#!/usr/bin/perl

###############################################################################
#
#                                IC-SiteSearch 1.7
#
# Program name: IC-SiteSearch
# Program version: 1.7
# Program name defined: 'Internet Coders Site Search' version 1.7
# Program purpose: Providing a way for customers to search your website.
#
#
# Written by: shawn@internetcoders.com
# Downloaded at: www.internetcoders.com
# Date released: 05/27/2002
# Last modified: never
# 
# System requirements: Linux,Perl
#
# Liability: InternetCoders.com and or all individuals of InternetCoders.com
#            are in no way liable for anything that this program may do while
#            in use modified/unmodified. You the person choosing to run this
#            program take full responsibilites for this code and everything
#            it does. It's simple, you don't want to take responsibility for
#            this code, then don't use it
#
# Support: If you have any questions about this program please refer to
#          the documentation/instructions that came with the program. Since
#          this program was free there is no e-mail or phone support. However
#          support is available if you pay for it, or if you pay for cusomizations.
#          To pay for support or cusomizations please visit our website at
#          http://www.internetcoders.com
#
# Distribution: Since you did not write this code, you are not allowed to distribute
#               it modified or unmodified. If you want someone to have this code then
#               tell them to download it from InternetCoders.com, if it is
#               not on the InternetCoders.com website at that time then that
#               person is out of luck and will not be able to download it until
#               we post it again, unless you get permission from InternetCoders.com
#
# Modifications: You are allowed to fully modify any of this code you like. Please
#                refer to the Distribution section above before giving away copies
#                of this code whether modified or unmodified.
#
###############################################################################





###############################################################################
# PROGRAM VARIABLES TO CONFIGURE BELOW
###############################################################################

#This should be the directory/path that you would like to have this program search.
#Put the exact path starting with the root directory /.
#Be sure to add a / at the end or else the program will not display the found
#results from the search properly.
$searchDirectory = '/home/httpd/cliff69/nobucks.com/';

#This should point to the same directory as above only as it would be typed in a
#web browser as a url. Point this url to the same directory that is being searched.
#This will be used to create a link to the pages being displayed by the search results.
#Be sure to add the trailing / at the end or else you the program will not display the found
#results from the search. Be sure to add a / at the end.
$hyperLink = 'http://www.nobucks.com/';

#This is the color that you would like the hypertext links to be. These are the links
#that point to the pages returned from the search results.
$hypertextLinkColor = 'blue';

#This is where you specify the types of files that you would like to have searched by this
#program. Seperate your extensions with a pipe symbol which looks like this | .
$searchFileTypes = '.html|.htm|.txt';

#This is the width of the search text field. Put the width that you want this text field
#to be displayed as.
$formFieldSize = '25';

#This is the text that you would like to have next to the search field box. The
#default word that shows up is Search.
$searchFieldText = 'Search Site ';

#To exclude the html tags from being searched set this variable to 1. If you want
#this program to search the html tags as well then set this variable to 0. Keep in
#mind that html tags contain words and this may throw off your search results. Also
#keep in mind that if you choose to not search the html and you have missing html tags
#then your search may not be as accurate either. Recommended that you set this to 0, don't
#remove the html from the search. Most people have missing html tags and this may rip the
#html incorrectly.
#1 turns html remove on. 0 turns remove off.
$removeHtml = '0';

#Point this to the header file if you have one that you would like to have displayed.
#Leave this blank if you do not have a header file to be displayed.
$header = '';

#Point this to the footer file if you have one that you would like to have displayed.
#Leave this blank if you do not have a footer file to be displayed.
$footer = '';

###############################################################################
# PARSE INCOMING DATA AND FIGURE OUT WHAT TO DO WITH IT
###############################################################################
&PARSE;
print"Content-type:text/html\n\n";
if($form{search} || $form{submit}){&SEARCH_FILES;}
else{&DISPLAY_SEARCH_FIELD;}

###############################################################################
# PROGRAM SUBROUTINES BELOW
###############################################################################
sub DISPLAY_SEARCH_FIELD{
  print qq~
    document.write("<form method='post' action='$ENV{SCRIPT_NAME}' name='searchForm'>");
    document.write("<font face='Times New Roman, Times, serif' size='3'><b><font size='2'>$searchFieldText</font></b></font>");
    document.write("<input type='text' name='search' value='$form{search}' size='$formFieldSize' class='input'><input type='submit' name='submit' value='go'>");
    document.write("</form>");
  ~;
}

sub SEARCH_FILES{
  #display the header
  open(R, "$header");
  @R = <R>;
  close R;
  print @R;

  opendir(DIR, "$searchDirectory") || ERROR("ERROR: Unable to open the directory:'$searchDirectory' for searching. Please make sure that the path to the directory is correct.");
  @files = readdir(DIR);
  close DIR;

  foreach $line(@files){
    #get the files extension
    $ext=$line;
    $ext =~ s|.*(\..*)|$1|g;
    
    #save the files with a valid extension
    if($ext =~ /$searchFileTypes/){
      push(@validFiles, "$line");
    }
  }

  #replace all spaces in the search string with pipes, this is the or symbol
  #this will allow for multiple word searching instead of exact string search
  #also add \b for word boundry control. This is used for a higher accuracy. This
  #will stop partial word matching and make sure that it is only matched by whole words.
  @searchWords = split(/\s+/, $form{search});
  foreach $line(@searchWords){
    $searchString .= "\\b$line\\b\|";
  }
  #remove begining and ending pipes since we don't want to match nothing ""
  $searchString =~ s/^\|+//g;
  $searchString =~ s/\|+$//g;

  #number of files searched
  if(@validFiles){$numSearched = $#validFiles;$numSearched++;}#remember arrays start at 0 so increement one

  #search valid files
  foreach $line(@validFiles){
    open(R, "$searchDirectory$line");
    @R = <R>;
    close R;
    $timesFound=0;
      foreach $nestedLine(@R){
      	#remove html if requested
      	if($removeHtml eq 1){
      	  $nestedLine =~ s/<(.|\n)*>?//g;
      	}
        if($nestedLine =~ /$searchString/i){
          $timesFound++;
        }
      }
    
    #save all results that returned positive
    if($timesFound != 0){
      push(@fileMatches, "$timesFound\|$line");
    }
  }
  
  #sort results, set the most positive results to the top of the list, and
  #the least positive resutls to the bottom of the list.
  @fileMatches = sort{lc($a) cmp lc($b)|$a<=>$b}@fileMatches;
  @fileMatches = reverse(@fileMatches);
  
  #print search box
  print"<script>";
  &DISPLAY_SEARCH_FIELD;
  print"</script>";
  
  #if there was no results then print this message
  if(!@fileMatches){
    print qq~
      <font size=3 face='Verdana, Arial, Helvetica, SunSans-Regular'>searched $numSearched page(s)<br><b>No matches were found.</b></font><br><br>\n
    ~;
  }else{
    #if there was positive results found then print the results
    print qq~
      <font size=3 face='Verdana, Arial, Helvetica, SunSans-Regular'>searched $numSearched page(s)<br><b>Your search returned the following pages.</b></font><br><br>\n
    ~;
    foreach $line(@fileMatches){
      ($numberResults,$fileName)=split(/\|/, $line);
      print"<font size=2 face='Verdana, Arial, Helvetica, SunSans-Regular'><b>Words matched:</b>$numberResults <b>File:</b><a href='$hyperLink$fileName'><font color='$hypertextLinkColor'>$fileName</font></a></font><br>\n";
    }
  }
  
  #display the header
  open(R, "$footer");
  @R = <R>;
  close R;
  print @R;
}

#This is the error subroutine
sub ERROR{
  print"@_";
  exit;
}

sub PARSE{
  for($i=0;$i<2;$i++){
    ($i eq 0)?($f=<stdin>):($f=$ENV{QUERY_STRING});
    @f=split(/\&/, $f);
      foreach $l(@f){
        ($n,$v)=split(/\=/, $l);
        $v=~s|\+| |g;
        $v=~s|%([a-fA-F0-9][a-fA-F0-9])|pack("C", hex($1))|ge;
        $v =~ s/<!--(.|\n)*-->//g;
        $form{$n}=$v;
      }
  }
}