一、前言

在使用php curl抓取網頁內容時，分析出來網頁內的meta資訊，一般情況下，只是會用到meta中的content-type或者charset、keywords、description。

二、實現

寫了一個函式，用來提出meta中的鍵值對，程式碼如下：

function getMetaOfWeb($html=NULL)
{
    if(is_null($html))
    {
        return NULL;
    }
    if(strlen($html)!=0)
    {
        $ret = array();

        $metapattern = '/<meta[^>]*?>/is';
        $kvpattern = '/([\w\-]+)=[\"\']?([^\s]+)["\']?/is';
        preg_match_all($metapattern, $html, $matches);
        if(!empty($matches))
        {
            $meta = $matches[0];
            foreach ($meta as $key => $value) {
                $value = preg_replace('/<meta\s+/is', "", $value);
                $value = rtrim($value,'/>');
                $value =trim($value);

                //替換;+空格
                $repattern = '/;\s+/is';
                $value = preg_replace($repattern, ';', $value);

                preg_match_all($kvpattern, $value, $kvmatches);
                if(!empty($kvmatches))
                {
                    $type = (int)count($kvmatches[0]);
                    switch ($type) {
                        case 1://meta裡面只有一條語句 <meta charset=UTF-8>
                            $metakey = rtrim($kvmatches[1][0],"\"'");
                            $metakey    =   ltrim($metakey,"\"'");
                            $metavalue = rtrim($kvmatches[2][0],"\"'");
                            $metavalue  =   ltrim($metavalue,"\"'");
                            $ret[$metakey]  =   $metavalue;
                            break;
                        case 2://meta裡面是引數名稱/引數值的方式
                            $metakey = rtrim($kvmatches[2][0],"\"'");
                            $metakey    =   ltrim($metakey,"\"'");
                            $metavalue = rtrim($kvmatches[2][1],"\"'");
                            $metavalue  =   ltrim($metavalue,"\"'");
                            $ret[$metakey]  =   $metavalue;
                            break;
                        case 3://meta裡面;+空格<meta http-equiv=mobile-agent content="format=wml;url=http://m.qidian.com
                            $metakey = rtrim($kvmatches[2][0],"\"'");
                            $metakey    =   ltrim($metakey,"\"'");
                            $tmp = rtrim($kvmatches[2][1],"\"'");
                            $tmp = ltrim($tmp,"\"'");
                            $tmp2 = rtrim($kvmatches[2][2],"\"'");
                            $tmp2 = ltrim($tmp2,"\"'");
                            $metavalue  =   $tmp.$tmp2;
                            $ret[$metakey]  =   $metavalue;
                            break;
                    }
                }
            }
            return $ret;
        }
        return NULL;
    }
}