<?PHPheader("content-type:text/html;charset=utf-8");/**Base Function***/// exit;/*** get远程文档* @access public* @param string $url 远程url* @return mixed*/function GetUrl($url) {if (function_exists('file_get_contents')){return file_get_contents($url);}$ch = curl_init();curl_setopt($ch, CURLOPT_RETURNTRANSFER, TRUE);curl_setopt($ch, CURLOPT_URL, $url);$result =  curl_exec($ch);curl_close($ch);return $result;}/*** 采集远程文件* @access public* @param string $remote 远程文件名* @param string $local 本地保存文件名* @return mixed*/function curlDownload($remote,$local) {$cp = curl_init($remote);$fp = fopen($local,"w");curl_setopt($cp, CURLOPT_FILE, $fp);curl_setopt($cp, CURLOPT_HEADER, 0);curl_exec($cp);curl_close($cp);fclose($fp);}/**采集代码开始干活的搞起**/$BaseUrl         =     'http://www.a-li.com.cn/';$EmotionUrl     =     $BaseUrl.'download/qq_emotion/index.php?page=';//采集页面的开始与结束$BeginPage     =     1;$EndPage        =    37;//总页面数；阿狸官方站更新后 总页面数可能会变化 在此处配置//$_SESSION['page'] = null;exit;//不注释该项重新初始化//采用session记录采集到哪个页面if(!$_SESSION['page']) {$Page = $BeginPage;$_SESSION['page'] = $Page;}else {$Page = $_SESSION['page'] + 1;$_SESSION['page'] = $Page;}if($Page>$EndPage) {exit('活已干完！');}//开始采集$url = $EmotionUrl.$Page;$string = GetUrl($url);while(!$string) { //while 循环保证采集无误sleep(3);$string = GetUrl($url);}//正则匹配采集到的页面中的阿狸gif图片信息$match  =    array();$string = preg_replace('/<!--.*\s*.*-->/i','',$string);//去除html里的注释段  可以删掉preg_match_all('/<img src="(data\/attachment.*\.gif)"\s+alt="(.*)"\s+.*\s+\/>/i',$string,$match,PREG_SET_ORDER);//采集后的文件存储位置 检测文件夹不存在就创建if(!is_dir('./ali/')) {mkdir('./ali/');}//检测匹配并循环下载采集到的gif图到与该PHP代码文件同级目录下的ali目录中if($match) {foreach($match as $key=>$value) {$imgUrl = $BaseUrl.$value[1];$imgName = './ali/'.preg_replace('/\./','',$value[2]).'.gif';//使用匹配到的img标签中的alt作为本地保存gif图的文件名curlDownload($imgUrl,$imgName);}}else {exit('匹配信息出错');}echo '<p>采集第'.$Page.'页中的阿狸gif图片完成。</p>';echo '<p>1秒后继续自动执行下一个页面</p>';echo '<script>setTimeout(function () {window.location.reload(true);},1000);</script>';?>

<?php
function curlDownload($remote,$local) {$cp = curl_init($remote);$fp = fopen($local,"w");curl_setopt($cp, CURLOPT_FILE, $fp);curl_setopt($cp, CURLOPT_HEADER, 0);curl_exec($cp);curl_close($cp);fclose($fp);}curlDownload('http://blog.jjonline.cn/Upload/image/201408/20140810151040.gif','aa.gif');?>

<?phpfunction GetUrl($url) {if (function_exists('file_get_contents')){return file_get_contents($url);}$ch = curl_init();curl_setopt($ch, CURLOPT_RETURNTRANSFER, TRUE);curl_setopt($ch, CURLOPT_URL, $url);$result =  curl_exec($ch);curl_close($ch);return $result;}echo GetUrl("http://blog.jjonline.cn/phptech/175.html");
?>