URL Spam Detector Function
by johnk Saturday, Oct. 21, 2006 at 1:08 AM

This is a function that returns true if URLs make up a big portion of the post.

Add this to the spam_class.inc

    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;
    }

        if ($this->too_many_urls( $_POST['article'] ))
            $ret = 1;

add your comments

fix
by johnk Saturday, Oct. 21, 2006 at 1:44 AM

Sorry, I messed it up. Here's one that works:

    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