首页 > 解决方案 > 用于文件上传/直接输入的 W3 Validator Api

问题描述

我正在尝试http://validator.w3.org/nu/通过 POST 方法使用 API 进行直接输入。

https://github.com/validator/validator/wiki/Service-%C2%BB-Input-%C2%BB-textarea

这是我尝试过但没有成功的方法

class frontend {
    public static function file_get_contents_curl($url)
    {
        $ch = curl_init();

        curl_setopt($ch, CURLOPT_HEADER, 0);
        $user_agent = self::random_user_agent();
        //var_dump($user_agent);
        curl_setopt($ch,CURLOPT_USERAGENT,$user_agent);
        curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
        curl_setopt($ch, CURLOPT_URL, $url);
        curl_setopt($ch, CURLOPT_FOLLOWLOCATION, 1);
        if (strpos($url, 'https') !== false) {
            curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, false);
        }
        $data = curl_exec($ch);
        curl_close($ch);
        return $data;
    }
}
$domain = 'yahoo.com';
$url = 'https://'.$domain;
$html = frontend::file_get_contents_curl($url);
libxml_use_internal_errors(true);
$doc = new DOMDocument;
$doc->loadHTML($html);
$html_file_output = $domain.'.html';
$dir = $_SERVER['DOCUMENT_ROOT'].'/tmp/';
if(!file_exists($dir)) {
    mkdir($dir);
}
$file_path = $dir.$html_file_output;
$doc->saveHTMLFile($file_path);
var_dump($file_path); // the filepath where the file is saved /www.domain.com/tmp/html_file.html
$url_validator = 'http://validator.w3.org/nu/';
$query = [
    'out' => 'json',
    'content' => $html // the HTML resulting from $url variable %3C%21DOCTYPE+html%3E%0....
    //'content' => $file_path tried also => /www.domain.com/tmp/the_file.html
];
$query_string = http_build_query($query);
var_dump($query_string); // returns string 'out=json&content=doctype html....' or 'out=json&content=F:/SERVER/www/www.domain.com/tmp/yahoo.com.html'
$ch = curl_init();
curl_setopt($ch, CURLOPT_POST, 1);
curl_setopt($ch, CURLOPT_POSTFIELDS, $query_string);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
$str_html = curl_exec($ch);
curl_close($ch);
$data = json_decode($str_html); 
var_dump($data); // returns null
unlink($file_path);

标签: phpapicurlposthttp-post

解决方案


首先,“直接输入”api 只接受multipart/form-data-format 的 POST 请求,但是当你通过它运行它时,http_build_query()你会将它转换为application/x-www-form-urlencoded-format,而该 api 不理解。(给 CURLOPT_POSTFIELDS 一个数组,它会自动转换为multipart/form-data

其次,此 API 会阻止缺少User-Agent标头的请求,并且 libcurl 没有默认 UA(cli 程序有,但 libcurl 没有),因此您必须自己提供一个,但您没有。

...修复这两个,并添加一些简单的错误消息解析,

<?php
$ch=curl_init();
$html=<<<'HTML'
<!DOCTYPE html>
<html lang="">
<head>
<title>Test</title>
</head><ERR&OR
<body>
<p></p>
</body>
</html>
HTML;
curl_setopt_array($ch,array(
    CURLOPT_URL=>'http://validator.w3.org/nu/',
    CURLOPT_ENCODING=>'',
    CURLOPT_USERAGENT=>'PHP/'.PHP_VERSION.' libcurl/'.(curl_version()['version']),
    CURLOPT_POST=>1,
    CURLOPT_POSTFIELDS=>array(
        'showsource'=>'yes',
        'content'=>$html
    ),
    CURLOPT_RETURNTRANSFER=>1,
));
$html=curl_exec($ch);
curl_close($ch);
$parsed=array();
$domd=@DOMDocument::loadHTML($html);
$xp=new DOMXPath($domd);
$res=$domd->getElementById("results");
foreach($xp->query("//*[@class='error']",$res) as $message){
    $parsed['errors'][]=trim($message->textContent);
}
var_dump($html);
var_dump($parsed);

印刷:

array(1) {
  ["errors"]=>
  array(4) {
    [0]=>
    string(156) "Error: Saw < when expecting an attribute name. Probable cause: Missing > immediately before.At line 6, column 1</head><ERR&ORâ©<body>â©<p></p>â©"
    [1]=>
    string(254) "Error: Element err&or not allowed as child of element body in this context. (Suppressing further errors from this subtree.)From line 5, column 8; to line 6, column 6e>â©</head><ERR&ORâ©<body>â©<p></Content model for element body:Flow content."
    [2]=>
    string(144) "Error: End tag for  body seen, but there were unclosed elements.From line 8, column 1; to line 8, column 7>â©<p></p>â©</body>â©</htm"
    [3]=>
    string(118) "Error: Unclosed element err&or.From line 5, column 8; to line 6, column 6e>â©</head><ERR&ORâ©<body>â©<p></"
  }
}

...并且unicode问题源于DOMDocument的默认字符集是.. idk,not-utf8,afaik没有用DOMDocument设置默认字符集的好方法,但是您可以通过这样做来解决它

$domd=@DOMDocument::loadHTML('<?xml encoding="UTF-8">'.$html);

这使它打印:

array(1) {
  ["errors"]=>
  array(4) {
    [0]=>
    string(147) "Error: Saw < when expecting an attribute name. Probable cause: Missing > immediately before.At line 6, column 1</head><ERR&OR↩&lt;body>↩&lt;p></p>↩"
    [1]=>
    string(245) "Error: Element err&or not allowed as child of element body in this context. (Suppressing further errors from this subtree.)From line 5, column 8; to line 6, column 6e>↩&lt;/head><ERR&OR↩&lt;body>↩&lt;p></Content model for element body:Flow content."
    [2]=>
    string(135) "Error: End tag for  body seen, but there were unclosed elements.From line 8, column 1; to line 8, column 7>↩&lt;p></p>↩&lt;/body>↩&lt;/htm"
    [3]=>
    string(109) "Error: Unclosed element err&or.From line 5, column 8; to line 6, column 6e>↩&lt;/head><ERR&OR↩&lt;body>↩&lt;p></"
  }
}

...哪个更好,但仍然包含网页上使用的箭头,可以用

foreach($xp->query("//*[@class='lf']") as $remove){
    $remove->parentNode->removeChild($remove);
}

这使它打印:

array(1) {
  ["errors"]=>
  array(4) {
    [0]=>
    string(138) "Error: Saw < when expecting an attribute name. Probable cause: Missing > immediately before.At line 6, column 1</head><ERR&OR<body><p></p>"
    [1]=>
    string(236) "Error: Element err&or not allowed as child of element body in this context. (Suppressing further errors from this subtree.)From line 5, column 8; to line 6, column 6e></head><ERR&OR<body><p></Content model for element body:Flow content."
    [2]=>
    string(126) "Error: End tag for  body seen, but there were unclosed elements.From line 8, column 1; to line 8, column 7><p></p></body></htm"
    [3]=>
    string(100) "Error: Unclosed element err&or.From line 5, column 8; to line 6, column 6e></head><ERR&OR<body><p></"
  }
}

推荐阅读