PHP定时任务通过CURL图片的抓取例子

<?php

  function getLink($url){

    include_once('simple_html_dom.php');

    $ch = curl_init();

    curl_setopt($ch,CURLOPT_URL,$url);

    curl_setopt($ch,CURLOPT_HEADER,false);

    curl_setopt($ch,CURLOPT_RETURNTRANSFER,1);

    $output = curl_exec($ch);

    curl_close($ch);

    $html = new simple_html_dom();

    $html->load($output);

 $links = array();

    $arr = array();

 $title = array();

    foreach($html->find('a') as $element){

      if( preg_match('#^\/content_[0-9]+_1\.html$#i',$element->href)){

            array_push($links,'http://www.111cn.net'.$element->href);

   array_push($title,$element->title);

  }

  

 } 

 $links = array_values(array_unique($links));

 $title = array_values(array_unique($title));

 $arr['links'] = $links;

 $arr['title'] = $title;

 return $arr;

  }

  

  function loadimg($url,$dirname){ 

  include_once('simple_html_dom.php');

  $ch = curl_init();

  curl_setopt($ch,CURLOPT_URL,$url);

  curl_setopt($ch,CURLOPT_HEADER,false);

  curl_setopt($ch,CURLOPT_RETURNTRANSFER,1);

  $output = curl_exec($ch);

  curl_close($ch);

  $html = new simple_html_dom();

  $html->load($output);

  $arr = array();

  foreach($html->find('img[w]') as $element){

    $image =  $element->src; 

  }

  $data = file_get_contents($image);

   $info = getimagesize($image);//获取图片信息，大小，格式

   switch($info[2]){

     case 1:

       $str = 'gif';

       break;

     case 2:

       $str = 'jpg';

       break;

     case 3:

       $str = 'png';

       break;

     default:

       continue;

       break;

   }

   if($info[1] < 10 || $info[0] < 10) continue;//图片太小，不是有价值的图片，跳过本次循环

   $filename = time().rand(1,999999).'.'.$str; 

   if(!is_dir($dirname)){

     mkdir($dirname,0777,true);

   }

   $fp = fopen($dirname.$filename,'w');

   fwrite($fp,$data);

   fclose($fp);

   return $dirname.$filename;

   

}

  do{

    set_time_limit(0);

    ignore_user_abort();

    $img = getLink('http://www.111cn.net /qutu_1.html');

    $count = count($img['links']);

    $arr = array();

    for($i=0;$i<$count;$i++){

   $arr[]=loadimg($img['links'][$i],'images/');

    }

    $img['url'] = $arr;

    echo '<br/>';

    $img['title'];

    $res = array();

    $len = count($img['title']);

    //重新将数据组装成我们常用的二维数组，方便数据的数据库处理

    for($i=0;$i<$len;$i++){

      $res[$i]['title'] = $img['title'][$i];

   $res[$i]['url'] = $img['url'][$i];

    }

    foreach($res as $item){

      echo '@@##@@'.$item["title"].'<br />';

 

    }

    $interval = 24*3600;

    sleep($interval);

   }while(true);

  

?>