Detect real URL behind an audio stream.

detect_real_audio_stream_url.php

<?php 

/**
 * The idea behind this is that I had to check several hundred streams
 * and validate their URLs if they contain valid stream or not. 
 * If they didn't - fix it.
 * 
 * It is best to automate such things, so the code below does that.
 * It's basic - detects mostly shoutcast/icecast streams and .as[fx], .m3u8? playlists
 */


function get_valid_stream_url($streamURL) {
    $ctx = stream_context_create(array('http'=>
        array(
            'timeout' => 10,  //1200 Seconds is 20 Minutes
        )
    ));
    
    // detect m3u
    if (preg_match('!\.m3u8?!', $streamURL)) {
        $url_streams = parsem3u(file_get_contents($streamURL,false, $ctx));
    }
    // detect asx
    elseif (preg_match('!\.as[fx]$!', $streamURL)) {
        $url_streams = parseasx(file_get_contents($streamURL,false, $ctx));
    }    
    else {
        $url_streams = array($streamURL);
    }
    
    $url_streams = array_unique($url_streams);
    
    foreach($url_streams as $stream) {
        $parts = parse_url($stream);
        $result = detectStream($parts['host'], $parts['port'], $parts['path']);        
    
        if (!empty($result)) {
            switch ($result) {
                case 'shoutcast':
                    $stream = preg_replace('!;.*$!uis', '', $stream);
                    return $stream.';1';
                    
                case 'icecast':
                case 'other':
                    return $stream;
                    
            }
        }
        else {
            continue;
        }
    }


    return reset($url_streams);
}

function detectStream($host, $port = 80, $path) {
    
    if (!$port) {
        $port = 80;
    }
    
    /* Get the IP address for the target host. */
    $address = gethostbyname($host);
    
    $socket = socket_create(AF_INET, SOCK_STREAM, SOL_TCP);
    socket_set_option($socket, SOL_SOCKET, SO_RCVTIMEO, array("sec" => 5, "usec" => 0));

    
    if ($socket === false) {
        return '';
    }
    
    $socket = fsockopen($host, $port, $errno, $err, 5);
    
    if ($socket === false) {
        return false;
    }
    
    $in = "GET ".$path." HTTP/1.1\r\n";
    $in .= "Connection: Close\r\n\r\n";
    $out = '';
    
    fwrite($socket, $in, strlen($in));
    
    
    // get headers 
    while (!feof($socket)){ 
        $line = fgets($socket, 4096); 
        if('' == trim($line)){
            break;
        }
        $theaders .= $line;
    }
    
    $headers = array();

    $theaders = preg_split('![\r\n]+!', $theaders);
    foreach ($theaders as $header){
        $t = explode(':', $header); 
        if (isset($t[0]) && trim($t[0]) != ''){
            $name = preg_replace('/[^a-z][^-a-z0-9]*/i','', strtolower(trim($t[0])));
            array_shift($t);
            $value = trim(implode(':', $t));
            if ($value != ''){
                if (is_numeric($value)){
                    $headers[$name] = (int)$value;
                }else{
                    $headers[$name] = $value;
                }
            }
        }
    }

    if (isset($headers['server']) && preg_match('!(icecast|MediaGateway)!uis',$headers['server'])) {
        return 'icecast';
    }

    // detect icecast
    foreach(array_keys($headers) as $header_name) {
        if (preg_match('!^icy!uis',$header_name)) {
            return 'shoutcast';
        }   
    }

    return 'other';

}


function parsem3u($file_content) {
    // should contain only urls
    if (!preg_match_all('!^(http://[^\n]+)!uim',$file_content, $matches)) {
        return array();
    }
    
    if (empty($matches[1])) {
        return array();
    }
    
    return array_unique($matches[1]);
}

function parseasx($file_content) {
    
    $file_content = trim($file_content);
    
    if (empty($file_content)) {
        return array();
    }
    
    $file_content = mb_strtolower($file_content);
    
    $file_content = preg_replace('@&(?!amp;)@uis', '&amp;', $file_content);

    $possible_urls = array();
    
    $doc = new DOMDocument();
    $doc->loadXML($file_content);
    
    
    $xpath = new DOMXpath($doc);

    $elements = $xpath->query("//ref");

    if (!is_null($elements)) {
      foreach ($elements as $element) {
          
        $url = $element->getAttribute('href');
        
        if (!empty($url)) {
            $possible_urls[] = $url;
        }    
      }
    }
    
    if (empty($possible_urls)) {
        
        echo $doc->saveXML();
        
        var_dump('Nothing found here :( ');
        echo $file_content;
        return array();
    }
    
    return $possible_urls;
}