PHP多线程批量采集下载美女图片的实现代码(续)

php  /  管理员 发布于 7年前   174

/**
     * curl 多线程
     * 
     * @param array $array 并行网址
     * @param int $timeout 超时时间
     * @return mix
     */
 public function Curl_http($array,$timeout='15'){
      $res = array();

      $mh = curl_multi_init();//创建多个curl语柄

      foreach($array as $k=>$url){
          $conn[$k]=curl_init($url);//初始化

          curl_setopt($conn[$k], CURLOPT_TIMEOUT, $timeout);//设置超时时间
          curl_setopt($conn[$k], CURLOPT_USERAGENT, 'Mozilla/5.0 (compatible; MSIE 5.01; Windows NT 5.0)');
          curl_setopt($conn[$k], CURLOPT_MAXREDIRS, 7);//HTTp定向级别 ，7最高
          curl_setopt($conn[$k], CURLOPT_HEADER, false);//这里不要header，加块效率
          curl_setopt($conn[$k], CURLOPT_FOLLOWLOCATION, 1); // 302 redirect
          curl_setopt($conn[$k], CURLOPT_RETURNTRANSFER,1);//要求结果为字符串且输出到屏幕上         
    curl_setopt($conn[$k], CURLOPT_HTTPGET, true);

          curl_multi_add_handle ($mh,$conn[$k]);
      }
       //防止死循环耗死cpu 这段是根据网上的写法
          do {
              $mrc = curl_multi_exec($mh,$active);//当无数据，active=true
          } while ($mrc == CURLM_CALL_MULTI_PERFORM);//当正在接受数据时
          while ($active and $mrc == CURLM_OK) {//当无数据时或请求暂停时，active=true
              if (curl_multi_select($mh) != -1) {
                  do {
                      $mrc = curl_multi_exec($mh, $active);
                  } while ($mrc == CURLM_CALL_MULTI_PERFORM);
              }
          }

      foreach ($array as $k => $url) {
            if(!curl_errno($conn[$k])){
             $data[$k]=curl_multi_getcontent($conn[$k]);//数据转换为array
             $header[$k]=curl_getinfo($conn[$k]);//返回http头信息
             curl_close($conn[$k]);//关闭语柄
             curl_multi_remove_handle($mh  , $conn[$k]);   //释放资源
            }else{
             unset($k,$url);
            }
          }

          curl_multi_close($mh);

          return $data;

   }

//参数接收
$callback = $_GET['callback'];
$hrefs = $_GET['hrefs'];
$urlarray = explode(',',trim($hrefs,','));
$date = date('Ymd',time());
//实例化
$img = new HttpImg();
$stime = $img->getMicrotime();//开始时间

$data = $img->Curl_http($urlarray,'20');//列表数据
mkdir('./img/'.$date,0777);
foreach ((array)$data as $k=>$v){
 preg_match_all("/(href|src)=(["|']?)([^ "'>]+.(jpg|png|PNG|JPG|gif))\2/i", $v, $matches[$k]);

 if(count($matches[$k][3])>0){
  $dataimg = $img->Curl_http($matches[$k][3],'20');//全部图片数据二进制
  $j = 0;
  foreach ((array)$dataimg as $kk=>$vv){
   if($vv !=''){
    $rand = rand(1000,9999);
    $basename = time()."_".$rand.".".jpg;//保存为jpg格式的文件
    $fname = './img/'.$date."/"."$basename";
    file_put_contents($fname, $vv);   
    $j++;
    echo "创建第".$j."张图片"."$fname"."
";
   }else{
    unset($kk,$vv);
   }
  }
 }else{
  unset($matches);
 }
}
$etime = $img->getMicrotime();//结束时间
echo "用时".($etime-$stime)."秒";
exit;