search.in

harvest是一个下载html网页得机器人

#!@PERL@
#############################################################################
#
#  search.cgi - Customizable WWW Interface to the Harvest Broker
#
#  Usage: Called as a CGI proccess from httpd
#
#############################################################################
#
#  Harvest Indexer http://harvest.sourceforge.net/
#  -----------------------------------------------
#
#  The Harvest Indexer is a continued development of code developed by
#  the Harvest Project. Development is carried out by numerous individuals
#  in the Internet community, and is not officially connected with the
#  original Harvest Project or its funding sources.
#
#  Please mail [e-mail] if you are interested in participating
#  in the development effort.
#
#  This program is free software; you can redistribute it and/or modify
#  it under the terms of the GNU General Public License as published by
#  the Free Software Foundation; either version 2 of the License, or
#  (at your option) any later version.
#
#  This program is distributed in the hope that it will be useful,
#  but WITHOUT ANY WARRANTY; without even the implied warranty of
#  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
#  GNU General Public License for more details.
#
#  You should have received a copy of the GNU General Public License
#  along with this program; if not, write to the Free Software
#  Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
#
#  Changes by Harald Weinreich <[e-mail]> for version 1.8 are
#  marked with "h.weinreich"

$ENV{'HARVEST_HOME'} = "@prefix@"
    unless defined($ENV{'HARVEST_HOME'});

# Uncomment this if you have problems with European 8-BIT characters on SunOS.
#$ENV{'LANG'} = "C";
#$ENV{'LC_CTYPE'} = "iso_8859_1";

# Set this to a location for search.log, or to /dev/null
#$BQLOG = "@prefix@/logs/search.log";
$BQLOG = "/dev/null";

$rcsid = '$Id: search.in,v 2.30 2002/08/30 12:25:00 sxw Exp $';

# $expire
# Temporary files are deleted after this many hours.  Or, more precisely, when
# this program is run, files older than $expire hours are deleted.
#
# $tmp_dir
# This is the location where temporary files are kept. This directory needs a
# permissive access mode, so that the 'nobody' user (or whatever
# user httpd uses) can write and delete files there.  It is a fatal error if
# this directory does not exist, or cannot be written in.
#

###### configure these locally! #######
$expire  = 1;				# expiration interval, in hours
$tmp_dir = "$ENV{HARVEST_HOME}/tmp";	# dir for temp files
######    end configuration     #######

$harvestIcon = "/Harvest/brokers/images/harvest-ic1.gif";    # logo
$weightIcon  = "/Harvest/brokers/images/harvest-weight.gif"; # weight image

@X = split ('/', $0);
$MYNAME = pop @X;
$DIR = join ('/', @X);
$ENV{'TMPDIR'} = "/tmp" unless defined($ENV{'TMPDIR'});
unshift(@INC, "$ENV{'HARVEST_HOME'}/lib");

not_configured() unless (-d $ENV{'HARVEST_HOME'});

require 'socket.ph';	# not sys/socket.ph, we use $HARVEST_HOME/lib/socket.ph

$debug		= 0;
$hp_url		= '';
$brokers	= $ENV{'HARVEST_HOME'} . '/brokers/Brokers.cf';

# show weight of result and navigation bar by Javier Masa Marin <[e-mail]>
$weight		= 0; # weight of current object
$maxWeight	= 0;
$totalPages	= 0; # number of Result pages

# ===== MAIN =================================================================

foreach $sig ('HUP', 'QUIT', 'TSTP', 'TERM', 'ABRT') {
	$SIG{$sig} = 'sigdie';
}

foreach $sig ('ALRM') {
	$SIG{$sig} = 'sigharddie';
}

# Parse the CGI request.
#
%RQ = &get_request;
$debug = 1 if defined $RQ{'debug'};
foreach $key (keys %RQ) { $RQ{$key} =~ s/\n/ /g; }

# use unbuffered output
#
$unbuffered = 1;

# If name is nph-search.cgi, more headers have to be sent.
#
$nph = $0 =~ /nph-/;

# Send the MIME header *now* if we're in debug mode this means it will always
# be text/html, but means when can see debugging information from parse_config
#
send_header() if $debug;

# Read in the 'master config' file.
#
&parse_config ("$DIR/lib/search.cf");
&parse_defaults;

$CFG{'rcsid'} = $rcsid;

# Parse a broker-specific config file (included as the following HTML - )
#  <INPUT TYPE="hidden" NAME="brokerqueryconfig" VALUE="foo.cf">
#
print "opening $DIR/lib/",&option('brokerqueryconfig'),"\n" if ($debug);
&parse_config ("$DIR/lib/".&option('brokerqueryconfig'))
	if ( -f "$DIR/lib/".&option('brokerqueryconfig') );
&parse_defaults;

# determine character set for the HTTP-header
#
# get the charset from search form or .cf-file
$charset = &option('charset');
	if ($charset eq "") {$charset = "ISO-8859-1";} # default

# We can't send the MIME header *until* we've read the broker-specific file
# (as they might change it in there) - so we have to wait until now to send
# it.
#
send_header() unless $debug;

# Now - once we're sure that a MIME header has been sent, we can validate
# their input
#
&fatal ('NoQuery')	if (%RQ == ());
&fatal ('rcsid')	if ($RQ{'version'} ne "");
&dump_array (%RQ)	if ($debug);
&dump_array (%DEF)	if ($debug);
#&dump_array (%ENV)	if ($debug);

# EXTRACT QUERY OPTIONS
#
$lifetime	= &option('lifetime');
$userquery	= &option('query');
$category	= &option('category');
$userclass	= &option('class');
$caseflag	= &option('caseflag')     eq 'on' ? 1 : 0; # case sensitivity
$wordflag	= &option('wordflag')     eq 'on' ? 1 : 0; # match on word boundary
$csumflag	= &option('csumflag')     eq 'on' ? 1 : 0; # show links to indexing data
$opaqueflag	= &option('opaqueflag')   eq 'on' ? 1 : 0; # return opaque data (matched lines)
$descflag	= &option('descflag')     eq 'on' ? 1 : 0; # return object description
$noregexflag	= &option('noregexflag')  eq 'on' ? 1 : 0; #
$maxresult	= &option('maxresultflag');                # max. num of result lines to be returned by broker
$maxfiles	= &option('maxobjflag');                   # max. num of objects to be returned by broker
$maxlines	= &option('maxlineflag');                  # max. num of matched lines per object
$perpage 	= &option('perpageflag');                  # show n objects per page (0 = don't split output)
$page		= &option('pageflag');                     # show page nr x
$errors		= &option('errorflag');                    # number of errors allowed
$weightflag	= &option('weightflag')   eq 'on' ? 1 : 0; # show weight balls
$broker 	= &option('broker') ||
		  &option('host');                         # name of broker
#$version	= &option('version')      eq "" ? 0 : 1;   # ?
$hp_url		= &option('hp_url');                       # URL of search page
$sort		= &option('sort');                         # sort options

@atts		= split (/\s+/, &option('attribute'));     # list of attributes to be displayed
foreach $a (@atts) {
	$attributes .= " #attribute \"$a\"";
}

# Build filter strings. Filters are e.g. used to specify a host.
# They are attached to the query but not displayed as query string
# Filters consist of a query string and a user hint to be displayed
# on the search results page, seperated by a pipe symbol. Example:
# <INPUT NAME="filter" value="type: html|<H3>Search only HTML documents</h3>" type="hidden">
# See query-glimpse.html for more examples - by h.weinreich
@filter     = split (/\0/, &option('filter'));
$filter     = "";             # The filter is attached to $query
$filterhint = "";             # Filterhint is the text shown in the output to the user...
foreach (@filter) {
	@t = split(/\|/);
	$filter     .= " AND $t[0]"   if $t[0];
	$filterhint .= "$t[1]\n"      if $t[1];
}

# SECURITY CHECKS AND TRANSLATION ON BROKER HOST,PORT
#
$errmsg = <<"EOF";
<TITLE>$broker not found</TITLE>
<PRE>
$MYNAME doesn't know the broker

    <B>$broker</B>

Either it is not in the allowed list, or perhaps
$brokers is not readable.
</PRE>
EOF
&fatal($errmsg)
	unless (@hostport = &get_host_port ($broker));

# HACKS FOR BROKEN LYNX BROWSER
#
$errors = 0 if ($errors eq 'None');
$errors = 1 if ($errors eq '1 Error');
$errors = 2 if ($errors eq '2 Errors');

# SET THE LIFETIME
#
$BQlife = $CFG{'Timeout'};
$BQlife = $lifetime + 300 if ($lifetime ne "");
alarm ($BQlife);

### # SANITY CHECKS
### #
### &fatal ('NoReplica')	if ($host eq "No Replicas");
### &fatal ('Misconfig')	if ($host eq "" || $port == 0);

# CHECK QUERY STRING FOR COMMON MISTAKES
#
$userquery =~ s/^\s+//;			# remove leading whitespace
$userquery =~ s/\s+$//;			# remove trailing whitespace

$userquery = &entities($userquery); # translate SGML entities

unless ( $userquery =~ /\s+and\s+/i || $userquery =~ /\s+or\s+/i
      || $userquery =~ /:\s+/  || $userquery =~ /\"/) {
    if ($category ne "anytext") {
	@X = split (/\s+/, $userquery);
	for ($i=0; $i<=$#X; $i++) {
	    # Support for + and - in front of search words - h.weinreich
	    $X[$i] =~ s/^\+(.*)/$1/;      # remove leading +
	    $X[$i] =~ s/^\-(.*)/NOT $1/;  # replace leading - by NOT
	    $X[$i] =~ s/\,//;             # remove commata: not supported by broker
	    # put quotes around unknown characters
	    $X[$i] = "\"$X[$i]\"" if ($X[$i] =~ /[^ \w\d-]/);
	    $X[$i] = "$category:$X[$i]" if (($category ne "any") && ($category ne ""));
	}
	$userquery = join (' AND ', @X);
    } else {
	@X = split (/\s+/, $userquery);
	for ($i=0; $i<=$#X; $i++) {
	    $X[$i] = "\"$X[$i]\"" if ($X[$i] =~ /\W/);
	    $X[$i] = "(keywords:$X[$i] OR title:$X[$i] OR body:$X[$i] OR headings:$X[$i] OR address:$X[$i])"
	}
	$userquery = join (' AND ', @X);
    }
}

# BUILD QUERY STRING
#
$query  = "";
$query .= $userclass . " AND " if ($userclass ne "");
$query .= $userquery;

# BUILD BROKER QUERY
#
$bquery = "#USER";
$bquery .= " #opaque"				if ($opaqueflag);
$bquery .= " #desc"				if ($descflag);
$bquery .= " #index timeout $lifetime"		if ($lifetime ne "");
$bquery .= " #index error $errors"		if ($errors ne "");
$bquery .= " #index maxresult $maxresult"	if ($maxresult ne "");
$bquery .= " #index maxfiles $maxfiles"		if ($maxfiles ne "");
$bquery .= " #index maxlines $maxlines"		if ($maxlines ne "");
$bquery .= " #index case";
$bquery .= $caseflag ? " insensitive" : " sensitive";
$bquery .= " #index matchword"			if ($wordflag);
$bquery .= " #index noregex"			if ($noregexflag);
$bquery .= $attributes;
$bquery .= " #END ";
$bquery .= $query;
$bquery .= $filter;

$simple_query = $1 if ($query =~ /^.*partial-text\s*:\s+"(.*)".*$/io);

# Call Init Function
#
eval $CFG{'InitFunction'} if (defined ($CFG{'InitFunction'}));

# DO THE QUERY
#
$html_query = &html_escape ($query);
print &expand ($CFG{'ResultHeader'});
$connected = 0;

# Split query into words for highlighting...
$searchwords = &option('query');
$searchwords =~ s/-|\.|\,|;/ /g;
$searchwords =~ s/ AND | OR | NOT |\(|\)|:|\"|\+/ /g;
@searchwords = split(/ +/,$searchwords);

&cleanup(); # remove outdated files...

&do_query ($bquery, @hostport);

exit 0;     # END OF PROGRAM

# ===== SUBROUTINES ==========================================================

# Send the appropriate HTTP headers

sub send_header {
    my $content = (defined $CFG{'ContentType'}) ?
	&expand($CFG{'ContentType'}) : "text/html";
    $content .= "; charset=".$charset if ($charset ne "");
    $|=1 if ($unbuffered);   # use unbuffered output.

    if ($nph) {
#	$|=1;

	print "HTTP/1.0 200 OK\n";
	print "Server: $ENV{SERVER_SOFTWARE}\n";
    }
    print "Content-Type: $content\n\n";
}

# broker_host_port:
#
# If given a name, return the corresponding (host,port) pair.
# If given a host:port string, make sure it is a valid broker.

sub get_host_port {
	local ($broker) = @_;
	local ($name,$host,$port);
	local ($pattern);
	local (@hostport) = ();

	if ($broker =~ /([^:]+):(\d+)/) {		# given host:port
		$host = $1;
		$port = $2;
		return ($host,$port) unless ( -r $brokers );
		$pattern = '\S+\s+' . $host . '\s+' . $port;
	}
	else {						# given a name
		return () unless ( -r $brokers );	# cant translate name
		$broker_re = $broker;
		$broker_re =~ s/\W/\\$&/g;		# escape specials
		$pattern = '^' . $broker_re . '\s+\S+\s+\d+';
	}

	open (brokers) 	|| &fatal ("$brokers: $!\n");
	while (<brokers>) {
		chop;
		s/#.*//;				# strip comments
		s/^\s+//;				# leading whitespace
		s/\s+$//;				# trailing whitespace
		next unless (/$pattern/io);
		($name,$host,$port) = split;		# found match
		push (@hostport, $host);		# add host to array
		push (@hostport, $port);		# add port to array
	}
	close brokers;
	return (@hostport);				# not found
}


#### cleanup: remove old temporary files - h.weinreich

sub cleanup {
    # (if $expire=12, then 12/24 sets a 12 hour expiration cycle)
    if (chdir "$tmp_dir/$broker") {
	foreach $file (<[a-z]*\.search>) {
	    unlink "$file" if -M $file > $expire/24;
	}
    }
}


#### do_query: send query to broker and write it to a temporary file
####           or read from temporary file if it already exists.

sub do_query {
	local ($bquery, @hostport) = @_;
	local ($nobjects) = 0;
	local ($nopaquelines) = 0;

	# The temporary files search results are used as buffers to reduce requests
	# to the broker, e.g. necessary for splitted result pages - h.weinreich

	local ($filename) = &get_hashcode($bquery); # create filename for the buffer file
	$filename = "$tmp_dir/$broker/$filename.search";
	local $tempfile_exists = 0;

	if (-e $filename) { # does temp file exist?
	    open($INPUT, "<$filename") || &fatal ("Could not open file for read: $filename\n");
	    my $oldquery = (<$INPUT>);
	    chop $oldquery;
	    if ($oldquery eq $bquery) { # hash code OK
		$tempfile_exists = 1;   # use tempfile as <INPUT>
	    } else {
		close $INPUT;
	    }
	}

	print "$tempfile_exists<BR>" if ($debug);
	print "$filename<BR>" if ($debug);

	if (!$tempfile_exists) { # tempfile does not exist: use broker as <INPUT>
	    # read from broker and write to tempfile - h.weinreich

	    while ($#hostport > $[) {
		$host = shift (@hostport);
		$port = shift (@hostport);
		if ($INPUT = &client_socket ($host, $port)) {
		    $connected = 1;
		    last;
		}
	    }
	    &broker_down ($host, $port) unless ($connected);

	    print "Sending <PRE> $bquery </PRE> to $host:$port<HR>\n" if ($debug);

	    alarm(300);    # after 5 minutes just kill it, CERN httpd won't do it

	    # Send query to broker
	    print $INPUT $bquery;

	    # Open temporary file for output
	    open(TEMPFILE, ">$filename") || \
		&fatal ("Could not open temporary file for write: $filename\n");

	    # Write query string first
	    print TEMPFILE "$bquery\n";
	}

	# Read input and pre-format for output.
	while (<$INPUT>) {

	    # print to tempfile if $INPUT is broker
	    print TEMPFILE;

	    if ($debug) {
		chop;
		print "|$_|\n";
	    }

	    # Ignore status messages
	    next if (/^200 -/o);

	    # Read Broker homepage URL (see file admin/broker.conf)
	    if (/^126 - (.*)$/o) {