#!/usr/bin/perl use CGI::Carp qw(fatalsToBrowser); # # csSearchPro - 061402 $basepath = './'; # 3.9 Multiple search terms with 'AND' fixed # 3.8 Duplicate search results fixed # 3.75 Changes # misc minor fixes # 3.6 Changes # -- Added ability to manually increase keywords with # # 3.5 Changes # -- Searches for multiple SSI methods when option is selected # 3.3 Changes # -- ability to use # -- partial searces or full word searches.. # -- ability for server side includes in the templates. # 3.0 Changes # -- Ability to log search terms # -- Improved Indexing ##################################################################### # # # Copyright © 1999-2001 CGISCRIPT.NET - All Rights Reserved # # # ##################################################################### # # # THIS COPYRIGHT INFORMATION MUST REMAIN INTACT # # AND MAY NOT BE MODIFIED IN ANY WAY # # # ##################################################################### # # When you downloaded this script you agreed to accept the terms # of this Agreement. This Agreement is a legal contract, which # specifies the terms of the license and warranty limitation between # you and CGISCRIPT.NET. You should carefully read the following # terms and conditions before installing or using this software. # Unless you have a different license agreement obtained from # CGISCRIPT.NET, installation or use of this software indicates # your acceptance of the license and warranty limitation terms # contained in this Agreement. If you do not agree to the terms of this # Agreement, promptly delete and destroy all copies of the Software. # # Versions of the Software # Only one copy of the registered version of CGISCRIPT.NET # may used on one web site. # # License to Redistribute # Distributing the software and/or documentation with other products # (commercial or otherwise) or by other than electronic means without # CGISCRIPT.NET's prior written permission is forbidden. # All rights to the CGISCRIPT.NET software and documentation not expressly # granted under this Agreement are reserved to CGISCRIPT.NET. # # Disclaimer of Warranty # THIS SOFTWARE AND ACCOMPANYING DOCUMENTATION ARE PROVIDED "AS IS" AND # WITHOUT WARRANTIES AS TO PERFORMANCE OF MERCHANTABILITY OR ANY OTHER # WARRANTIES WHETHER EXPRESSED OR IMPLIED. BECAUSE OF THE VARIOUS HARDWARE # AND SOFTWARE ENVIRONMENTS INTO WHICH CGISCRIPT.NET MAY BE USED, NO WARRANTY # OF FITNESS FOR A PARTICULAR PURPOSE IS OFFERED. THE USER MUST ASSUME THE # ENTIRE RISK OF USING THIS PROGRAM. ANY LIABILITY OF CGISCRIPT.NET WILL BE # LIMITED EXCLUSIVELY TO PRODUCT REPLACEMENT OR REFUND OF PURCHASE PRICE. # IN NO CASE SHALL CGISCRIPT.NET BE LIABLE FOR ANY INCIDENTAL, SPECIAL OR # CONSEQUENTIAL DAMAGES OR LOSS, INCLUDING, WITHOUT LIMITATION, LOST PROFITS # OR THE INABILITY TO USE EQUIPMENT OR ACCESS DATA, WHETHER SUCH DAMAGES ARE # BASED UPON A BREACH OF EXPRESS OR IMPLIED WARRANTIES, BREACH OF CONTRACT, # NEGLIGENCE, STRICT TORT, OR ANY OTHER LEGAL THEORY. THIS IS TRUE EVEN IF # CGISCRIPT.NET IS ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. IN NO CASE WILL # CGISCRIPT.NET' LIABILITY EXCEED THE AMOUNT OF THE LICENSE FEE ACTUALLY PAID # BY LICENSEE TO CGISCRIPT.NET. # # Credits: # Andy Angrick - Programmer - angrick@cgiscript.net # Mike Barone - Design - mbarone@cgiscript.net # # For information about this script or other scripts see # http://www.cgiscript.net # # Thank you for trying out our script. # If you have any suggestions or ideas for a new innovative script # please direct them to suggest@cgiscript.net. Thanks. # ######################################################################## # Config Variables # ######################################################################## ##################################################################### # # # Configuration variables # # # ##################################################################### (! -e "$basepath/setup.cgi")?($nosetup=1):(require("$basepath/setup.cgi")); (!$htmlpath)&&($htmlpath=$cgipath); (!$htmlurl)&&($htmlurl=$cgiurl); $in{'htmlurl'} = $htmlurl; $in{'cgiurl'} = $cgiurl."/csSearchPro.cgi"; $flock=1; $in{'cinfo'} = qq|

Powered by csSearchPro - © 2000,2001 CGIScript.net |; ##################################################################### # # # End Configuration Variables. # # # ##################################################################### require("$basepath/libs.cgi"); $| = 1; &main; exit; sub main{ print "Content-type: text/html\n\n"; ($ARGV[0] eq 'index')&&($ARGV[1] eq $username)&&(&CreateIndexStart); &getdata(); $in{'terms'} =~ s/([\\\|\(\)\[\]\{\}\^\$\*\+\?\.])/\\$1/g; ($in{'command'} eq '')&&($nosetup)&&(&Setup); ($in{'command'} eq '')&&(!$nosetup)&&(&ShowSearchForm); ($in{'command'} eq 'login')&&(&Login); ($in{'command'} eq 'query')&&(&Query); ($in{'command'} eq 'info')&&(&Info); ($in{'command'} eq 'savesetup')&&(($nosetup)?(&SaveSetup):(&PError("Error. Permission denied."))); #all require password below &GetLogin; ($in{'command'} eq 'manage')&&(&Manage); ($in{'command'} eq 'aae')&&(&AddExtExcludes); ($in{'command'} eq 'ade')&&(&AddDirectoryExcludes); ($in{'command'} eq 'ae')&&(&AddIgnore); ($in{'command'} eq 'as')&&(&AddSkip); ($in{'command'} eq 'gu')&&(&GoUp); ($in{'action'} eq '-=Skip/UnSkip All Selected=-')&&(&SkipSelected); ($in{'action'} eq '-=Ignore/UnIgnore All Selected=-')&&(&IgnoreSelected); ($in{'command'} eq "showadv")&&(&ShowAdv); ($in{'command'} eq "setstyles")&&(&SetStyles); ($in{'command'} eq "showcolor")&&(&ShowColor); ($in{'command'} eq "setcolor")&&(&SetColor); ($in{'command'} eq "index")&&(&CreateIndexStart); ($in{'command'} eq "clearlog")&&(&ClearLog); ($in{'command'} eq "viewlogs")&&(&ViewQueryLog); exit; } sub ShowAdv{ &GetAdvSettings; $in{'S'.$in{'style'}}='checked'; (!$in{'extinc'})?($in{'EXTY'}='checked'):($in{'EXT'.$in{'extinc'}}='checked'); $in{'directlinks'} = qq| |; $in{'directlinksnojs'} = qq| |; &PageOut("$htmlpath/t_search_advanced_settings.htm"); exit; } sub GetAdvSettings{ (-e "$htmlpath/styles.pl")&&(require "$htmlpath/styles.pl"); } sub ShowColor{ $in{'c'.$in{'cc'}} = checked; &PageOut("$htmlpath/color_selector.htm"); exit; } sub SetColor{ open(ADV,"<$htmlpath/styles.pl"); open(TMP,">$htmlpath/styles.pl.tmp"); while(){ (!/'\$in{'$in{'fieldname'}'}'/)&&(!/1;/)&&(print TMP $_); } $in{'colorselect'} =~ s/\\//g; $in{'colorselect'} =~ s/\"//g; print TMP "\$in{'$in{'fieldname'}'} = \"$in{'colorselect'}\";\n1;\n"; close ADV; close TMP; @fi = stat("styles.pl.tmp"); rename("styles.pl.tmp","styles.pl") unless ($fi[7] < 1); print <<"EOF"; EOF exit; } sub SetStyles{ open(STYLES,">$htmlpath/styles.pl")||die print "$!: styles.pl"; foreach $i (keys(%in)){ next if (($i eq 'command')||($i eq 'cgiurl')); next if (($i eq 'UserName')||($i eq 'Password')); next if (($i eq 'basemanageurl')||($i eq 'cinfo')); next if (($i eq 'database')||($i eq 'imagedir')); next if (($i eq 'imagerealdir')||($i eq 'images2')); next if (($i eq 'format')||($i eq 'managementname')); next if (($i eq 'managementuser')||($i eq 'managementemail')); next if (($i eq 'exportdir')||($i eq 'managementemail')); next if (($i eq 'htmlurl')||($i eq 'cgiurl')); next if ($i eq 'terms'); $in{$i} =~ s/'//g; print STYLES "\$in{'$i'}=\'$in{$i}\';\n"; } print STYLES "1;\n"; close STYLES; print<<"EOF"; EOF exit; } sub Login{ &PageOut("$htmlpath/t_login.htm"); exit; } sub GetLogin{ &GetCookies; $in{'UserName'} = $cookie{'UserName'}; $in{'PassWord'} = $cookie{'PassWord'}; if(!$in{'UserName'}){ &PageOut("$htmlpath/t_login.htm"); exit; } else{ (($in{'UserName'} ne $username)||(($in{'PassWord'} ne $password)))&&(&PError("Error. Invalid username or password")); } } sub QueryI{ &GetAdvSettings; if($in{'parsessi'}){ require("$cgipath/ssi.cgi"); } (!$in{'page'})&&($in{'page'} = 1); $start = ($in{'page'}*$in{'entriesperpage'})-$in{'entriesperpage'}; $end = $start + $in{'entriesperpage'}; (!$in{'terms'})&&(&PError("Error. Please enter something to search for.")); &SearchByIndex; (!$in{'count'})&&($in{'count'} = '0'); (!$in{'searchresults'})&&($in{'searchresults'} = '
No se han encontrado resultados para su consulta.'); $in{$in{'mcase'}} = 'selected'; $in{$in{'mbool'}} = 'selected'; ($count)&&(&GetNlinks); $in{'count'} = $count; $in{'count'} = "Su busqueda retornó $in{'count'} resultados."; $in{'terms'} =~ s/\\//g; ($in{'template'})?(&InsertPage($in{'template'},"$htmlpath/t_searchresults.htm")):(&PageOut("$htmlpath/t_searchresults.htm")); exit; } sub Query{ &GetAdvSettings; &LogQuery; ($in{'useindex'})&&(&QueryI); if($in{'parsessi'}){ require("$cgipath/ssi.cgi"); } (!$in{'page'})&&($in{'page'} = 1); $start = ($in{'page'}*$in{'entriesperpage'})-$in{'entriesperpage'}; $end = $start + $in{'entriesperpage'}; (!$in{'terms'})&&(&PError("Error. Please enter something to search for.")); &LoadVars; &GetGlobalIgnore; chdir($rootpath); &GetFiles("."); ##start actual search foreach $i (sort keys %ftsearch){ &SearchFile($i); } (!$in{'count'})&&($in{'count'} = '0'); (!$in{'searchresults'})&&($in{'searchresults'} = '
No se han obtenido resultados de su búsqueda'); $in{$in{'mcase'}} = 'selected'; $in{$in{'mbool'}} = 'selected'; ($count)&&(&GetNlinks); $in{'count'} = $count; $in{'count'} = "Su busqueda retornó $in{'count'} resultados."; ($in{'template'})?(&InsertPage($in{'template'},"$htmlpath/t_searchresults.htm")):(&PageOut("$htmlpath/t_searchresults.htm")); exit; } sub CreateIndexStart{ &GetAdvSettings; &GetGlobalIgnore; $cid=0; $pid=0; &LoadVars; if($in{'parsessi'}){ require("$cgipath/ssi.cgi"); } print "

Indexando archivos.. por favor paciencia\n";

chdir($rootpath);
undef $/;

&GetFiles(".");

foreach $i (keys %ftsearch){
  &CreateIndex($i);
  }

#uninit
open(IDX,">$htmlpath/site.idx");
foreach my $i (keys %site){
print IDX "$i:$site{$i}\n";
}
close IDX;

open(IDX,">$htmlpath/page.idx");
foreach my $i (keys %page){
print IDX "$page{$i}:$i\n";
}
close IDX;

dbmclose %site;
dbmclose %page;

close DB;

print <<"EOF";

EOF
exit;
}


sub CreateIndex{
local($file) = @_;
$cg=0;
undef @words;
print "Indexing file $file =->";
open(FILE,"<$rootpath/$file")||print "$!:$rootpath/$file
";
$data = ;
close FILE;

if($in{'parsessi'}){
  &CheckForSSI;
  }
&FormatPage;
#index the page
if(!$page{$file}){
  $pid++;
  $page{$file} = $pid;
  $cid = $pid;
  }
else{
  $cid = $page{$file};
  }

my @words = split(/\s/,$data);

foreach my $m (@words){
next if (!$m);
$m =~ s/[\,\.\/\!\@\#\$\%\^\&\*\(\)\_\=\+\{\}\[\]\|\\\"\:\;\?\/\>\<\~\`]//g;
next if (length("$m") < 2);
$mw{$m}++;
}

foreach my $k (keys %mw){
$cg++;
($site{$k})?($site{$k} .= "\t$cid-$mw{$k}"):($site{$k} .= "$cid-$mw{$k}");
}
undef %mw;
print " $cg words\n";
}

sub SearchByIndex{
@aq = split(/\s+/,$in{'terms'});

#load in vars
open(DB,"<$htmlpath/site.idx");
while(){
chomp;
($k,$v) = split(":",$_);
$index{$k} = $v;
}
close DB;

open(DB,"<$htmlpath/page.idx");
while(){
chomp;
($k,$v) = split(":",$_);
$page{$k} = $v;
}
close DB;

foreach $j (keys %index){

foreach $i (@aq){
  if($in{'mcase'} eq 'Insensitive'){
    
  ($in{'partial'})?(($j =~ /$i/i)&&($fnd{$i} .= "\t$index{$j}")):(($j =~ /\b$i\b/i)&&($fnd{$i} .= "\t$index{$j}"));
    
    }
  else{
    ($in{'partial'})?(($j =~ /$i/)&&($fnd{$i} .= "\t$index{$j}")):(($j =~ /\b$i\b/)&&($fnd{$i} .= "\t$index{$j}"));
    }

  }

}#end while


##got all the pages

if($in{'mbool'} eq 'AND'){
  &SortAndPages;
  }
else{
  &SortOrPages;
  }

}

sub SortOrPages{

foreach my $i (keys %fnd){
(@pages) = split("\t",$fnd{$i});
foreach my $s (@pages){
  my($pid,$n) = split("-",$s);
  my $z = sprintf("%.10d",$n).$page{$pid};
  $fndt{$z}=1;
  }
}

foreach my $x (sort {$b cmp $a} keys %fndt){
$x =~ s/^\d\d\d\d\d\d\d\d\d\d//;
next if(!$x);
&DisplayOneFromIndex("$rootpath/$x");
}

}


sub SortAndPages{

foreach my $i (keys %fnd){
(@pages) = split("\t",$fnd{$i});
foreach my $s (@pages){
  my($pid,$n) = split("-",$s);
  my $z = sprintf("%.10d",$n).$page{$pid};
  $i =~ tr/a-z/A-Z/;
  $fndt{$z}=1;
  $fndtx{$page{$pid},$i}=1;
  }
}

foreach my $x (sort {$b cmp $a} keys %fndt){
$getcha=1;
foreach $z (@aq){
$z =~ tr/a-z/A-Z/;
$x =~ s/^\d\d\d\d\d\d\d\d\d\d//;
(!$fndtx{$x,$z})&&($getcha=0)&&(last);
}
next if(!$x);
&DisplayOneFromIndex("$rootpath/$x") if ($getcha==1);
}

}

sub DisplayOneFromIndex{
local($file) = @_;

return if ($printed{$file});
$printed{$file} = 1;

##start pagination;
$count++;
if(($count > $start) & ($count <= $end)){

open(FILE,"<$file")||print "$!:$file
";
while(){
$data .= $_;
}
close FILE;

&OutputPage;

}##end pagination

}

sub SearchFile{

undef $/;
local($file) = @_;

open(FILE,"<$file")||print "$!:$file
";
$data = ;
close FILE;
$tdata = $data;

if($in{'parsessi'}){
&CheckForSSI;
}

&FormatPage;

(!$in{'partial'})&&($b='\b');
if($in{'mbool'} eq 'AND'){
  if($in{'mcase'} eq 'Insensitive'){

    if($data =~ /$b$in{'terms'}$b/i){
      &DisplayFound;
      #print "Found: $in{'terms'}
";
      }
  }
  else{
    if($data =~ /$b$in{'terms'}$b/){
      &DisplayFound;
      #print "Found: $in{'terms'}
";
      }
  }
}
else{
  @aq = split(/\s+/,$in{'terms'});
  foreach $k (@aq){
    if($in{'mcase'} eq 'Insensitive'){
    
      if($data =~ /$b$k$b/i){
        &DisplayFound;
        #print "Found: $in{'terms'}
";
        }
    }
    else{
      if($data =~ /$b$k$b/){
        &DisplayFound;
        #print "Found: $in{'terms'}
";
        }
   }
  }
}

}

sub DisplayFound{
$data = $tdata;
$count++;
if(($count > $start) & ($count <= $end)){
  &OutputPage
  }# end next page
}

sub GetNlinks{
((!$in{'page'})||($in{'lpage'} < 0))&&($in{'lpage'} = 1);
$nstart = ($in{'lpage'}+10)-10;
$nend = $nstart + 10;
$sstring = "command=$in{'command'}&terms=$in{'terms'}&mbool=$in{'mbool'}&mcase=$in{'mcase'}";
$sstring =~ s/([^\w&=])/'%'.sprintf("%.2x",ord($1))/ge;

$npages = int($count/$in{'entriesperpage'});
(($count/$in{'entriesperpage'}) > $npages)&&($npages++);

for $i (1..$npages){
  if(($i > $nstart) & ($i <= $nend)){
    ($i eq $in{'page'})?($link .= " $i "):($link .= " $i ");
    }
  }
$nend;
$pend = $nend - 20;
($npages > $nend)&&($link .= " [Next] ");
($nend > 11)&&($link = " [Prev] " . $link);
(!$link)&&($link='0');
$link = "Páginas con resultados: " . $link;
$in{'link'} = $link;
}

sub DeHTML{

    #########################################################
    # remove anything between  
    #########################################################
    $data =~ s/.*//gsi;
    #########################################################
    # next we'll remove all the