There are plenty of scenarios when you get passed ugly HTML and need to parse it, modify it and re-write it to how it looked before. Take for example HTML E-Mails - a magic box of strange behaviours across all the important clients like Outlook 2007+ and Webmailers.

PHPricot to the rescue, using the pecl/html_parse extension based on the EKHtml Parser to build an Abstract Syntax Tree of your HTML, no matter how ugly it is, and re-writes this AST into your ugly HTML with only minor changes:

• All attributes will be enclosed by ""
• Tags get lower-cased
• Self-Closing Tags (currently) always get /> endings
• Multi-line tags are wrapped into a single line
• Really broken tags like <a href="foo<a href="bar.html"> are separated into <a href="foo"><a href="bar.html">
• Closing tags with no matching opening tag are omitted.
• Comments always get enclosed by <!-- -->

Using the AST you can then access the nodes using CSS Selectors, cloning parts of the jQuery API.

PHPricot is motivated by Rubys HPricot but only a distantly related port.

This currently a linux only installation guide (please contribute Windows or MacOS if you find the time).

1. Install EKHtml

2. Install pecl/html_parse

   pecl install html_parse

$query = new PHPricot_Query($html);
echo $query->toHtml(); // render the AST to html
echo $query->getDocument()->toText(); // echo a text representation of your html

PHPricot supports CSS Selectors to find elements in the AST. A subset of the CSS 3 specification is supported (excluding namespaces, pseudo elements and classes aswell as the sibling operator ~).

$query = new PHPRicot_Query($html);
$links = $query->find('a.selected');

$query = new PHPRicot_Query($html);
$query->find('p')->addClass('foo')
                 ->removeClass('bar')
                 ->toggleClass('baz');
$html = $query->toHtml();

$query = new PHPricot_Query($html);
$val = $query->find('p#foo')->attr('align', 'center')->attr('align');

$query = new PHPricot_Query($html);
$query->find('.inner')->append('<p>Test</p>');
$query->find('.inner')->prepend('<p>Test</p>');

Replace the matched elements with the given HTML or PHPRicot_Query instance.

$html = '<div class="container">
  <div class="inner first">Hello</div>
  <div class="inner second">And</div>
  <div class="inner third">Goodbye</div>
</div>';

Using this PHPricot Code:

$query = new PHPricot_Query($html);
$query->find('.second')->replaceWith('<h2>New heading!</h2>');

Leads to this result:

<div class="container">
  <div class="inner first">Hello</div>
  <h2>New heading</h2>
  <div class="inner third">Goodbye</div>
</div>

• Port all the parts of the jQuery API that apply to this use-case
• Add HTML context details (what tags are are really HTML tags)
• Attempt to fix some broken HTML if possible