Add support for indexing fictionbook documents
This commit is contained in:
Jean-Francois Dockes
parent
62da0f8b02
commit
47f76c6c42
140
src/filters/rclfb2
Executable file
140
src/filters/rclfb2
Executable file
@ -0,0 +1,140 @@
|
||||
#!/bin/sh
|
||||
# @(#$Id: rclopxml,v 1.3 2008-10-08 08:27:34 dockes Exp $ (C) 2004 J.F.Dockes
|
||||
#================================================================
|
||||
# rclfb2
|
||||
# Extract text from an fb2 ebook (xml)
|
||||
#================================================================
|
||||
|
||||
# set variables
|
||||
LANG=C ; export LANG
|
||||
LC_ALL=C ; export LC_ALL
|
||||
progname=rclfb2
|
||||
filetype=fb2
|
||||
|
||||
|
||||
#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
|
||||
|
||||
xsltproc - $infile <<EOF
|
||||
<?xml version="1.0"?>
|
||||
<xsl:stylesheet version="1.0"
|
||||
xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
|
||||
xmlns:fb="http://www.gribuser.ru/xml/fictionbook/2.0"
|
||||
exclude-result-prefixes="fb"
|
||||
>
|
||||
|
||||
<xsl:output method="html" encoding="UTF-8"/>
|
||||
|
||||
<xsl:template match="/fb:FictionBook">
|
||||
<html>
|
||||
<xsl:apply-templates select="fb:description"/>
|
||||
<xsl:apply-templates select="fb:body"/>
|
||||
</html>
|
||||
</xsl:template>
|
||||
|
||||
<xsl:template match="fb:description">
|
||||
<head>
|
||||
<xsl:apply-templates select="fb:title-info"/>
|
||||
</head><xsl:text>
|
||||
</xsl:text>
|
||||
</xsl:template>
|
||||
|
||||
<xsl:template match="fb:description/fb:title-info">
|
||||
<xsl:apply-templates select="fb:book-title"/>
|
||||
<xsl:apply-templates select="fb:author"/>
|
||||
</xsl:template>
|
||||
|
||||
<xsl:template match="fb:description/fb:title-info/fb:book-title">
|
||||
<title> <xsl:value-of select="."/> </title>
|
||||
</xsl:template>
|
||||
|
||||
<xsl:template match="fb:description/fb:title-info/fb:author">
|
||||
<meta>
|
||||
<xsl:attribute name="name">author</xsl:attribute>
|
||||
<xsl:attribute name="content">
|
||||
<xsl:value-of select="fb:first-name"/><xsl:text> </xsl:text>
|
||||
<xsl:value-of select="fb:middle-name"/><xsl:text> </xsl:text>
|
||||
<xsl:value-of select="fb:last-name"/>
|
||||
</xsl:attribute>
|
||||
</meta>
|
||||
</xsl:template>
|
||||
|
||||
<xsl:template match="fb:body">
|
||||
<body>
|
||||
<xsl:apply-templates select="fb:section"/>
|
||||
</body>
|
||||
</xsl:template>
|
||||
|
||||
<xsl:template match="fb:body/fb:section">
|
||||
<xsl:for-each select="fb:p">
|
||||
<p><xsl:value-of select="."/></p>
|
||||
</xsl:for-each>
|
||||
</xsl:template>
|
||||
|
||||
</xsl:stylesheet>
|
||||
EOF
|
||||
@ -89,6 +89,7 @@ text/html = internal
|
||||
text/plain = internal
|
||||
text/rtf = exec unrtf --nopict --html;charset=iso-8859-1;mimetype=text/html
|
||||
text/x-c = internal
|
||||
text/x-fictionbook = exec rclfb2
|
||||
text/x-gaim-log = exec rclgaim
|
||||
text/x-html-sidux-man = exec rclsiduxman
|
||||
text/x-mail = internal
|
||||
@ -151,6 +152,7 @@ text/html = html
|
||||
text/plain = txt
|
||||
text/x-c = source
|
||||
text/x-c++ = source
|
||||
text/x-fictionbook = document
|
||||
text/x-html-sidux-man = sidux-book
|
||||
text/x-mail = message
|
||||
text/x-man = document
|
||||
@ -187,6 +189,7 @@ text = \
|
||||
text/rtf \
|
||||
text/x-c \
|
||||
text/x-c++ \
|
||||
text/x-fictionbook \
|
||||
text/x-html-sidux-man \
|
||||
text/x-man \
|
||||
text/x-python \
|
||||
|
||||
@ -107,6 +107,8 @@
|
||||
.tiff = image/tiff
|
||||
.tif = image/tiff
|
||||
|
||||
.fb2 = text/x-fictionbook
|
||||
|
||||
# A list of suffixes (name endings) that we don't want to touch at all.
|
||||
# Having these explicitely listed speeds things up a bit by avoiding
|
||||
# unneeded decompression or 'file' calls. File names still get indexed if
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user