快捷導(dǎo)航

發(fā)布一個用PHP fsockopen寫的HTTP下載的類

更新時間：2007年02月22日 00:00:00 作者：

如果支持打開遠(yuǎn)程內(nèi)容的選項(xiàng)的話，實(shí)際上php用fopen或file_get_contents都能獲得一個網(wǎng)頁的內(nèi)容，但是默認(rèn)的函數(shù)有個不足的地方就是無法獲取HTTP頭，這在一些特殊的應(yīng)用中很不方便，如，有一個鏈接：

http://www.abc.com/showvd.asp?id=18

假如它返回的是一個圖片，用默認(rèn)函數(shù)就很難識別，但如果通過HTTP應(yīng)答頭來判斷就簡單多了，此外如果對方通過 Refer 來防盜鏈的話，也是無法獲取的，用HTTP類就能完美解決這些問題，而且速度也相差無幾。

使用方法：

$hd = new DedeHttpDown();
$hd->OpenUrl("http://www.dedecms.com");
echo $hd->GetHtml();
//如果保存為文件則用 $hd->SaveBin("dede.html");
$hd->Close();

獲得http請求頭用
$hd->GetHead("key")
設(shè)置請求頭
$hd->SetHead(key,value); (必須在調(diào)用 OpenUrl 之前設(shè)定)

代碼如下:

<?
/* ---------------------------------------------------------------------
//織夢Http下載類V1.0
//出自：織夢之旅 http://www.dedecms.com
//作者: IT柏拉圖
//時間: 2005-11-13 12:39
//聲明: 首發(fā)在落伍者網(wǎng)站，轉(zhuǎn)載請保留版權(quán)信息
--------------------------------------------------------------------- */
class DedeHttpDown
{
 var $m_url = "";
 var $m_urlpath = "";
 var $m_scheme = "http";
 var $m_host = "";
 var $m_port = "80";
 var $m_user = "";
 var $m_pass = "";
 var $m_path = "/";
 var $m_query = "";
 var $m_fp = "";
 var $m_error = "";
 var $m_httphead = "" ;
 var $m_html = "";
 var $m_puthead = "";
 var $BaseUrlPath = "";
 var $HomeUrl = "";
 var $JumpCount = 0;//防止多重重定向陷入死循環(huán)
 //
 //初始化系統(tǒng)
 //
 function PrivateInit($url)
 {
 if($url=="") return ;
 $urls = "";
 $urls = @parse_url($url);
 $this->m_url = $url;
 if(is_array($urls))
 {
 $this->m_host = $urls["host"];
 if(!empty($urls["scheme"])) $this->m_scheme = $urls["scheme"];

 if(!empty($urls["user"])){
 $this->m_user = $urls["user"];
 }

 if(!empty($urls["pass"])){
 $this->m_pass = $urls["pass"];
 }

 if(!empty($urls["port"])){
 $this->m_port = $urls["port"];
 }

 if(!empty($urls["path"])) $this->m_path = $urls["path"];
 $this->m_urlpath = $this->m_path;

 if(!empty($urls["query"])){
 $this->m_query = $urls["query"];
 $this->m_urlpath .= "?".$this->m_query;
 }
 $this->HomeUrl = $urls["host"];
 $this->BaseUrlPath = $this->HomeUrl.$urls["path"];
 $this->BaseUrlPath = ereg_replace("/([^/]*)\.(.*)$","/",$this->BaseUrlPath);
 $this->BaseUrlPath = ereg_replace("/$","",$this->BaseUrlPath);
 }
 }
 //
 //打開指定網(wǎng)址
 //
 function OpenUrl($url)
 {
 //重設(shè)各參數(shù)
 $this->m_url = "";
 $this->m_urlpath = "";
 $this->m_scheme = "http";
 $this->m_host = "";
 $this->m_port = "80";
 $this->m_user = "";
 $this->m_pass = "";
 $this->m_path = "/";
 $this->m_query = "";
 $this->m_error = "";
 $this->JumpCount = 0;
 $this->m_httphead = Array() ;
 //$this->m_puthead = "";
 $this->m_html = "";
 $this->Close();
 //初始化系統(tǒng)
 $this->PrivateInit($url);
 $this->PrivateStartSession();
 }
 //
 //打開303重定向網(wǎng)址
 //
 function JumpOpenUrl($url)
 {
 //重設(shè)各參數(shù)
 $this->m_url = "";
 $this->m_urlpath = "";
 $this->m_scheme = "http";
 $this->m_host = "";
 $this->m_port = "80";
 $this->m_user = "";
 $this->m_pass = "";
 $this->m_path = "/";
 $this->m_query = "";
 $this->m_error = "";
 $this->JumpCount++;
 $this->m_httphead = Array() ;
 $this->m_html = "";
 $this->Close();
 //初始化系統(tǒng)
 $this->PrivateInit($url);
 $this->PrivateStartSession();
 }
 //
 //獲得某操作錯誤的原因
 //
 function printError()
 {
 echo "錯誤信息：".$this->m_error;
 echo "具體返回頭： ";
 foreach($this->m_httphead as $k=>$v)
 { echo "$k => $v \r\n"; }
 }
 //
 //判別用Get方法發(fā)送的頭的應(yīng)答結(jié)果是否正確
 //
 function IsGetOK()
 {
 if( ereg("^2",$this->GetHead("http-state")) )
 { return true; }
 else
 {
 $this->m_error .= $this->GetHead("http-state")." - ".$this->GetHead("http-describe")." ";
 return false;
 }
 }
 //
 //看看返回的網(wǎng)頁是否是text類型
 //
 function IsText()
 {
 if(ereg("^2",$this->GetHead("http-state"))
 && eregi("^text",$this->GetHead("content-type")))
 { return true; }
 else
 {
 $this->m_error .= "內(nèi)容為非文本類型或網(wǎng)址重定向 ";
 return false;
 }
 }
 //
 //判斷返回的網(wǎng)頁是否是特定的類型
 //
 function IsContentType($ctype)
 {
 if(ereg("^2",$this->GetHead("http-state"))
 && $this->GetHead("content-type")==strtolower($ctype))
 { return true; }
 else
 {
 $this->m_error .= "類型不對 ".$this->GetHead("content-type")." ";
 return false;
 }
 }
 //
 //用Http協(xié)議下載文件
 //
 function SaveToBin($savefilename)
 {
 if(!$this->IsGetOK()) return false;
 if(@feof($this->m_fp)) { $this->m_error = "連接已經(jīng)關(guān)閉！"; return false; }
 $fp = fopen($savefilename,"w");
 while(!feof($this->m_fp)){
 fwrite($fp,fread($this->m_fp,1024));
 }
 fclose($this->m_fp);

 fclose($fp);
 return true;
 }
 //
 //保存網(wǎng)頁內(nèi)容為Text文件
 //
 function SaveToText($savefilename)
 {
 if($this->IsText()) $this->SaveBinFile($savefilename);
 else return "";
 }
 //
 //用Http協(xié)議獲得一個網(wǎng)頁的內(nèi)容
 //
 function GetHtml()
 {
 if(!$this->IsText()) return "";
 if($this->m_html!="") return $this->m_html;
 if(!$this->m_fp||@feof($this->m_fp)) return "";
 while(!feof($this->m_fp)){
 $this->m_html .= fgets($this->m_fp,256);
 }
 @fclose($this->m_fp);
 return $this->m_html;
 }
 //
 //開始HTTP會話
 //
 function PrivateStartSession()
 {

 if(!$this->PrivateOpenHost()){
 $this->m_error .= "打開遠(yuǎn)程主機(jī)出錯!";
 return false;
 }

 if($this->GetHead("http-edition")=="HTTP/1.1") $httpv = "HTTP/1.1";
 else $httpv = "HTTP/1.0";

 //發(fā)送固定的起始請求頭GET、Host信息
 fputs($this->m_fp,"GET ".$this->m_urlpath." $httpv\r\n");
 $this->m_puthead["Host"] = $this->m_host;

 //發(fā)送用戶自定義的請求頭
 if(!isset($this->m_puthead["Accept"])) { $this->m_puthead["Accept"] = "*/*"; }
 if(!isset($this->m_puthead["User-Agent"])) { $this->m_puthead["User-Agent"] = "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.2)"; }
 if(!isset($this->m_puthead["Refer"])) { $this->m_puthead["Refer"] = "http://".$this->m_puthead["Host"]; }
 foreach($this->m_puthead as $k=>$v){
 $k = trim($k);
 $v = trim($v);
 if($k!=""&&$v!=""){
 fputs($this->m_fp,"$k: $v\r\n");
 }
 }

 //發(fā)送固定的結(jié)束請求頭
 //HTTP1.1協(xié)議必須指定文檔結(jié)束后關(guān)閉鏈接,否則讀取文檔時無法使用feof判斷結(jié)束
 if($httpv=="HTTP/1.1") fputs($this->m_fp,"Connection: Close\r\n\r\n");
 else fputs($this->m_fp,"\r\n");

 //獲取應(yīng)答頭狀態(tài)信息
 $httpstas = explode(" ",fgets($this->m_fp,256));
 $this->m_httphead["http-edition"] = trim($httpstas[0]);
 $this->m_httphead["http-state"] = trim($httpstas[1]);
 $this->m_httphead["http-describe"] = "";
 for($i=2;$i<count($httpstas);$i++){
 $this->m_httphead["http-describe"] .= " ".trim($httpstas[$i]);
 }
 //獲取詳細(xì)應(yīng)答頭
 while(!feof($this->m_fp)){
 $line = trim(fgets($this->m_fp,256));
 if($line == "") break;
 $hkey = "";
 $hvalue = "";
 $v = 0;
 for($i=0;$i<strlen($line);$i++){
 if($v==1) $hvalue .= $line[$i];
 if($line[$i]==":") $v = 1;
 if($v==0) $hkey .= $line[$i];
 }
 $hkey = trim($hkey);
 if($hkey!="") $this->m_httphead[strtolower($hkey)] = trim($hvalue);
 }
 //判斷是否是3xx開頭的應(yīng)答
 if(ereg("^3",$this->m_httphead["http-state"]))
 {
 if($this->JumpCount > 3) return;
 if(isset($this->m_httphead["location"])){
 $newurl = $this->m_httphead["location"];
 if(eregi("^http",$newurl)){
 $this->JumpOpenUrl($newurl);
 }
 else{
 $newurl = $this->FillUrl($newurl);
 $this->JumpOpenUrl($newurl);
 }
 }
 else
 { $this->m_error = "無法識別的轉(zhuǎn)移應(yīng)答！"; }
 }//
 }
 //
 //獲得一個Http頭的值
 //
 function GetHead($headname)
 {
 $headname = strtolower($headname);
 if(isset($this->m_httphead[$headname]))
 return $this->m_httphead[$headname];
 else
 return "";
 }
 //
 //設(shè)置Http頭的值
 //
 function SetHead($skey,$svalue)
 {
 $this->m_puthead[$skey] = $svalue;
 }
 //
 //打開連接
 //
 function PrivateOpenHost()
 {
 if($this->m_host=="") return false;
 $this->m_fp = @fsockopen($this->m_host, $this->m_port, &$errno, &$errstr,10);
 if(!$this->m_fp){
 $this->m_error = $errstr;
 return false;
 }
 else{
 return true;
 }
 }
 //
 //關(guān)閉連接
 //
 function Close(){
 @fclose($this->m_fp);
 }
 //
 //補(bǔ)全相對網(wǎng)址
 //
 function FillUrl($surl)
 {
 $i = 0;
 $dstr = "";
 $pstr = "";
 $okurl = "";
 $pathStep = 0;
 $surl = trim($surl);
 if($surl=="") return "";
 $pos = strpos($surl,"#");
 if($pos>0) $surl = substr($surl,0,$pos);
 if($surl[0]=="/"){
 $okurl = "http://".$this->HomeUrl."/".$surl;
 }
 else if($surl[0]==".")
 {
 if(strlen($surl)<=2) return "";
 else if($surl[0]=="/")
 {
 $okurl = "http://".$this->BaseUrlPath."/".substr($surl,2,strlen($surl)-2);
 }
 else{
 $urls = explode("/",$surl);
 foreach($urls as $u){
 if($u=="..") $pathStep++;
 else if($i<count($urls)-1) $dstr .= $urls[$i]."/";
 else $dstr .= $urls[$i];
 $i++;
 }
 $urls = explode("/",$this->BaseUrlPath);
 if(count($urls) <= $pathStep)
 return "";
 else{
 $pstr = "http://";
 for($i=0;$i<count($urls)-$pathStep;$i++)
 { $pstr .= $urls[$i]."/"; }
 $okurl = $pstr.$dstr;
 }
 }
 }
 else
 {
 if(strlen($surl)<7)
 $okurl = "http://".$this->BaseUrlPath."/".$surl;
 else if(strtolower(substr($surl,0,7))=="http://")
 $okurl = $surl;
 else
 $okurl = "http://".$this->BaseUrlPath."/".$surl;
 }
 $okurl = eregi_replace("^(http://)","",$okurl);
 $okurl = eregi_replace("/{1,}","/",$okurl);
 return "http://".$okurl;
 }
}
?>

您可能感興趣的文章: