<?php
preg_replace("/[^\p{L}0-9\+\-\_:\.@ ]/u", "", $_string);
$html  = file_get_contents('yandex.ru');
$url   = 'yandex.ru';
$vnut  = array();
$vnech = array();
preg_match_all('~<a [^<>]*href=[\'"]([^\'"]+)[\'"][^<>]*>(((?!~si', $html, $matches);
foreach ($matches[1] as $val) {
    if (!preg_match("~^[^=]+://~", $val) || preg_match("~^[^://]+://(www\.)?" . $url . "~i", $val)) {
        $vnut[] = $val;
    } else
        $vnech[] = $val;
}
$vnut  = array_unique($vnut);
$vnech = array_unique($vnech);
print_r($vnut);
print_r($vnech);
?>