<?php
include 'curl2.php';
include 'phpQuery.php';
include 'QueryList.php';
set_time_limit(0);
use QL\QueryList;
function getFile($path,$url)
{
$ch = curl_init();
$fp=fopen($path, 'w');
curl_setopt($ch, CURLOPT_URL, $url);
curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 60);
curl_setopt($ch, CURLOPT_FILE, $fp);
$output = curl_exec($ch);
$info = curl_getinfo($ch);
fclose($fp);
$size = filesize($path);
if ($size != $info['size_download']) {
echo $url.$path."下载的数据不完整,请重新下载<br/>";
} else {
echo $url.$path."下载数据完整<br/>";
}
curl_close($ch);
}
$domain = 'http://demo.sucaihuo.com/';
$resourceUrl = $domain.'129/';
$saveRoot = '../Ninthpalacedraw/';
$c = new cURL();
$content = $c->get($resourceUrl);
file_put_contents('fetch.txt', $content);
$content = file_get_contents('fetch.txt');
$rules = array(
//采集id为one这个元素里面的纯文本内容
'link' => array('link','href'),
'img'=>array('img ','src'),
'script'=>array('script ','src')
);
$data = QueryList::Query($content,$rules)->getData();
//print_r($data);
$resource = '';
foreach ($data as $key => $value)
{
foreach ($value as $k => $v)
{
if(!$v) continue;
if(strpos($v, 'http') !== false)
{
//$resource = $c->get($v);
//$resource =file_get_contents($v);
$url= $v;
$uri = mb_substr($v, strpos($v, '.com/')+4);
}
else
{
//$resource = $c->get($domain.$v);
//$resource = file_get_contents($domain.$v);
$url = $resourceUrl.$v;
$uri = $v;
}
$uri = explode('/',$uri);
$filename = end($uri);
unset($uri[key($uri)]);
reset($uri);
$path = $saveRoot;
foreach ($uri as $uk => $uv) {
$path.=$uv.'/';
if(!file_exists($path))
{
if(!mkdir($path))
{
echo $path.'目录创建失败<br />';
}
}
}
$resource = $c->get($url);
file_put_contents($path.$filename, $resource);
//getFile($path.$filename, $url);
}
}
file_put_contents($saveRoot.'lottery.html', $content);
//print_r($data);