Wa72\HtmlPageDom is a PHP library for easy manipulation of HTML documents using DOM. It requires DomCrawler from Symfony2 components for traversing the DOM tree and extends it by adding methods for manipulating the DOM tree of HTML documents.

It's useful when you need to not just extract information from an HTML file (what DomCrawler does) but also to modify HTML pages. It is usable as a template engine: load your HTML template file, set new HTML content on certain elements such as the page title, div#content or ul#menu and print out the modified page.

Wa72\HtmlPageDom consists of two classes:

• HtmlPageCrawler extends Symfony\Components\DomCrawler by adding jQuery inspired, HTML specific DOM manipulation functions such as html(), before(), append(), wrap(), addClass() or css(). It's like jQuery for PHP: simply select elements of an HTML page using CSS selectors and change their attributes and content.

• HtmlPage represents one complete HTML document and offers convenience functions like setTitle($title), setMeta('description', $description), getBody(). It uses HtmlPageCrawler for filtering and manipulation DOM Elements.

• PHP 5.3+
• Symfony\Components\DomCrawler
• Symfony\Components\CssSelector

• using [composer] (http://getcomposer.org): require "wa72/htmlpagedom":"~1.1"

• using other [PSR-4] (http://www.php-fig.org/psr/psr-4/) compliant autoloader: clone this project to where your included libraries are and point your autoloader to look for the "\Wa72\HtmlPageDom" namespace in the "src" directory of this project

HtmlPageCrawler is a wrapper around DOMNodes. HtmlPageCrawler objects can be created using new or the static function HtmlPageCrawler::create(), which accepts an HTML string or a DOMNode (or an array of DOMNodes or a DOMNodeList) as arguments.

Afterwards you can select nodes from the added DOM tree by calling filter() (equivalent to find() in jQuery) and alter the selected elements using the following jQuery-like manipulation functions:

• addClass(), hasClass(), removeClass(), toggleClass()
• after(), before()
• append(), appendTo()
• makeClone() (equivalent to clone() in jQuery)
• css() (alias getStyle() / setStyle())
• html() (alias getInnerHtml() / setInnerHtml())
• attr() (alias getAttribute() / setAttribute()), removeAttr()
• insertAfter(), insertBefore()
• makeEmpty() (equivalent to empty() in jQuery)
• prepend(), prependTo()
• remove()
• replaceAll(), replaceWith()
• text()
• unwrap()
• wrap(), wrapInner(), wrapAll()

To get the modified DOM as HTML code use html() (returns innerHTML of the first node in your crawler object) or saveHTML() (returns combined "outer" HTML code of all elements in the list).

Example:

use \Wa72\HtmlPageDom\HtmlPageCrawler;

// create an object from a fragment of HTML code as you would do with jQuery's $() function
$c = HtmlPageCrawler::create('<div id="content"><h1>Title</h1></div>');

// the above is the same as calling:
$c = new HtmlPageCrawler('<div id="content"><h1>Title</h1></div>');

// filter for h1 elements and wrap them with an HTML structure
$c->filter('h1')->wrap('<div class="innercontent">');

// return the modified HTML
echo $c->saveHTML();
// or simply:
echo $c; // implicit __toString() calls saveHTML()
// will output: <div id="content"><div class="innercontent"><h1>Title</h1></div></div>

Advanced example: remove the third column from an HTML table

use \Wa72\HtmlPageDom\HtmlPageCrawler;
$html = <<<END
<table>
    <tr>
        <td>abc</td>
        <td>adsf</td>
        <td>to be removed</td>
    </tr>
    <tr>
        <td>abc</td>
        <td>adsf</td>
        <td>to be removed</td>
    </tr>
    <tr>
        <td>abc</td>
        <td>adsf</td>
        <td>to be removed</td>
    </tr>
</table>    
END;  

$c = HtmlPageCrawler::create($html);
$tr = $c->filter('table > tr > td')
    ->reduce(
        function ($c, $j) {
            if (($j+1) % 3 == 0) {
                return true;
            }
            return false;
        }
    );
$tr->remove();
echo $c->saveHTML();

Usage examples for the HtmlPage class:

use \Wa72\HtmlPageDom\HtmlPage;

// create a new HtmlPage object with an empty HTML skeleton
$page = new HtmlPage();

// or create a HtmlPage object from an existing page
$page = new HtmlPage(file_get_contents('http://www.heise.de'));

// get or set page title
echo $page->getTitle();
$page->setTitle('New page title');
echo $page->getTitle();


// add HTML content
$page->filter('body')->setInnerHtml('<div id="#content"><h1>This is the headline</h1><p class="text">This is a paragraph</p></div>');

// select elements by css selector
$h1 = $page->filter('#content h1');
$p = $page->filter('p.text');

// change attributes and content of an element
$h1->addClass('headline')->css('margin-top', '10px')->setInnerHtml('This is the <em>new</em> headline');

$p->removeClass('text')->append('<br>There is more than one line in this paragraph');

// add a new paragraph to div#content
$page->filter('#content')->append('<p>This is a new paragraph.</p>');

// add a class and some attribute to all paragraphs
$page->filter('p')->addClass('newclass')->setAttribute('data-foo', 'bar');


// get HTML content of an element
echo $page->filter('#content')->saveHTML();

// output the whole HTML page
echo $page->save();
// or simply:
echo $page;

• HtmlPageDom builds on top of PHP's DOM functions and uses the loadHTML() and saveHTML() methods of the DOMDocument class. That's why it's output is always HTML, not XHTML.

• The HTML parser used by PHP is built for HTML4. It throws errors on HTML5 specific elements which are ignored by HtmlPageDom, so HtmlPageDom is usable for HTML5 with some limitations.

• HtmlPageDom has not been tested with character encodings other than UTF-8.

• According to failing Travis-CI tests, it does not work with HHVM (fixes welcome) Update 2015-05-21: [Travis CI test with HipHop VM 3.5.0] (https://travis-ci.org/wasinger/htmlpagedom/jobs/63323545) now passes

When I discovered how easy it was to modify HTML documents using jQuery I looked for a PHP library providing similar possibilities for PHP.

Googling around I found [SimpleHtmlDom] (http://simplehtmldom.sourceforge.net) and later [Ganon] (http://code.google.com/p/ganon) but both turned out to be very slow. Nevertheless I used both libraries in my projects.

When Symfony2 appeared with it's DomCrawler and CssSelector components I thought: the functions for traversing the DOM tree and selecting elements by CSS selectors are already there, only the manipulation functions are missing. Let's implement them! So the HtmlPageDom project was born.

It turned out that it was a good choice to build on PHP's DOM functions: Compared to SimpleHtmlDom and Ganon, HmtlPageDom is lightning fast. In one of my projects, I have a PHP script that takes a huge HTML page containing several hundreds of article elements and extracts them into individual HTML files (that are later on demand loaded by AJAX back into the original HTML page). Using SimpleHtmlDom it took the script 3 minutes (right, minutes!) to run (and I needed to raise PHP's memory limit to over 500MB). Using Ganon as HTML parsing and manipulation engine it took even longer, about 5 minutes. After switching to HtmlPageDom the same script doing the same processing tasks is running only about one second (all on the same server). HtmlPageDom is really fast.

© 2015 Christoph Singer, Web-Agentur 72. Licensed under the MIT License.