<?php
/*
author: ssh_kobe
date: 20110602
shortage: 如果網頁中的圖片路徑不是絕對路徑，就無法抓取
*/
set_time_limit(0);//抓取不受時間限制
$URL='http://pp.baidu.com/';//任意網址 get_pic($URL); function get_pic($pic_url) {
//擷取圖片二進位流
$data=CurlGet($pic_url);
/*利用Regex得到圖片連結*/
$pattern_src = '/<[img|IMG].*?src=['|"](.*?(?:[.gif|.jpg]))['|"].*?[/]?>/';
$num = preg_match_all($pattern_src, $data, $match_src);
$arr_src=$match_src[1];//獲得圖片數組
get_name($arr_src); echo "<br>finished!!!";
return 0;
} /*得到圖片類型，並將其儲存到與該檔案同一目錄*/
function get_name($pic_arr)
{
//圖片類型
$pattern_type = '/(/.(jpg|bmp|jpeg|gif|png))/'; foreach($pic_arr as $pic_item){//迴圈取出每幅圖的地址
$num = preg_match_all($pattern_type, $pic_item, $match_type);
$pic_name = get_unique().$match_type[1][0];//改時微秒時間戳記命名
//以流的形式儲存圖片
$write_fd = @fopen($pic_name,"wb");
@fwrite($write_fd, CurlGet($pic_item));
@fclose($write_fd);
echo "[OK]..!";
}
return 0;
} //通過微秒時間獲得唯一ID
function get_unique(){
list($msec, $sec) = explode(" ",microtime());
return $sec.intval($msec*1000000);
} //抓取網頁內容
function CurlGet($url){
$url=str_replace('&','&',$url);
$curl = curl_init();
curl_setopt($curl, CURLOPT_URL, $url);
curl_setopt($curl, CURLOPT_HEADER, false); //curl_setopt($curl, CURLOPT_REFERER,$url);
curl_setopt($curl, CURLOPT_USERAGENT, "Mozilla/4.0 (compatible; MSIE 6.0; SeaPort/1.2; Windows NT 5.1; SV1; InfoPath.2)");
curl_setopt($curl, CURLOPT_COOKIEJAR, 'cookie.txt');
curl_setopt($curl, CURLOPT_COOKIEFILE, 'cookie.txt');
curl_setopt($curl, CURLOPT_RETURNTRANSFER, 1);
curl_setopt($curl, CURLOPT_FOLLOWLOCATION, 0);
$values = curl_exec($curl);
curl_close($curl);
return $values;
}
?> |
