PHP简单网页图片抓取类

使用例子 ?php include imgCatcher.class.php; header(content-type:text/html;charset=utf-8); $catcher = New imgCatcher(http://www.oschina.net/,images,jpg); $catcher-get_pic(); ? 无 ?php class imgCatcher{const FLODER_DIR = 'image/';const IMG_TY

url = $url;
			
			$url_arr = parse_url($url);

			$this->domain = array_shift($url_arr);
			
			if(trim($floder) == ""){
				$this->folder = FLODER_DIR;
			}else{
				$this->folder = $floder.'//m.sbmmt.com/m/';
			}

			if(trim($type)==""){
				$this->type = IMG_TYPE;
			}else{
				$this->type = $type;	
			}
			
			$this->pic_arr = array();
			if(!is_dir(__DIR__.'//m.sbmmt.com/m/'.$this->folder))
				mkdir(__DIR__.'//m.sbmmt.com/m/'.$this->folder);
		}

		public function get_pic(){
			set_time_limit(0);//抓取不受时间限制  
	        //获取图片二进制流  
	        $data = self::CurlGet($this->url);  
	        //利用正则表达式得到图片链接  
	        $pattern_src = '/type.')).*?>/';// 
	        $num = preg_match_all($pattern_src, $data, $match_src);  

	        $this->pic_arr=$match_src[1];//获得图片数组


	        $this->get_name();   

	        return 0;  
		} 

		public function get_name(){  
			$pic_arr = $this->pic_arr;
			
	      	//图片编号和类型  
	      	$pattern_type = '/.*\/(.*?)$/';  
	        
	      	foreach($pic_arr as $pic_item){//循环取出每幅图的地址  
	       	
	       		$num = preg_match_all($pattern_type,$pic_item,$match_type);  
	       		//以流的形式保存图片  
	       		$write_fd = fopen($this->folder.$match_type[1][0],"wb");  

	       		fwrite($write_fd, self::CurlGet($pic_item,$this->url));  
	       		fclose($write_fd);  
	      	}

      	    return 0;  
     	} 


	     //抓取网页内容  
	    static function CurlGet($url,$domain=""){
	        if(substr($url, 0, 1) == "//m.sbmmt.com/m/"){
	        	$url =  $domain.$url;
	        }

	        if(substr($url, 0, 1) == "."){
	        	$url =  $domain.substr($url, 1);
	        }

	        if(substr($url, 0, 2) == ".."){
	        	$url =  $domain.substr($url, 2);
	        }	        
	        $url=str_replace('&','&',$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;  
	    }     

	}


 ?>