Spam class refactoring
by johnk Saturday, Oct. 21, 2006 at 1:21 AM

This isn't a full refactor...

This is a slightly refactored version of spam_class.inc, called spamb_class.inc. It just splits the Detect() code into individual methods. That way, it can be used to check spam on other parts of the site (or I guess in general).
<pre>
include_once("shared/global.cfg");

/**
* This is a lot like Spam, except it'll operate on arbitrary strings, not just articles.
* The different spam detection methods are in separate functions.
*/
class SpamB
{

function SpamB ()
{
// Nothing
}

function detect_strings( &$data )
{
$stringstext = file_get_contents(SF_CACHE_PATH.'/spam_strings.txt');
$strings = explode("n",$stringstext);
foreach ( $strings as $string )
{
$string = rtrim( $string );
if (! $string)
continue;
$lines = explode( "n", $data );
foreach ( $lines as $line )
{
if ( preg_match( "/$string/i", $line ) )
{
return 1;
}
}
}
return 0;
}
function detect_ip( $ip )
{
$lines = file(SF_CACHE_PATH.'/next_ip_to_block.txt' ,'r');
$user_ip=trim($_SERVER['REMOTE_ADDR']);
for ( $i=0 ; $i < count($lines) ; $i++)
{
$saved_ip = rtrim(trim($lines[$i]));

if ( preg_match( "/^$saved_ip$/", $user_ip ) )
{
$this->Log( $_SERVER['REMOTE_ADDR']."|".
date("m-d-y g:ia")."|".
$_SERVER['HTTP_USER_AGENT']."|".
$_SERVER['HTTP_REFERER']."|".
$_SERVER['REQUEST_URI'] );
return 1;
}
}
return 0;
}
/** not tested */
function remember_content( &$data )
{
if ( $GLOBALS['spam_filter_time'] < (time()-filectime(SF_CACHE_PATH.'/hashes_time')) ) {
unlink(SF_CACHE_PATH.'/hashes.txt');
touch(SF_CACHE_PATH.'/hashes.txt');
touch(SF_CACHE_PATH.'/hashes_time');
unlink(SF_CACHE_PATH.'/hashes_content.txt');
touch(SF_CACHE_PATH.'/hashes_content.txt');
}
$hashes = fopen(SF_CACHE_PATH.'/hashes.txt','a');
fputs( $hashes, md5($data)."n" );
fclose ($hashes);
}
/**
* @return the number of matches
* not tested
*/
function detect_repeated_content( &$data )
{
$hashes = file(SF_CACHE_PATH.'/hashes_content.txt','r');
$dataHash = md5($data);
$matched_hashes = 0;
foreach( $hashes as $hash )
{
if ( $hash == $dataHash )
$matched_hashes++;
}
return $matched_hashes;
}

function Log ($text) {

$log = fopen(SF_CACHE_PATH."/ipblock.log","a");
fwrite($log, date("m-d-y g:ia").": ".$text."n");
fclose($log);
}

function too_many_urls( $str ) {
$sansUrls = preg_replace( '/http:[a-zA-Z0-9.\/]+?/', '', $str );
$sansHref = preg_replace( '/href/', '', $sansUrls );
$sansUrl = preg_replace( '/url/', '', $sansHref );
$sansWww = preg_replace( '/www[a-zA-Z0-9.\/]+?/', '', $sansUrl );
if ((sizeof($sansWww)/sizeof($str)) > 0.5)
return 1;
else
return 0;
}
}
</pre>

add your comments

Ooops - here it is with proper indenting
by johnk Saturday, Oct. 21, 2006 at 1:21 AM

include_once("shared/global.cfg");

/**
 * This is a lot like Spam, except it'll operate on arbitrary strings, not just articles.
 * The different spam detection methods are in separate functions.
 */
class SpamB
{

    function SpamB ()
    {
        // Nothing
    }

    function detect_strings( &$data )
    {
        $stringstext = file_get_contents(SF_CACHE_PATH.'/spam_strings.txt');
        $strings = explode("\n",$stringstext);
        foreach ( $strings as $string )
        {
            $string = rtrim( $string );
            if (! $string)
                continue;
            $lines = explode( "\n", $data );
            foreach ( $lines as $line )
            {
                if ( preg_match( "/$string/i", $line ) )
                {
                    return 1;
                }
            }
        }
        return 0;
    }
    function detect_ip( $ip )
    {
        $lines = file(SF_CACHE_PATH.'/next_ip_to_block.txt' ,'r');
        $user_ip=trim($_SERVER['REMOTE_ADDR']);
        for ( $i=0 ; $i < count($lines) ; $i++)
        {
            $saved_ip = rtrim(trim($lines[$i]));

            if ( preg_match( "/^$saved_ip\$/", $user_ip ) )
            {
                $this->Log( $_SERVER['REMOTE_ADDR']."|".
                    date("m-d-y g:ia")."|".
                    $_SERVER['HTTP_USER_AGENT']."|".
                    $_SERVER['HTTP_REFERER']."|".
                    $_SERVER['REQUEST_URI'] );
                return 1;
            }
        }
        return 0;
    }
    /** not tested */
    function remember_content( &$data )
    {
        if ( $GLOBALS['spam_filter_time'] < (time()-filectime(SF_CACHE_PATH.'/hashes_time')) ) {
            unlink(SF_CACHE_PATH.'/hashes.txt');
            touch(SF_CACHE_PATH.'/hashes.txt');
            touch(SF_CACHE_PATH.'/hashes_time');
            unlink(SF_CACHE_PATH.'/hashes_content.txt');
            touch(SF_CACHE_PATH.'/hashes_content.txt');
        }
        $hashes = fopen(SF_CACHE_PATH.'/hashes.txt','a');
        fputs( $hashes, md5($data)."\n" );
        fclose ($hashes);
    }
    /**
     * @return the number of matches
     * not tested
     */
    function detect_repeated_content( &$data )
    {
        $hashes = file(SF_CACHE_PATH.'/hashes_content.txt','r');
        $dataHash = md5($data);
        $matched_hashes = 0;
        foreach( $hashes as $hash )
        {
            if ( $hash == $dataHash )
                $matched_hashes++;
        }
        return $matched_hashes;
    }

    function Log ($text) {

        $log = fopen(SF_CACHE_PATH."/ipblock.log","a");
        fwrite($log, date("m-d-y g:ia").": ".$text."\n");
        fclose($log);
    }

    function too_many_urls( $str ) {
        $sansUrls = preg_replace( '/http:[a-zA-Z0-9.\\/]+?/', '', $str );
        $sansHref = preg_replace( '/href/', '', $sansUrls );
        $sansUrl = preg_replace( '/url/', '', $sansHref );
        $sansWww = preg_replace( '/www[a-zA-Z0-9.\\/]+?/', '', $sansUrl );
        if ((sizeof($sansWww)/sizeof($str)) > 0.5)
            return 1;
        else
            return 0;
    }
}

add your comments

Detect strings
by johnk Saturday, Oct. 21, 2006 at 1:24 AM

I forgot to add that there's a content spam detection method in there. It checks against a file of spam keywords.

I think the code to check for repeeated content and IPs isn't quite right. I never used spamb to replace spam_class.inc. It was used for checking calendar spam.

add your comments

New version of too_many_urls
by johnk Saturday, Oct. 21, 2006 at 1:45 AM

I messed it up. Here's a good one:

    function too_many_urls( $str ) {
        $sansUrls = preg_replace( '/http:[a-zA-Z0-9.\\/]+/', '', $str );
        $sansHref = preg_replace( '/href/', '', $sansUrls );
        $sansUrl = preg_replace( '/url/', '', $sansHref );
        $sansWww = preg_replace( '/www[a-zA-Z0-9.\\/]+/', '', $sansUrl );
        if ((strlen($sansWww)/strlen($str)) < 0.5)
            return 1;
        else
            return 0;
    }

add your comments