<?php

/***  DOCUMENTATION LAYER

Last Google Query Validation Library

Last Update: Jul 2007
Author: Tom at klenwell@gmail.com

FUNCTIONS
    lg_clean_query($query)
    lg_is_spam($string)
    lg_on_spam_list($query)
    lg_spam_preg_match($query)
    lg_spam_redirect()

NOTES

______________________________________________________________________________*/

// GLOBALS (LGSS: Last Google Site Security)
$LGSS['spam_trigger'] = '';
$LGSS['default_redirect'] = 'http://klenwell.net/is/SpamTrap';

$LGSS['TERM'] = array
    (
        'hydrocodone', 'phentermine',       // spam terms
        '[url'      // forum spam
    );

$LGSS['regex'] = '%'        // begin regex
    . '(lo\d+)'         // e.g. lo1239 -- something to do with loopbacks?
    . '|(<\s*br[^\>]*>)'        // don't need line breaks
    . '|(viagra)'           // common spam terms
    . '%iU';            // end regex

// lg_clean_query
/*____________________________________________________________________________*/
function lg_clean_query($query)
{
// *** DATA

    // return
    $clean_query = '';

// *** MANIPULATE

    // trim
    $query = trim($query);

    // chop at newline
    $nl_pos = strpos($query, "\n");
    if ( $nl_pos > 0 ) $query = substr($query, 0, $nl_pos);
    $clean_query = strtolower($query);
   
    // clean up extra spaces
    $clean_query = preg_replace('%\s+%', ' ', $clean_query);
    $clean_query = trim($clean_query);


// *** RETURN

    return $clean_query;
}
/*____________________________________________________________________________*/



// lg_is_spam
/*____________________________________________________________________________*/
function lg_is_spam($string)
{
// *** DATA

    // Global
    global $PROJECT;

    // Return
    $is_spam = 0;


// *** MANIPULATE

    // sanity check
    if ( empty($string) ) return 0;

    // * Cases (add as needed)
    // css-hidden text field (bot bait) source: http://it.slashdot.org/comments.pl?sid=247229&cid=19797293
    // note: not actual field name
    if ( $_POST['bot_trap'] ) $is_spam = 1;

    // check spam list
    elseif ( lg_on_spam_list($string) ) $is_spam = 1;
   
    // regex check
    elseif ( lg_spam_preg_match($string) ) $is_spam = 1;
   
    // too many urls
    elseif ( substr_count($string, 'http://') > 1 ) $is_spam = 1;
   
    // endline signal bot in this case (via spoofed blogspot form)
    // lg_clean_query should fix this
    #elseif ( strpos($string,"\n") !== FALSE ) $is_spam = 1;
   
    // clean
  else $is_spam = 0;

// *** RETURN

    return $is_spam;

} // end Fx
/*____________________________________________________________________________*/


// is spam term
/*____________________________________________________________________________*/
function lg_on_spam_list($query)
{
// *** DATA

    // globals
    global $LGSS;
   
    // internal
    $_QUERY = array();

    // return
    $is_spam = 0;


// *** MANIPULATE

    // split query
    $query = trim($query);
    $_QUERY = explode(' ', $query);
   
    // cycle
    foreach ( $_QUERY as $_term )
    {
        if ( in_array($_term, $LGSS['TERM']) )
        {
            #trigger_notice(__FUNCTION__ . ' ' . $_term);
            $is_spam = 1;
            $LGSS['spam_trigger'] = $_term;
            break;
        }
    }

// *** RETURN

    return $is_spam;
}
/*____________________________________________________________________________*/


// in spam regex
/*____________________________________________________________________________*/
function lg_spam_preg_match($query)
{
// *** DATA

    // globals
    global $LGSS;

    // return
    $is_spam = 0;

// *** MANIPULATE

    if ( preg_match($LGSS['regex'], $query, $MATCH) )
    {
        $is_spam = 1;
        $LGSS['spam_trigger'] = htmlspecialchars($MATCH[0]);
        #trigger_notice($MATCH); trigger_notice(__FUNCTION__ . ' : ' . $query);
    }

// *** RETURN

    return $is_spam;
}
/*____________________________________________________________________________*/


// Testbed
/*____________________________________________________________________________*/

if ( 0 )
{
    $_TEST[] = "test site <br />";
    $_TEST[] = "test site br />";
    $_TEST[] = "lo831l";
    $_TEST[] = <<<TEST
Very fabulos site!
idiot test
idiot test
[URL=http://idiotest.9999mb.com]idiot test[/URL]
idiot test http://idiotest.9999mb.com
Thank you!
TEST;
    $_TEST[] = ' PhenTermine ';
    $_TEST[] = 'this is a legit query';
   
    // test clean query
    foreach ( $_TEST as $test )
    {
        $result = lg_clean_query($test);
        $test = htmlspecialchars($test);
        echo "<p><b>$test</b> --lg_clean_query--> $result</p>";
    }
   
    echo "<h2>begin spam test</h2>";

    foreach ( $_TEST as $test )
    {
        $test = lg_clean_query($test);
        $result = lg_is_spam($test) ? 'spam' : 'ok';
        $test = htmlspecialchars($test);
        echo "<p><b>$test</b> --lg_is_spam--> $result --LGSS['spam_trigger']--> {$LGSS['spam_trigger']}</p>";
        $LGSS['spam_trigger'] = '';
    }
    die(__FILE__);
}

/*____________________________________________________________________________*/

?>