Home >php教程 >PHP源码 >Blankspider PHP 爬虫

Blankspider PHP 爬虫

PHP中文网
PHP中文网Original
2016-05-25 17:09:591225browse

php代码

function spider($url, $spider = 'Blankspider', $port = 80, $timeout = 15) {
	$content= '';
	$resolve = parse_url($url);
	$host = $resolve['host'];
	$path = empty($resolve['path']) ? '/' : $resolve['path'].(!empty($resolve['query']) ? '?'.$resolve['query'] : '');
	if(empty($host)) { return 'Requested host name can\'t be empty'; }
	$fp = fsockopen($host, $port, $errno, $errstr, $timeout);
	if (!$fp) {
		return $errstr;
	} else {
		$fputs = "GET $path HTTP/1.1\r\n";
		$fputs.= "Accept: */*\r\n";
		$fputs.= "Host: $host\r\n";
		$fputs.= "Spider: $spider\r\n";
		$fputs.= "Connection: Close\r\n\r\n";
		stream_set_blocking($fp, 1);
		stream_set_timeout($fp, $timeout);
		fputs($fp, $fputs);
		while(!feof($fp)) {
			if(($return = fgets($fp)) && ($return == "\r\n" || $return == "\n")) {
				break;
			}
		}
		while(!feof($fp)) { $content .= fgets($fp, 8192); }
		fclose($fp);
		return $content;
	}

}

function cron2spider($iptable, $sleep = 5) {
	set_time_limit(0);
	$i = 0;
	date_default_timezone_set('PRC');
	if(!file_exists($iptable)) {
		return json_encode(array(
		'status'=> 'error',
		'description'=> 'iptable.conf file not exists'));
	}
	$file = file($iptable);
	if(empty($file)) {
		return json_encode(array(
		'status'=> 'error',
		'description'=> 'iptable.conf can\'t be empty'));
	}
	while($i< count($file)) {
		if(!file_exists(&#39;robots&#39;) || !is_writable(&#39;robots&#39;)) {
			return json_encode(array(
			&#39;status&#39;=> &#39;error&#39;,
			&#39;description&#39;=> &#39;directory doesn\&#39;t exist or don\&#39;t have write permissions&#39;));
		}
		$dir = &#39;robots/&#39;.preg_replace(&#39;/(http\:\/\/)|(\s)|(www\.)/&#39;, &#39;&#39;, $file[$i]);
		if(!file_exists($dir)){ mkdir($dir); }
		file_put_contents($dir.&#39;/&#39;.date(&#39;Y.m.d.H.i.s&#39;, time()).&#39;.txt&#39;, spider(preg_replace(&#39;/\s/&#39;, &#39;&#39;, $file[$i])));
		$i++;
		sleep($sleep);
	}
	return json_encode(array(
	&#39;status&#39;=> &#39;ok&#39;,
	&#39;description&#39;=> &#39;robots program execution success&#39;));
}

echo cron2spider(&#39;iptable.conf&#39;);

Statement:
The content of this article is voluntarily contributed by netizens, and the copyright belongs to the original author. This site does not assume corresponding legal responsibility. If you find any content suspected of plagiarism or infringement, please contact admin@php.cn