php 利用php抓取批量关键词百度推广广告中网址保存在txt文件中

结合服务器的定时任务可以定时查找关键词广告的竞争程度

<?php

$fp = @fopen ( "semallurl.txt", "a+" );
$kws1 = "上海酒店,北京酒店,广州酒店,天津酒店,广州酒店";
$kws = explode ( ",", $kws1 );
foreach ( $kws as $kw ) {
    $keywords = $kw;
    $enkeywords = urlencode ( $keywords );
    $pageURL = "http://www.baidu.com/s?word=$enkeywords";
    $contents = fetch ( $pageURL ); /* 抓取页面 */
    $contents = preg_replace ( '/<script[^>]*?>.*?<\\/script>/', "", $contents ); /*
                                                                               *
                                                                               * 去掉js代码
                                                                               */
    $contents_left = "";
    $contents_right = "";
    $ads_left_green = "";
    $ads_left_white = "";
    $contentsbytwoside = "";
    $ads_right = ""; /* 变量初始化 */
    $contentsbytwoside = explode ( '<div id="content_', $contents );
    $contents_right = $contentsbytwoside [1];
    $contents_right = '<div id="content_' . $contents_right; /* 搜索结果右边部分 */
    $contents_left = $contentsbytwoside [2];
    $contents_left = '<div id="content_' . $contents_left; /* 搜索结果左边部分 */
    preg_match_all ( '/(<div id=\\"[0-9]*\\" class=\\"ec_pp_f ec_pp_top.*?)<a href=\\"http:\\/\\/e\\.baidu\\.com\\//', $contents_left, $ads_left_white );
    preg_match_all ( '/(<table class=\\"EC_mr15 EC_ppim_top ec_pp_f.*?<\\/table>)/', $contents_left, $ads_left_green );
    preg_match_all ( '(<div id=\\"bdfs[^>]*class=\\"EC_im EC_fr EC_PP  EC_idea1017 \\">.*?<a class=\\"EC_BL EC_desc\\".*?<\\/a>)', $contents_right, $ads_right );
    echo "------------Keywords ads for" . $kw . "start ------------------------------------<br>";
    fwrite ( $fp, "----------" . $kw . " ads start------------------------- \\r\\n" );

    echo "left ads with green background is<br>";
    /* print_r($ads_left_green[0]); */
    foreach ( $ads_left_green [0] as $tg1 ) {
        preg_match ( '/<span>.*?<\\/span>/', $tg1, $tg11 );
        fwrite ( $fp, strip_tags ( $tg11 [0] ) . "\\r\\n" );
        echo $tg11 [0] . "<br>";
    }
    ;

    echo "<p>-------------<br>";
    echo "left ads with white background is<br>";
    /* print_r($ads_left_white[0]); */
    foreach ( $ads_left_white [0] as $tg2 ) {

        preg_match ( '/<span class=\\"ec_url\\">.*?<\\/span>/', $tg2, $tg22 );
        fwrite ( $fp, strip_tags ( $tg22 [0] ) . "\\r\\n" );
        echo $tg22 [0] . "<br>";
    }
    ;

    echo "<p>-------------<br>";
    echo "right ads with is<br>";
    /* print_r($ads_right[0]); */

    foreach ( $ads_right [0] as $tg3 ) {

        preg_match ( '/(<font size=\\"-1\\" class=\\"EC_url\\">.*?<\\/font>)/', $tg3, $tg33 );
        fwrite ( $fp, strip_tags ( $tg33 [0] ) . "\\r\\n" );
        echo $tg33 [0] . "<br>";
    }
    ;

    echo "---------------Keywords ads for" . $kw . "END ------------------------------------<br>";
    fwrite ( $fp, "----------" . $kw . " ads End------------------------- \\r\\n" );
}
;

fwrite ( $fp, date ( "Y-m-d H:i:s" ) . " PHP代码自动运行！\\r\\n" );

fclose ( $fp );

function fetch($Date) {
    $ch = curl_init ();
    $timeout = 5;
    curl_setopt ( $ch, CURLOPT_URL, "$Date" );
    curl_setopt ( $ch, CURLOPT_RETURNTRANSFER, 1 );
    curl_setopt ( $ch, CURLOPT_USERAGENT, "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1)" );
    curl_setopt ( $ch, CURLOPT_CONNECTTIMEOUT, $timeout );
    $contents = curl_exec ( $ch );
    curl_close ( $ch );
    return $contents;
}
?>
//该片段来自于http://outofmemory.cn