Parent:
[bf0f72]
(diff)
Child:
[17393b]
(diff)
Download this file
rclsvg
138 lines (114 with data), 3.4 kB
#!/bin/sh
# @(#$Id: rclsvg,v 1.3 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 a Scalable Vector Graphics file
#================================================================
# set variables
LANG=C ; export LANG
LC_ALL=C ; export LC_ALL
progname="rclsvg"
filetype=svg
#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 iconv sed
encoding=`sed -ne '/<?xml/s/.*encoding="\([^"]*\).*/\1/p' < $infile`
if test X$encoding = X ; then encoding=UTF-8;fi
# We use several sed instances to make our life easier. Not good for
# performance, and a sed guru might be able to do better.
#
# The first sed makes sure each tag starts on a new line
# The second one selects the tags we're interested in.
# The last strips the tags, leaving only text.
#
# The whole thing wholly ignore issues like '<' inside quoted strings.
#
# We could/should add code to explicitly separate title and other
# metadata elements.
# Insert new line before each tag
sptagonline='s/</\
</g'
# Select tags
spselecttags='/<title/,/<\/title>/p
/<desc/,/<\/desc>/p
/<metadata/,/<\/metadata>/p
/<text/,/<\/text>/p'
# Strip tags
spstriptags='#n
/</{
:c
/>/!{
N
b c
}
/>/s/<.*>//g
}
/^[ ]*$/!p'
content=`sed -e "$sptagonline" < $infile | sed -ne "$spselecttags" | \
sed -ne "$spstriptags"`
(echo '<html><head>'
echo '<meta http-equiv="Content-Type" content="text/html;charset=UTF-8">'
echo '</head><body><pre>'
echo "$content"
echo '</pre></body></html>') \
| iconv -f $encoding -t UTF-8 -c -s
# exit normally
exit 0