正则表达式可查找HTML字符串中的所有路径

Question

I have a string, with a htmlentities encoded HTML code.

What I want to do is find all the paths in the document, between:

href="XXX", src="XXX".

I do have a regex expression that find all the links starting by http, https, ftp and file, and lest me iterate over it:

"/\b(?:(?:https?|ftp|file):\/\/|www\.|ftp\.)[-A-Z0-9+&@#\/%=~_|$?!:,.]*[A-Z0-9+&@#\/%=~_|$]/i"

Any idea?

Answer 1

Update: Doing it with a regex isn't reliable. The src=".." or href=".." statement can be part of a comment or a javascript statement. To reliable obtain the links I would suggest to use XPath:

<?php

$html = file_get_contents('http://stackoverflow.com/questions/14782334/regex-expression-to-find-all-paths-in-a-html-string/14782594#14782594');
$doc = new DOMDocument();
@$doc->loadHTML($html);
$selector = new DOMXPath($doc);

$result = $selector->query('//a/@href | //@src');
foreach($result as $link) {
    echo $link->value, PHP_EOL;
}

---

If using a regex I would try to grab the content between the = " of the href or src attribute. Here comes an example how to get the links from this page using a regex:

<?php

$html = file_get_contents('http://stackoverflow.com/questions/14782334/regex-expression-to-find-all-paths-in-a-html-string');

preg_match_all('/href="(?P<href>.*)"|src="(?P<src>.*)"/U', $html, $m);
                                                        <--- note the U to make the 
                                                             pattern ungreedy
var_dump($m['href']);
var_dump($m['src']);

Answer 2

You could use the DOM to find all links in specific tags. For example to get urls from anchor tags do something like this (untested, but it should point you in the right direction):

function findPaths($url)
{
   $dom = new DOMDocument();

   //$url of page to search, the "@' is there to suppress warnings
   @$dom->loadHTMLFile($url) 

   $paths = array();
   foreach($dom->getElementsByTagName('a') as $path)
   {
     $paths[] = array('url' => $path->getAttribute('href'), text => $path->nodeValue);
   }
   return $paths;
}

You could make it even easier using XPath to load and evaluate the DOM.