<form action="index.php" method="post"> <input type="text" /> <input type="submit" value="Get Keywords" /> </form> <?php header('Content-Type:text/html;charset=gbk'); class ComBaike{ private $o_String=NULL; public function __construct(){ include('cls.StringEx.php'); $this->o_String=new StringEx(); } public function getItem($word){ $url = "http://www.baidu.com/s?wd=".$word; // 构造包头，模拟浏览器请求 $header = array ( "Host:", "Content-Type:application/x-www-form-urlencoded",//post请求 "Connection: keep-alive", 'Referer:', 'User-Agent: Mozilla/5.0 (compatible; MSIE 9.0; Windows NT 6.1; WOW64; Trident/5.0; BIDUBrowser 2.6)' ); $ch = curl_init (); curl_setopt ( $ch, CURLOPT_URL, $url ); curl_setopt ( $ch, CURLOPT_HTTPHEADER, $header ); curl_setopt ( $ch, CURLOPT_RETURNTRANSFER, 1 ); $content = curl_exec ( $ch ); if ($content == FALSE) { echo "error:" . curl_error ( $ch ); } curl_close ( $ch ); //输出结果echo $content; $this->o_String->string=$content; $s_begin='<div>'; $s_end='</div>'; $summary=$this->o_String->getPart($s_begin,$s_end); $s_begin='<div>相关搜索</div><table cellpadding="0"><tr><th>'; $s_end='</th></tr></table></div>'; $content=$this->o_String->getPart($s_begin,$s_end); return $content; } public function __destruct(){ unset($this->o_String); } } if($_POST){ $com = new ComBaike(); $q = $_POST['q']; $str = $com->getItem($q); //获取搜索内容 $pat = '/<a(.*?)href="(.*?)" (.*?)>(.*?)<\/a>/i'; preg_match_all($pat, $str, $m); //print_r($m[4]); 链接文字 $con = implode(",", $m[4]); //生成文件夹 $dates = date("Ymd"); $path="./Search/".$dates."https://www.jb51.net/"; if(!is_dir($path)){ mkdir($path,0777,true); } //生成文件 $file = fopen($path.iconv("UTF-8","GBK",$q).".txt",'w'); if(fwrite($file,$con)){ echo $con; echo '<script>alert("success")</script>'; }else{ echo '<script>alert("error")</script>'; } fclose($file); } ?>
cls.StringEx.php：
<?php header('Content-Type: text/html; charset=UTF-8'); class StringEx{ public $string=''; public function __construct($string=''){ $this->string=$string; } public function pregGetPart($s_begin,$s_end){ $s_begin==preg_quote($s_begin); $s_begin=str_replace('https://www.jb51.net/','\/',$s_begin); $s_end=preg_quote($s_end); $s_end=str_replace('https://www.jb51.net/','\/',$s_end); $pattern='https://www.jb51.net/'.$s_begin.'(.*?)'.$s_end.'https://www.jb51.net/'; $result=preg_match($pattern,$this->string,$a_match); if(!$result){ return $result; }else{ return isset($a_match[1])?$a_match[1]:''; } } public function strstrGetPart($s_begin,$s_end){ $string=strstr($this->string,$s_begin); $string=strstr($string,$s_end,true); $string=str_replace($s_begin,'',$string); $string=str_replace($s_end,'',$string); return $string; } public function getPart($s_begin,$s_end){ $result=$this->pregGetPart($s_begin,$s_end); if(!$result){ $result=$this->strstrGetPart($s_begin,$s_end); } return $result; } } ?>
更多关于PHP相关内容感兴趣的读者可查看本站专题：《php curl用法总结》、《PHP网络编程技巧总结》、《PHP数组(Array)操作技巧大全》、《php字符串(string)用法总结》、《PHP数据结构与算法教程》及《PHP中json格式数据操作技巧汇总》