淺析織夢cms的php采集類

2024-07-12 08:54:08

字體：大中小

來源：轉載

供稿：網友

以下為引用的內容：
<?php
@set_time_limit(0);
/*=======================================
//織夢Http下載類V1.1版，本版增加了POST發送數據的方式
//把要POST的數據像GET一樣跟在網址后，程序會自動處理，并用POST方式發送
//POST僅支持application/x-www-form-urlencoded方式
=======================================*/
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 $JumpCount = 0;//防止多重重定向陷入死循環
//
//初始化系統
//
function PrivateInit($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"];
$this->BaseUrlPath = preg_replace("///([^//]*)/.(.*)$/","/",$this->BaseUrlPath);
$this->BaseUrlPath = preg_replace("///$/","",$this->BaseUrlPath);
}
}
function ResetAny()
{
//重設各參數
$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 = "";
$this->m_error = "";
}
//
//打開指定網址
//
function OpenUrl($url,$requestType="GET")
{
$this->ResetAny();
$this->JumpCount = 0;
$this->m_httphead = Array() ;
$this->m_html = "";
$this->Close();
//初始化系統
$this->PrivateInit($url);
$this->PrivateStartSession($requestType);
}
//
//轉到303重定向網址
//
function JumpOpenUrl($url)
{
$this->ResetAny();
$this->JumpCount ;
$this->m_httphead = Array() ;
$this->m_html = "";
$this->Close();
//初始化系統
$this->PrivateInit($url);
$this->PrivateStartSession('GET');
}
//
//獲得某操作錯誤的原因
//
function printError()
{
echo "錯誤信息：".$this->m_error;
echo " 具體返回頭： ";
foreach($this->m_httphead as $k=>$v){ echo "$k => $v /r/n"; }
}
//
//判別用Get方法發送的頭的應答結果是否正確
//
function IsGetOK()
{
if( ereg("^2",$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
{
$this->m_error .= "內容為非文本類型或網址重定向 ";
return false;
}
}
//
//判斷返回的網頁是否是特定的類型
//
function IsContentType($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 = "連接已經關閉！"; return false; }

$fp = fopen($savefilename,"w");
while(!feof($this->m_fp)){
fwrite($fp,fread($this->m_fp,1024));
}
fclose($this->m_fp);

fclose($fp);
return true;
}
//
//保存網頁內容為Text文件
//
function SaveToText($savefilename)
{
if($this->IsText()) $this->SaveBinFile($savefilename);
else return "";
}

function SaveFile($savefilename,$somecontent=''){
$filename=$savefilename;
$somecontent = $somecontent==''?$this->GetHtml():$somecontent;
// if (is_writable($filename)) {
// 在這個例子里，我們將使用添加模式打開$filename，
// 因此，文件指針將會在文件的開頭，
// 那就是當我們使用fwrite()的時候，$somecontent將要寫入的地方。
if (!$handle = fopen($filename, 'w')) {
echo "can not open $filename";
return false;
//exit;
}
// 將$somecontent寫入到我們打開的文件中。
if (!fwrite($handle, $somecontent) ) {
echo "can not write $filename";
//exit;
return false;
}
//echo "成功地將 $somecontent 寫入到文件$filename";
fclose($handle);
//die( $savefilename);
return true;
/* } else {
echo "file $filename cannot";
return false;
}*/

}
//
//用Http協議獲得一個網頁的內容
//
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($requestType="GET")
{

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

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

$ps = explode('?',$this->m_urlpath);
//發送固定的起始請求頭GET、Host信息
if($requestType=="GET") fputs($this->m_fp,"GET ".$this->m_urlpath." $httpv/r/n");
else fputs($this->m_fp,"POST ".$ps[0]." $httpv/r/n");
$this->m_puthead["Host"] = $this->m_host;

//發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(!empty($GLOBALS['RfUrl'])) $this->m_puthead["Refer"] = $GLOBALS['RfUrl'];
else if(!isset($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");
}
}

if($requestType=="POST"){
$postdata = "";
if(count($ps)>1){
for($i=1;$i<count($ps);$i ) $postdata .= $ps[$i];
}
else{ $postdata = "OK"; }
$plen = strlen($postdata);
fputs($this->m_fp,"Content-Type: application/x-www-form-urlencoded/r/n");
fputs($this->m_fp,"Content-Length: $plen/r/n");
}

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

if($requestType=="POST"){
fputs($this->m_fp,$postdata);
}

//獲取應答頭狀態信息
$httpstas = explode(" ",fgets($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[$i]);
}
//獲取詳細應答頭
while(!feof($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
{ $this->m_error = "無法識別的答復！"; }
}//
}
//
//獲得一個Http頭的值
//
function GetHead($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;
$errno = "";
$errstr = "";
$this->m_fp = @fsockopen($this->m_host, $this->m_port, $errno, $errstr,10);
if(!$this->m_fp){
$this->m_error = $errstr;
return false;
}
else{ return true; }
}
//
//關閉連接
//
function Close(){
@fclose($this->m_fp);
}
//
//補全相對網址
//
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($surl)<=1) return "";
else if($surl[1]=="/")
{
$okurl = "http://".$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;
}
$okurl = eregi_replace("^([url]http://)[/url]","",$okurl);
$okurl = eregi_replace("/{1,}","/",$okurl);
return "http://".$okurl;
}
}
?>

上一篇：實現dedecms圖集單擊圖片翻頁的功能

下一篇：解析dedecms空間遷移步驟詳解