Extrakt Meta-Beschreibung Inhalt unter Verwendung von DOMDocument() und XPath in PHP

Question

01.00 versuchen, Meta-Beschreibung Inhalt zu extrahieren, von Seiten &-Anzeige in den Suchergebnissen. Aber seine Anzeige es wie:

content = „Lorem Ipsum Karotten, verbessern ökologische Umwelt Jasmin Bequemlichkeit Reservierte Bedürfnisse Jasmin Massen Wenn Pulls Rays Super Bowl Berge sofort.... "

Während nur 1 wollen:

Lorem ipsum Karotten, verbessern ökologische Umwelt. Reserviert Jasmin Bequemlichkeit muss. Jasmin Masse. Wenn Pulls Rays Super Bowl Berge sofort.

Jede Vermutung, was in meinem Code falsch?

Code:

$doc = new DOMDocument(); 
    @$doc->loadHTMLFile($page_path); 
    $xpath = new DOMXPath($doc); 

    $body = $xpath->query('//meta[@name="description"]/@content'); 
    $page_title = @$doc->getElementsByTagName('title')->item(0)->textContent; 
    $page_title = $page_title ? $page_title : $page_path; 
    $page_body = html2text($doc->saveXml($body->item(0)));// this is meta-description, which i want 

    Functions : 

    function html2text($html) 
    { 
    $text = $html; 
    static $search = array(
    '@<script.+?</script>@usi', // Strip out javascript content 
    '@<style.+?</style>@usi', // Strip style content 
    '@<!--.+?-->@us',   // Strip multi-line comments including CDATA 
    '@</?[a-z].*?\>@usi',   // Strip out HTML tags 
); 
    $text = preg_replace($search, ' ', $text); 
    /* 
    * normalize common entities 
    */ 
    $text = normalizeEntities($text); 
    /* 
    * decode other entities 
    */ 
    $text = html_entity_decode($text, ENT_QUOTES, 'utf-8'); 
    /* 
    * normalize possibly repeated newlines, tabs, spaces to spaces 
    */ 
    $text = preg_replace('/\s+/u', ' ', $text); 
    $text = trim($text); 
    return $text; 
    } 


    /** 
    * Replace encoded and double encoded entities to equivalent unicode character 
    * @param string $text 
    * @return string - the same as $text but without encoded entries 
    * @access public 
    */ 
    function normalizeEntities($text) 
    { 
    static $find = array(); 
    static $repl = array(); 
    if (!count($find)) { 
    /* 
    * build $find and $replace from map one time 
    */ 
    $map = array(
    array('\'', 'apos', 39, 'x27'), // Apostrophe 
    array('\'', '‘', 'lsquo', 8216, 'x2018'), // Open single quote 
    array('\'', '’', 'rsquo', 8217, 'x2019'), // Close single quote 
    array('"', '“', 'ldquo', 8220, 'x201C'), // Open double quotes 
    array('"', '”', 'rdquo', 8221, 'x201D'), // Close double quotes 
    array('\'', '‚', 'sbquo', 8218, 'x201A'), // Single low-9 quote 
    array('"', '„', 'bdquo', 8222, 'x201E'), // Double low-9 quote 
    array('\'', '′', 'prime', 8242, 'x2032'), // Prime/minutes/feet 
    array('"', '″', 'Prime', 8243, 'x2033'), // Double prime/seconds/inches 
    array(' ', 'nbsp', 160, 'xA0'), // Non-breaking space 
    array('-', '‐', 8208, 'x2010'), // Hyphen 
    array('-', '–', 'ndash', 8211, 150, 'x2013'), // En dash 
    array('--', '—', 'mdash', 8212, 151, 'x2014'), // Em dash 
    array(' ', ' ', 'ensp', 8194, 'x2002'), // En space 
    array(' ', ' ', 'emsp', 8195, 'x2003'), // Em space 
    array(' ', ' ', 'thinsp', 8201, 'x2009'), // Thin space 
    array('*', '•', 'bull', 8226, 'x2022'), // Bullet 
    array('*', '‣', 8227, 'x2023'), // Triangular bullet 
    array('...', '…', 'hellip', 8230, 'x2026'), // Horizontal ellipsis 
    array('°', 'deg', 176, 'xB0'), // Degree 
    array('€', 'euro', 8364, 'x20AC'), // Euro 
    array('¥', 'yen', 165, 'xA5'), // Yen 
    array('£', 'pound', 163, 'xA3'), // British Pound 
    array('©', 'copy', 169, 'xA9'), // Copyright Sign 
    array('®', 'reg', 174, 'xAE'), // Registered Sign 
    array('™', 'trade', 8482, 'x2122') // TM Sign 
); 
    foreach ($map as $e) { 
    for ($i = 1; $i < count($e); ++$i) { 
    $code = $e[$i]; 
    if (is_int($code)) { 
    // numeric entity 
    $regex = "/&(amp;)?#0*$code;/"; 
    } elseif (preg_match('/^.$/u', $code)/* one unicode char*/) { 
    // single character 
    $regex = "/$code/u"; 
    } elseif (preg_match('/^x([0-9A-F]{2}){1,2}$/i', $code)) { 
    // hex entity 
    $regex = "/&(amp;)?#x0*" . substr($code, 1) . ";/i"; 
    } else { 
    // named entity 
    $regex = "/&(amp;)?$code;/"; 
    } 
    $find[] = $regex; 
    $repl[] = $e[0]; 
    } 
    } 
    } 
    return preg_replace($find, $repl, $text); 
    } 

Answer 1

Sie einen Attributknoten als XML speichern. Bloß nicht! Lesen Sie einfach ihren Wert.

: Attributknoten (DOMAttr) einen Eigenschaftswert ab, den Attributwert zurückgibt. Der Attributwert ist ein Textwert.

$html = <<<'HTML' 
<meta name="description" content="Some description"> 
HTML; 

$document = new DOMDocument(); 
$document->loadHTML($html); 
$xpath = new DOMXPath($document); 

$description = $xpath->evaluate('//meta[@name="description"]/@content')->item(0); 
var_dump($description->value); 

Ausgang:

string(16) "Some description" 

Aber XPath können den Wert direkt als String zurück. Nur warf das Ergebnis (in der XPath). Dies wird nur mit DOMXpath::evaluate() arbeiten. DOMXpath::query() kann nur Knotenliste zurückzukehren.

$description = $xpath->evaluate('string(//meta[@name="description"]/@content)'); 
var_dump($description); 

Ausgang:

string(16) "Some description" 

Es wird auch für andere Knoten arbeiten. Wenn Sie ein Element Knoten werfen (wie body) nur in ihrem Textinhalt zurückgeschickt werden, Organisationen wie © erhalten decodiert.

$html = <<<'HTML' 
<html> 
    <head> 
    <title>The Title</title> 
    <meta name="description" content="Some description"> 
    </head> 
    <body> 
    <p>Some content &amp; entities &copy;</p> 
    </body> 
</html> 
HTML; 

$document = new DOMDocument(); 
$document->loadHTML($html); 
$xpath = new DOMXPath($document); 

$title = $xpath->evaluate('string(//head/title)'); 
$description = $xpath->evaluate('string(//meta[@name="description"]/@content)'); 
$content = $xpath->evaluate('string(//body)'); 

var_dump($title, $description, $content); 

Ausgang:

string(9) "The Title" 
string(16) "Some description" 
string(36) " 
    Some content & entities © 
    "