kagekiri

Shadow DOM-piercing query APIs. Supports:

API	Interface	Support
querySelector	Element, Document	✅
querySelectorAll	Element, Document	✅
getElementsByClassName	Element, Document	✅
getElementsByTagName	Element, Document	✅
getElementsByTagNameNS	Element, Document	✅
getElementById	Document	✅
getElementsByName	Document	✅
matches	Element	✅
closest	Element	✅

Usage

Install:

npm install --save kagekiri

Query the document or a specific element:

import { querySelector, querySelectorAll } from 'kagekiri'

// query the document
const elements = querySelectorAll('.container button')
const element = querySelector('.container button')

// or a specific element
const elements = querySelectorAll('button', otherElement)
const element = querySelector('button', otherElement)

Example

A custom element:

<my-component></my-component>
<script>
  class MyComponent extends HTMLElement {
    constructor() {
      super()
      const shadowRoot = this.attachShadow({mode: 'open'})
      shadowRoot.innerHTML = '<span class="hello">Hello</span>'
    }
  }
  customElements.define('my-component', MyComponent)
</script>

It renders as:

<my-component>
  <!-- shadow root (open) -->
  <span class="hello">Hello</span>
</my-component>

You can't query the .hello element:

document.querySelector('.hello')    // undefined 😞
document.querySelectorAll('.hello') // empty 😞

But with kagekiri you can!

kagekiri.querySelector('.hello')    // <span> 😃
kagekiri.querySelectorAll('.hello') // [<span>] 😃

Your can even query across the shadow boundary!

kagekiri.querySelector('my-component .hello')   // <span> 😃
kagekiri.querySelector('my-component > .hello') // <span> 😃

API

closest

▸ closest(selector: string, element: Node): Element | null

Find the closest ancestor of an element (or the element itself) matching the given CSS selector. Analogous to Element.closest

Parameters:

Name	Type	Description
selector	string	CSS selector
element	Node	target element to match against, and whose ancestors to match against

Returns: Element | null

---

getElementById

▸ getElementById(id: string, context?: DocumentOrShadowRoot): Element | null

Query for an element matching the given ID, or null if not found. Analogous to Document.getElementById

The default context is document. Choose another DocumentOrShadowRoot to query within that context.

Parameters:

Name	Type	Description
id	string	element ID
context?	DocumentOrShadowRoot	context to query in, or document by default

Returns: Element | null

---

getElementsByClassName

▸ getElementsByClassName(names: string, context?: Node): Element[]

Query for all elements matching a given class name, or multiple if a whitespace-separated list is provided. Analogous to Document.getElementsByClassName.

Unlike the standard API, this returns a static array of Elements rather than a live HTMLCollection.

The default context is document. Choose another node to query within that context.

Parameters:

Name	Type	Description
names	string	class name or whitespace-separated class names
context?	Node	context to query in, or document by default

Returns: Element[]

---

getElementsByName

▸ getElementsByName(name: string, context?: DocumentOrShadowRoot): Element[]

Query for all elements matching a given name. Analogous to Document.getElementsByName

The default context is document. Choose another DocumentOrShadowRoot to query within that context.

Unlike the standard API, this returns a static array of Elements rather than a live NodeList.

Parameters:

Name	Type	Description
name	string	element name attribute
context?	DocumentOrShadowRoot	context to query in, or document by default

Returns: Element[]

---

getElementsByTagName

▸ getElementsByTagName(tagName: string, context?: Node): Element[]

Query for all elements matching a given tag name. Analogous to Document.getElementsByTagName. The "*" query is supported.

Unlike the standard API, this returns a static array of Elements rather than a live HTMLCollection.

The default context is document. Choose another node to query within that context.

Parameters:

Name	Type	Description
tagName	string	name of the element tag
context?	Node	context to query in, or document by default

Returns: Element[]

---

getElementsByTagNameNS

▸ getElementsByTagNameNS(namespaceURI: string, localName: string, context?: Node): Element[]

Query for all elements matching a given tag name and namespace. Analogous to Document.getElementsByTagNameNS. The "*" query is supported.

Unlike the standard API, this returns a static array of Elements rather than a live NodeList.

The default context is document. Choose another node to query within that context.

Parameters:

Name	Type	Description
namespaceURI	string	namespace URI, or "*" for all
localName	string	local name, or "*" for all
context?	Node	context to query in, or document by default

Returns: Element[]

---

matches

▸ matches(selector: string, element: Node): boolean

Return true if the given Node matches the given CSS selector, or false otherwise. Analogous to Element.closest

Parameters:

Name	Type	Description
selector	string	CSS selector
element	Node	element to match against

Returns: boolean

---

querySelector

▸ querySelector(selector: string, context?: Node): Element | null

Query for a single element matching the CSS selector, or return null if not found. Analogous to Document.querySelector

The default context is document. Choose another element or DocumentOrShadowRoot to query within that context.

Parameters:

Name	Type	Description
selector	string	CSS selector
context?	Node	context to query in, or document by default

Returns: Element | null

---

querySelectorAll

▸ querySelectorAll(selector: string, context?: Node): Element[]

Query for all elements matching a CSS selector. Analogous to Document.querySelectorAll

The default context is document. Choose another node to query within that context.

Parameters:

Name	Type	Description
selector	string	CSS selector
context?	Node	context to query in, or document by default

Returns: Element[]

How it works

kagekiri parses the CSS selector using postcss-selector-parser. Then it queries the entire DOM tree, traverses any shadowRoots it may find, and checks the selector from children to ancestors (the same way a browser would).

Note that it only works on open shadow DOM. Closed shadow DOM cannot be traversed.

Slotted elements are considered to be children of their slots (inside the shadow DOM) rather than children of their host components. If you don't want this behavior, you can use the normal DOM APIs (e.g. document.querySelector() or document.querySelectorAll()).

See the tests for full supported CSS features.

The name

陰 kage (shadow) + 切り kiri (cut).

Roughly, "shadow-cutter."

Build

npm run build

Build TypeScript-based API docs using kagekiri.d.ts, inject them into the README:

npm run typedoc

Test

npm test

Debug:

npm run test:debug

Lint:

npm run lint

Fix most lint issues:

npm run lint:fix

Test bundle size:

npm run test:bundlesize