Attention: New version can break compatibility, in that case use previous version under the v1.0 branch or tag which supports even php 5.4
\nokogiri class is left for compatibility
This library is a fast HTML parser, which can work with invalid code (errors are ignored).
Under the hood is used LibXML.
As the input you can use HTML string in UTF-8 encoding or DOMDocument.
For the querying elements CSS selectors are used, which are transformed to XPath expressions internally.
HTML errors are ignored
- From HTML string
$saw = new \nokogiri($html);$saw = \nokogiri::fromHtml($html); - From DOM elements
$saw = new \nokogiri($dom);$saw = \nokogiri::fromDom($dom);
$cssSelector elements have the following format:
tagName[attribute=value]#elementId.className:pseudoSelector(expression)
$saw->get('div > a[rel=bookmark]')->toArray();Returns underlying DOM structure as an array.
Values are attributes, text content under #text key and child elements under numeric keys
Returns HTML string
Returns DOMDocument. Given true as the first argument - can also return DOMNodeList or DOMElement
foreach ($saw->get('#sidebar a.topic') as $link){
var_dump($link['#text']);
}- tag
- .class
- #id
- [attr]
- [attr=value]
- :root
- :empty
- :first-child
- :last-child
- :first-of-type
- :last-of-type
- :only-of-type
- :nth-child(a)
- :nth-child(an b)
- :nth-child(even/odd)
- DOM
- libxml >=2.9.0
- PHP >= 7.3
MIT
- Minimal PHP version 7.3
- Minimal LibXML version 2.9.0
- Complete refactoring
- Partially changed behaviour, can break compatibility
- HTML loading behaviour changed
- Test coverage
- Fixed work of nth-child and other selectors
- Incorrect selectors now throw exceptions
- New selectors added
- First version, 2011
- Minimal PHP version 5.4