; export LANG
#LC_ALL=C ; export LC_ALL
progname="rcllyx"
filetype=lyx
#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 lyx iconv
# 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/rcllyx_tmp$$
mkdir $tmpdir || exit 1
mkdir $tmpdir/rcllyxtmp || exit 1
cleanup()
{
# Note that we're using a constant part (rcllyxtmp), that hopefully
# guarantees that we can't do big mistakes here.
rm -rf $tmpdir/rcllyxtmp
rmdir $tmpdir
}
trap cleanup EXIT HUP QUIT INT TERM
workdir=$tmpdir/rcllyxtmp
case "$infile" in
*/*) ;;
*) infile=`pwd`/$infile;;
esac
binfile=`basename "$infile"`
ln -s "$infile" "$workdir/$binfile" || exit 1
lyxfile=$workdir/"$binfile"
textfile=$workdir/`basename "$binfile" .lyx`.txt
#echo binfile: $binfile;echo lyxfile: $lyxfile ; ls -l $lyxfile; echo textfile: $textfile
# Run lyx --export
lyx --export text "$lyxfile"
# Charset and language
formatline=`egrep '^\\\lyxformat ' "$lyxfile"`
if test -n "$formatline" ; then
set $formatline
format=$2
fi
charsetline=`egrep '^\\\inputencoding ' "$lyxfile"`
if test -n "$charsetline" ; then
set $charsetline
charset=$2
fi
langline=`egrep '^\\\language ' "$lyxfile"`
if test -n "$langline" ; then
set $langline
lang=$2
fi
#echo format: [$format] charset: [$charset] lang [$lang]
if test "$format" -ge 249 ; then
charset=utf-8
else
# try to guess the charset from the language: this is in no way guaranteed
# to work, the logic has built-in inconsistencies even beyond the numerous
# external ones (what if the ukrainian writer prefers koi8-r ?). This is a
# joke.
if test -z "$charset" -o "$charset" = default -o "$charset" = auto ; then
case "$lang" in
american|afrikaans|basque|catalan|danish|dutch|english|faeroese|finnish|french|galician|german|icelandic|irish|italian|norwegian|portuguese|spanish|swedish)
charset=iso-8859-1;;
czech|german|hungarian|polish|romanian|croatian|slovak|slovene)
charset=iso-8859-2;;
esperanto|galician|maltese|Turkish)
charset=iso-8859-3;;
estonian|latvian|lithuanian)
charset=iso-8859-4;;
bulgarian|byelorussian|macedonian|russian|serbian|ukrainian)
charset=iso-8859-5;;
arabic)
charset=iso-8859-6;;
greek)
charset=iso-8859-7;;
hebrew)
charset=iso-8859-8;;
#ISO-8859-9 - Latin 5 Same as 8859-1 except for Turkish instead of
#Icelandic. ? What is one to do :)
#ISO-8859-10 - Latin 6
lappish|nordic|eskimo|inuit|sami)
charset=iso-8859-10;;
albanian|german|english|basque|breton|catalan|danish|spanish|estonian|esthonian|faeroese|faroese|finnish|french|frisian|friesian|scottish|goidelic|irish|gaelic|galician|welsh|greenlandic|inuit|icelandic|italian|latin|dutch|norvegian|portuguese|romansch|romansh|friulian|ladin|swedish)
charset=iso-8859-15;;
*)
charset=iso-8859-1;;
esac
fi
fi
if test -n "$charset" ; then
inputcmd="iconv -f $charset -t UTF-8 -c -s"
else
inputcmd=cat
fi
#echo inputcmd: [$inputcmd]
cat <<EOF
<html>
<head>
<title>$title</title>
<meta http-equiv="Content-Type" content="text/html;charset=UTF-8">
</head>
<body>
<pre>
EOF
$inputcmd < "$textfile"
cat <<EOF
</pre>
</body>
</html>
EOF