Reduce (UTF-8) strings to alphanumeric

urlify.php

<?php

/**
 * Normalize a string to only contain alphanumeric characters and dashes.
 *
 * Replace accents by their entities. 
 * Replace everything else by - (dash).
 * @note mb_internal_charset() must be set to whatever encoding $string had originally
 * @param string $string String to normalize
 * @param boolean $trim Trim string to not start/end with a dash and not contain dash sequences
 * @param array $allow List of Characters that are to be ignored while urlifying array( unicode => true )
 * @param array $replace List of Characters that are to be replaced while urlifying array( unicode => "character" )
 * @return string normalized string
 * @author Christian Kruse <[email protected]>
 * @author Rodney Rehm <[email protected]>
 */
function urlify($string, $trim=true, $allow=null, $replace=null)
{
    if (!is_string($string)) {
        throw new Exception('$string must be a string');
    }

    $_replace = array(
        0xE4 => 'ae',
        0xC4 => 'Ae',
        0xF6 => 'oe',
        0xD6 => 'Oe',
        0xFC => 'ue',
        0xDC => 'Ue',
        0xDF => 'ss',
    );
    
    if ($replace) {
        $_replace = array_merge($_replace, $replace);
    }

    $res = '';
    $encoding = mb_internal_encoding();
    $string = mb_convert_encoding($string, "UTF-32BE");
    $unicodes = unpack("N*", $string);
    $i = -1;

    foreach ($unicodes as $code) {
        $i++;
        $character = mb_substr($string, $i, 1, "UTF-32BE");

        if (($code >= 97 && $code <= 122) || ($code >= 65 && $code <= 90) || ($code >= 48 && $code <= 57) || $code == 95 || $code == 45) {
            // skip normalization for alphanumeric characters [a-zA-Z0-9_-]
            $res .= mb_convert_encoding($character, $encoding, "UTF-32BE");
        } elseif ($allow && isset($allow[$code])) {
            // skip normalization for allowed characters
            $res .= mb_convert_encoding($character, $encoding, "UTF-32BE");
        } elseif (isset($_replace[$code])) {
            // replace as defined
            $res .= $_replace[$code];
        } elseif (ctype_space($character)) {
            // replace spaces by dash
            $res .= '-';
        } elseif (($code >= 0xC0 && $code <= 0xC6) || ($code >= 0xE0 && $code <= 0xE6) || ($code >= 0x100 && $code <= 0x105)) {
            $res .= 'a';
        } elseif ($code == 0xC7 || $code == 0xE7 || ($code >= 0x106 && $code <= 0x10D)) {
            $res .= 'a';
        } elseif ($code == 0xD0 || ($code >= 0x10E && $code <= 0x111)) {
            $res .= 'd';
        } elseif (($code >= 0xC8 && $code <= 0xCB) || ($code >= 0xE8 && $code <= 0xEB) || ($code >= 0x112 && $code <= 0x11B)) {
            $res .= 'e';
        } elseif (($code >= 0xCC && $code <= 0xCF) || ($code >= 0xEC && $code <= 0xEF)) {
            $res .= 'i';
        } elseif ($code == 0xD1 || $code == 0xF1) {
            $res .= 'n';
        } elseif (($code >= 0xD2 && $code <= 0xD8) || ($code >= 0xF2 && $code <= 0xF8)) {
            $res .= 'o';
        } elseif (($code >= 0xD9 && $code <= 0xDB) || ($code >= 0xF9 && $code <= 0xFB)) {
            $res .= 'u';
        } elseif ($code == 0xDD || $code == 0xFD || $code == 0xFF) {
            $res .= 'y';
        } else {
            $res .= '-';
        }
    }
    
    if (!$trim) {
        return $res;
    }
    
    $res = preg_replace('#-{2,}#', '-', $res);
    return trim($res, '-');
}

urlify.test.php

<?php

include dirname(__FILE__) . '/urlify.php';

mb_internal_encoding('UTF-8');

$test = array(
    'hällö wörld' => 'haelloe-woerld',
    'hällö wörldß' => 'haelloe-woerldss',
    'hällö wörldß' => 'haelloe-woerldss',
    'hällö wörld ' => 'haelloe-woerld',
    'hällö wörld %' => 'haelloe-woerld',
    'héllò peôplë ÑO?' => 'hello-people-nO',
);

echo '<pre>';
foreach ($test as $o => $r){
    $_r = urlify($o);
    echo $o .' - '. $_r .' --- '. ($r == $_r ? 'OK' : 'FAILED') .' --- '. bin2hex($_r) ."\n";
}
echo '</pre>';