#!/usr/bin/perl


#!c:/programme/perl/bin/perl

#use File::stat;


## Variablendeklaration ##


# Angabe des Basisverzeichnis (genaue Angaben beim Admin erfragen)
$basedir = $ENV{DOCUMENT_ROOT}."/imv-center/prueftest/";

# Zum Basisverzeichnis korrespondierender URL
$baseurl = 'http://www.pruefziffer.de/';


# Durchsuchbare Inhalte/ Dateien
@files = ('*.html');


# URL und Titel der Website
$link_url = 'http://www.pruefziffer.de';
$link_title = 'PrüfzifferValidate - Bankverbindungen schnell und einfach prüfen';

# Verzeichnis, in dem das Skript abgelegt ist
$cgi_url = '/cgi-bin/search.cgi';


## Ende Variablendeklaration ##



# Optionen für die Gewichtung vom Dateiname, Titel, 
# Meta-Schlüsselwörter Meta-Beschreibung

($name_x, $title_x, $keywords_x, $description_x) = (1,2,4,2);




#########################################################
##  Nachfolgender Code muss nur dann bearbeitet werden, #
##  wenn das Ausgabedokumente anpepasst werden soll.    #
#########################################################

read(STDIN,$buffer,$ENV{'CONTENT_LENGTH'});
@pairs = split(/&/,$buffer);
foreach $pair (@pairs)
	{
	($name,$value) = split(/=/,$pair);
	$value =~ tr/+/ /;
	$value =~ s/%([a-fA-F0-9][a-fA-F0-9])/pack("C", hex($1))/eg;
	$FORM{$name} = $value;
	}


# Umlaute
$FORM {'queryphrase'} =~ s/ä/\&auml\;/g; 
$FORM {'queryphrase'} =~ s/ö/\&ouml\;/g; 
$FORM {'queryphrase'} =~ s/ü/\&uuml\;/g; 
$FORM {'queryphrase'} =~ s/Ä/\&Auml\;/g; 
$FORM {'queryphrase'} =~ s/Ö/\&Ouml\;/g; 
$FORM {'queryphrase'} =~ s/Ü/\&Uuml\;/g; 
$FORM {'queryphrase'} =~ s/ß/\&szlig\;/g; 

@sp = split (/\+/, $FORM{'queryphrase'});
	






if ($FORM{'queryphrase'})
	{
	&get_files;
	&search;
	&return_html;
	}
else
	{
	&prompt;
	}


sub prompt
{
	
&output_top;

print <<EOM;

<H1>Suche</H1>
<FORM METHOD=POST ACTION="$cgi_url" NAME="searchform">
<P>Bitte geben Sie einen Suchbegriff oder eine Suchphrase ein: <INPUT TYPE=TEXT NAME="queryphrase" SIZE=30>
<INPUT TYPE=SUBMIT VALUE="Suchen"><BR>

EOM

&output_bottom;

}






sub get_files
{
&bad_base unless (-e $basedir);
chdir($basedir);
foreach $file (@files)
 {
#	$ls = `ls $file`;
#   @ls = split(/\s+/,$ls);
 @ls = glob $file;
      foreach $temp_file (@ls) {
         if (-d $file) {
            $filename = "$file$temp_file";
            if (-T $filename) {
               push(@FILES,$filename);
            }
         }
         elsif (-T $temp_file) {
            push(@FILES,$temp_file);
         }
      }
   }
}


sub search
{
# Konvertiert mehrere Leerzeichen in eines:
$FORM{'queryphrase'} =~ s/\s+/ /g;
$FORM{'queryphrase'} = " $FORM{'queryphrase'} ";

# Konvertiert NOT-Statements in  Minuszeichen:
$FORM{'queryphrase'} =~ s/ not / -/ig;

# Konvertiert AND-Statements in Pluszeichen:
$FORM{'queryphrase'} =~ s/ and / \+/ig;

# OR-statements:
$FORM{'queryphrase'} =~ s/ or / /ig;

# Platzhalter:
$check = 'true' unless ($FORM{'queryphrase'} =~ /\*/);

# Gruppierte Einträge:
@terms = split(/\"/,$FORM{'queryphrase'});
$iterator = 0;
$FORM{'queryphrase'} = "";
$placeholder = '%%%==%%%';
foreach $term (@terms)
	{
	if ($iterator == 1)
		{$iterator--;}
	else
		{$iterator++;}

	$term =~ s/ /$placeholder/g unless $iterator;
	$FORM{'queryphrase'} .= $term;
	}


@terms = split(/\s+/,$FORM{'queryphrase'});
foreach $term (@terms)
	{
	next if ($term eq '');

	$term =~ s/$placeholder/ /g;

	if ($term =~ /^\+/)
		{
		$term =~ s/\+//o;
		$term = '\W' . $term. '\W' unless ($term =~ /^\$/);
		$term =~ s/^\$//;
		push(@required,$term);
		$required_terms_present = "you bet";
		}
	elsif ($term =~ /^-/)
		{
		$term =~ s/-//o;
		$term = '\W' . $term. '\W' unless ($term =~ /^\$/);
		$term =~ s/^\$//;
		push(@forbidden,$term);
		$forbidden_terms_present = "fraid so";
		}
	else
		{
		$term = '\W' . $term. '\W' unless ($term =~ /^\$/);
		$term =~ s/^\$//;
		push(@optional,$term);
		}
	}


foreach $FILE (@FILES)
	{
open(FILE,"$FILE");
@LINES = <FILE>;
close(FILE);
$string = join(' ',@LINES);
$string =~ s/\n//g;


# Extrahiert Titel, sofern vorhanden:
if ($string =~ /<title>(.*)<\/title>/i)
	{
	$titles{$FILE} = $1;
	for ($i=1;$i<$title_x;$i++)
		{
		$string .= $titles{$FILE};
		}
	}
$titles{$FILE} = $FILE unless $titles{$FILE};

# Extrahiert Beschreibung, sofern vorhanden:
if ($string =~ /<meta\s+name="description"\s+content="(.*)>/i)
	{
	@cut = split(/\">/,$1);
	$description{$FILE} = $cut[0];
	for ($i=0;$i<$description_x;$i++)
		{
		$string .= $description{$FILE};
		}
	}
else
	{
	$string2 = $string;
	$string2 =~ s/<title>(.*)<\/title>//ig;
	$string2 =~ s/<([^>]|\n)*>//g;
	@words = split(/\s+/,$string2);
	for ($i=0;$i<25;$i++)
		{$description{$FILE} .= "$words[$i] ";}
	$description{$FILE} .= "...";
	}


# Extrahiert Keywords, sofern vorhanden:
if ($string =~ /<meta\s+name="keywords"\s+content="(.*)>/i)
	{
	@cut = split(/\">/,$1);
	$keywords = $cut[0];
	for ($i=0;$i<$keywords_x;$i++)
		{
		$string .= $keywords;
		}
	}


# Gewichtet den Dateinamen:
for ($i=0;$i<$name_x;$i++)
	{
	$string .= "$baseurl$FILE";
	}


# 
$string =~ s/<([^>]|\n)*>//g;


# Überprüft optionale Ausdrücke:
foreach $term (@optional)
	{
	$lowercase = $term;
	$lowercase =~ tr/[A-Z]/[a-z]/;
	$lowercase =~ tr/\\w/\\W/;

	if (($term eq $lowercase) && ($check))
		{
		$include{$FILE} = 'yes' if ($string =~ /$term/i);
		@count = split(/$term/i,$string);
		}
	elsif ($check)
		{
		$include{$FILE} = 'yes' if ($string =~ /$term/);
		@count = split(/$term/,$string);
		}
	$word_count = @count;
	$relevance{$FILE} = $relevance{$FILE} + $word_count;
	}



# Prüft erforderliche Begriffe:
if ($required_terms_present eq "you bet")
{
foreach $term (@required)
	{
	$lowercase = $term;
	$lowercase =~ tr/[A-Z]/[a-z]/;
	$lowercase =~ tr/\\w/\\W/;
	if (($term eq $lowercase) && ($check))
		{
		if (($string =~ /$term/i) && ($include{$FILE} ne 'no'))
			{
			$include{$FILE} = 'yes';
			@count = split(/$term/i,$string);
			}
		else
			{
			$include{$FILE} = 'no';
			last;
			}
		}
	elsif ($check)
		{
		if (($string =~ /$term/) && ($include{$FILE} ne 'no'))
			{
			$include{$FILE} = 'yes';
			@count = split(/$term/,$string);
			}
		else
			{
			$include{$FILE} = 'no';
			last;
			}
		}
	$word_count = @count;
	$relevance{$FILE} = $relevance{$FILE} + $word_count;
	}
}



# Prüft verbotene Begriffe:
if (($forbidden_terms_present eq "fraid so") && ($check))
{
foreach $term (@forbidden)
	{
	$lowercase = $term;
	$lowercase =~ tr/[A-Z]/[a-z]/;
	$lowercase =~ tr/\\w/\\W/;
	if (($term eq $lowercase) && ($string =~ /$term/i))
		{
		$include{$FILE} = 'no';
		last;
		}
	elsif ($string =~ /$term/)
		{
		$include{$FILE} = 'no';
		last;
		}
	}
}



# Platzhalter
$include{$FILE} = 'yes' unless ($check);


# Format für die Relevanz der Suchergebnisse:
if ($include{$FILE} eq 'yes')
	{
	$relevance = sprintf("%.3f",($relevance{$FILE}/1000));
	$HITS{"$relevance$FILE"} = "$FILE";
	$hitcount++;
	}


} # Schleifenende - alle Dateien sind durch.
} # Suchprozedur abgeschlossen.



####  Ergebnisausgabe ###

sub return_html
{
	
# Erzeugt eine Zusammenfassung für den Suchenden: 

$docstring = "$hitcount Dokumente";
$docstring = "Ein Dokument" if ($hitcount == 1);
$docstring = "Keine Dokumente" unless ($hitcount);

$summary = "<H3> $docstring gefunden</H3> \n";
$summary .= "<BLOCKQUOTE>\n<PRE>\n";
$summary .= "Optionale Begriffe:  " if (@optional);
$i = 0;
foreach $term (@optional)
	{
	$term = "<I>$term</I>" unless ($term =~ /^\\W/);
	$term =~ s/\\W//g;
	$summary .= "$term";
	$i++;
	$summary .= ", " unless ($i == @optional);
	}
$summary .= "\n     Erzungene Begriffe:   " if (@required);
$i = 0;
foreach $term (@required)
	{
	$term = "<I>$term</I>" unless ($term =~ /^\\W/);
	$term =~ s/\\W//g;
	$summary .= "$term";
	$i++;
	$summary .= ", " unless ($i == @required);
	}
$summary .= "\n    Ausgeschlossene Begriffe:  " if (@forbidden);
$i = 0;
foreach $term (@forbidden)
	{
	$term = "<I>$term</I>" unless ($term =~ /^\\W/);
	$term =~ s/\\W//g;
	$summary .= "$term";
	$i++;
	$summary .= ", " unless ($i == @forbidden);
	}
$summary .= "\n</PRE></BLOCKQUOTE>\n";

open(SUMMARY,">>$summary_file");
print SUMMARY "Suche nach $ENV{'REMOTE_HOST'}:<BR>\n";
print SUMMARY $summary;
close(SUMMARY);



# Es folgt die Ausgabe der Suchergebnisse: 



&output_top;

print <<EOM;
<h1>Ergebnis der Suche</h1>
<P><font face="Verdana, Arial, Helvetica, sans-serif">
    $summary
<DL>
EOM

if ($hitcount > 0)
{
foreach $key (reverse sort keys %HITS)
	{
	$file = $HITS{$key};
	$size = -s "$basedir$file";
	if ($size > 1500)
		{$size = int($size/1000) . " K";}
	else
		{$size = "$size bytes";}
	$last = &Last_Modified("$basedir$file");
	print "<P><DT><a href=\"$baseurl$file\"><STRONG>$titles{$file}</STRONG></a></DT>\n";
	print "<DD>$description{$file}<BR>\n";
	print "<CITE><A HREF=\"$baseurl$file\">$baseurl$file</A>";
#	print "<FONT SIZE=-1> - $size - $last</FONT>";
	print "</CITE></DD></P>\n";
	}
}

&output_bottom;
}


sub Last_Modified
{

$filename = shift;
#$sb = stat($filename);

($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) = localtime((stat($filename))[9]);
#($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) = scalar localtime($sb->mtime);
@months = ("Januar","Februar","M&auuml;rz","April","Mai","Juni","Juli","August","September","Oktober","November","Dezember"); 
$year += 1900; 
return  "$mday. $months[$mon] $year"; 

} 


sub output_top
{
print "Content-type: text/html\n\n";
print <<EOM;
<HTML>
<HEAD>
<STYLE type="text/css">
  A:link,A:visited,A:active{color:#000099;text-decoration:none;font-weight:bold;font-family:Arial,Helvetica,Verdana,sans serif}
  A:hover{color:#990000;text-decoration:underline;font-weight:bold;font-family:Arial,Helvetica,Verdana,sans serif}
  p,td,li{color:#000000;font-family:Arial,Helvetica,Verdana,sans serif;font-size:10pt}
  h1{color:#0000AA;font-family:Arial,Helvetica,Verdana,sans serif;font-size:16pt}
  h2{color:#0000AA;font-family:Arial,Helvetica,Verdana,sans serif;font-size:14pt}
  h3{color:#0000AA;font-family:Arial,Helvetica,Verdana,sans serif;font-size:12pt}
  .field{color:#08386D;font-size:8pt;font-family:Arial,Helvetica,Verdana,sans serif;background-color:#F3F5F8;padding:0pt;top:-5pt;}
  body{margin-left:0px;margin-right:0px;margin-top:0px;margin-bottom:0px}
</STYLE>
<TITLE>Suche auf den Seiten von pruefziffer.de</TITLE>
<META NAME="author" content="Semjon Krolzik">
<META NAME="publisher" content="imMEDIAtely Medienversand Gebr. Krolzik oHG">
<META NAME="copyright" content="(c) 2001 imMEDIAtely Medienversand, Germany">
<META NAME="keywords"
content="Prüfziffer, Pruefziffer, Pruefung, Bank, Bankverbindung, BLZ, Prüfung, 
Bankkonto, Konto, Girokonto, Kreditkarte, Kreditkartennummer, Check, ISBN, EAN, 
EAN13, Kontonummer, Kontoprüfung, Kreditkartenprüfung, Validierung, Berechnung, 
Prüfzifferberechnung, Prüfverfahren, Plausibilität, Sicherheit, ecommerce,
e-commerce, download, downloaden, sicheres 
online-shopping, online-shopping, shop, VISA, Eurocard, Mastercard, Amex, 
American Express, Lastschrift, elv, Bundesbank, Bankleitzahl, PrüfzifferValidate, 
Java, Perl, ActiveX, Software, Internet, Krolzik, imMEDIAtely, Medienversand, 
iMV, Kto-Nr">
<META NAME="description" content="Suche auf den Seiten von pruefziffer.de">
<META NAME="page-topic" content="Wirtschaft">
<meta name="Content-language" content="DE">
<META NAME="page-type" content="Produktinfo">
<META NAME="audience" content="Alle">
<META NAME="robots" content="NOINDEX,NOFOLLOW">
<SCRIPT language="javascript">
<!-- 
function nullstate(){status='';return(true);} 
//-->
</SCRIPT>
</HEAD>
<BODY background="/images/back.gif" bgcolor="white" topmargin="0" leftmargin="0" bottommargin="0" marginwidth="0" marginheight="0">
<table width="100%" height="100%" cellspacing="0" cellpadding="0" border="0">
  <tr height="56" bgcolor="#0BAEFE">
    <td width="166"><A href="/index.html"><IMG src="/images/lefthead.gif" alt="Homepage pruefziffer.de" border="0"></A></td>
    <td colspan="4" align="center" width="1600"><IMG src="/images/header.gif" alt="Pr&uuml;fzifferValidate"></td>
  </tr>
  <tr height="10">
    <td valign="top"><IMG src="/images/left1.gif" width="166" height="20"></td>
    <td rowspan="4"><IMG src="/images/spacer.gif" width="15"></td>
    <td rowspan="2" width="95%" valign="top">
 	  <FORM method="POST" action="/cgi-bin/search.cgi"><INPUT type="image" src="/images/s_suche.gif" border="0" width="43" height="20" alt="Suche" align="top">&nbsp;<INPUT type="text" name="queryphrase" size="10" class="field">&nbsp;&nbsp;
	  <A href="/news.html"><IMG src="/images/s_news.gif" border="0" width="38" height="20" alt="News" align="top"></A>&nbsp;&nbsp;
	  <A href="/sitemap.html"><IMG src="/images/s_sitemap.gif" border="0" width="59" height="22" alt="Site Map" align="top"></A>&nbsp;&nbsp;
	  <A href="/kontakt.html"><IMG src="/images/s_kontakt.gif" border="0" width="53" height="20" alt="Kontakt" align="top"></A>&nbsp;&nbsp;
	  <A href="/impressum.html"><IMG src="/images/s_impressum.gif" border="0" width="73" height="22" alt="Impressum" align="top"></A>&nbsp;&nbsp;</FORM>
	</td>
    <td rowspan="4"><IMG src="/images/spacer.gif" width="15"></td>
    <td rowspan="4">&nbsp;</td>
  </tr>
  <tr height="100">
    <td rowspan="3" valign="top">
	  <A href="/produkt.html"><IMG src="/images/produkt.gif" width="166" height="34" alt="Produktbeschreibung" border="0"></A><BR>
	  <A href="/funktionen.html"><IMG src="/images/funktion.gif" width="166" height="34" alt="Funktionen testen" border="0"></A><BR>
	  <A href="/download.html"><IMG src="/images/downloads.gif" width="166" height="34" alt="Downloads" border="0"></A><BR>
	  <A href="/preise.html"><IMG src="/images/preise.gif" width="166" height="34" alt="Preise" border="0"></A><BR>
	  <A href="/bestellung.html"><IMG src="/images/bestellung.gif" width="166" height="34" alt="Pr&uuml;fzifferValidate bestellen" border="0"></A><BR>
<!--	  <A href="/referenzkunden.html"><IMG src="/images/referenzkunden.gif" width="166" height="34" alt="Referenzkunden" border="0"></A><BR>-->
	  <A href="/faq.html"><IMG src="/images/faq.gif" width="166" height="34" alt="Frequently Asked Questions" border="0"></A><BR>
	</td>
  </tr>
  <tr height="5">
    <td><IMG src="/images/spacer.gif" width="1" height="5"></td>
  </tr>
  <tr>
    <td valign="top">
	  <TABLE border="0" width="500">
	    <TR>
		  <TD>
EOM
}


sub output_bottom
{
	
print <<EOM;
</DL>
<BR>
    </font></P>
    
		  </TD>
		</TR>
	  </TABLE>
    </td>
  </tr>
  <tr>
    <td colspan="5" valign="bottom" align="left">&nbsp;<BR>
      <IMG src="/images/footer.gif" width="700" height="16" border="0" style="vertical-align:bottom" usemap="#creator" alt="&copy; 1999-2001 imMEDIAtely Medienversand, Germany">
      <map name="creator">
        <area shape=rect coords="659,2,676,14" href="/info.html" onmousemove="return nullstate();" onmouseover="return nullstate();" onmouseout="return nullstate();">
      </map>
    </td>
  </tr>
</table>
</BODY>
</HTML>
EOM
	
}