[Relative to absolute links with regex and DOMDocument] #domdocument #regex

relabsolut.php

<?php

function regexToAbs($text, $base)
	{
		
		if (empty($base))
		return $text;
		// base url needs trailing /
		if (substr($base, -1, 1) != "/")
		$base .= "/";

		$text = (string) trim($text);

		if (empty($text)) return $text;

		// http://wintermute.com.au/bits/2005-09/php-relative-absolute-links/
		$text=preg_replace('#(href|src)="([^:"]*)(?:")#','$1="'.$base.'$2"',$text);

		//dd($text);
		return $text;

	}
	
	
	
	function domdocToAbs($text, $base)
	{
		
		if (empty($base))
		return $text;
		// base url needs trailing /
		if (substr($base, -1, 1) != "/")
		$base .= "/";

		$text = (string) trim($text);

		if (empty($text)) return $text;
	
		// parsing HTML with RegEx is bad they say
		// now compare this sensitive pile of crap, that chokes on 
		// string encodings with the one line of regex on top of the page

		$html = $text;
		
		// you should know what you feed into
		$DOM = new DOMDocument();
		$DOM->encoding = 'utf-8';
    
		$DOM->loadHTML(utf8_decode($html));

    // promised to leave out the additional DOCTYPE>html>body the result is wrapped with
		// but results in faulty HTML
		//$DOM->loadHTML($html, LIBXML_HTML_NOIMPLIED | LIBXML_HTML_NODEFDTD);

		$imgs = $DOM->getElementsByTagName('img');
		foreach($imgs as $img){
		    $src = $img->getAttribute('src');
		    if(strpos($src, 'http') !== 0 || strpos($src, '/') !== 0){
		        $img->setAttribute('src', $base.$src);
		    }
		}

		// remove the redundant doctype/html/body DOMDocument added with Regex
		// http://stackoverflow.com/a/10023094
		$html = preg_replace('~<(?:!DOCTYPE|/?(?:html|body))[^>]*>\s*~i', '', $DOM->saveHTML());

		return $html;
	}
	
	$input = <<<HTM
	<img src="relative.html" />
	HTM;
	
	var_dump(regexToAbs($input,"http://www.de"), domdocToAbs($input,"http://www.de"));