@open-wc/semantic-dom-diff

Semantic Dom Diff

Part of Open Web Components

Open Web Components provides a set of defaults, recommendations and tools to help facilitate your web component project. Our recommendations include: developing, linting, testing, building, tooling, demoing, publishing and automating.

Manual Setup

yarn add @open-wc/semantic-dom-diff --dev

semantic-dom-diff allows diffing chunks of dom or HTML for semanticaly equality:

• whitespace and newlines are normalized
• tags and attributes are printed on individual lines
• comments are removed
• style, script and svg contents are removed
• tags, attributes or element's light dom can be ignored through configuration

Chai plugin

While semantic-dom-diff can be used standalone (see below) you will most commonly use this through our chai plugin.

Registering the plugin

If you are using @open-wc/testing this is already done for you.

import { chai } from '@bundled-es-modules/chai';
import { chaiDomDiff } from '@open-wc/semantic-dom-diff';

chai.use(chaiDomDiff);

Setting up your dom for diffing

You can set up our chai plugin to diff different types of DOM:

class MyElement extends HTMLElement {
  constructor() {
    super();
    this.attachShadow({ mode: 'open' });
  }

  connectedCallback() {
    this.shadowRoot.innerHTML = '<p> shadow content </p>';
  }
}

customElements.define('my-element', MyElement);

it('my test', async () => {
  const el = await fixture(`
    <my-element>
      <div> light dom content </div>
    </my-element>
  `);

  expect(el).dom // dom is <my-element><div>light dom content</div></my-element>
  expect(el).lightDom // dom is <div>light dom content</div>
  expect(el).shadowDom // dom is <p>shadow content</p>
});

Manual diffing

You can use the chai plugin to manually diff chunks of dom. The dom is diffed semantically: whitespace, newlines etc. are normalized.

class MyElement extends HTMLElement {
  constructor() {
    super();
    this.attachShadow({ mode: 'open' });
  }

  connectedCallback() {
    this.shadowRoot.innerHTML = '<p> shadow content </p>';
  }
}

customElements.define('my-element', MyElement);

it('my test', async () => {
  const el = await fixture(`
    <my-element>
      <div> light dom content </div>
    </my-element>
  `);

  expect(el).dom.to.equal('<my-element><div>light dom content</div></my-element>');
  expect(el).lightDom.to.equal('<div>light dom content</div>');
  expect(el).shadowDom.to.equal('<p>shadow content</p>');
});

Snapshot testing

The most powerful feature of semantic-dom-diff is the ability to test and manage snapshots of your web components.

If you are not using @open-wc/testing-karma, you need to manually install karma-snapshot and karma-mocha-snapshot.

Setting up a snapshot

Snapshots are created by setting up your component in a specific state, and then calling .to.equalSnapshot(). You can use .dom, .lightDom or .shadowDom to set up the dom of your element:

import { fixture } from '@open-wc/testing';

describe('my-message', () => {
  it('renders message foo correctly', () => {
    const element = await fixture(`
      <my-message message="Foo"></my-element>
    `);

    expect(element).shadowDom.to.equalSnapshot();
  });

  it('renders message bar correctly', () => {
    const element = await fixture(`
      <my-message message="Bar"></my-element>
    `);

    expect(element).shadowDom.to.equalSnapshot();
  });

  it('renders a capitalized message correctly', () => {
    const element = await fixture(`
      <my-message message="Bar" capitalized></my-element>
    `);

    expect(element).shadowDom.to.equalSnapshot();
  });

  it('allows rendering a message from a slot', () => {
    const element = await fixture(`
      <my-message capitalized>Bar</my-element>
    `);

    expect(element).lightDom.to.equalSnapshot();
  });
});

Snapshots are stored in the __snapshots__ folder in your project, using the most top level describe as the name for your snapshots file.

Updating a snapshot

If you are not using the standard @open-wc/testing-karma configuration, see the documentation of karma-snapshot how to pass the update/prune flags.

When your tests run for the first time the snapshot files are generated. On subsequent test runs your element is compared with the stored snapshots. If the element and the snapshots differ the test fails.

If the difference was an intended change, you can update the snapshots by passing the --update-snapshots flag.

Cleaning up unused snapshots

After refactoring there might be leftover snapshot files which are unused. You can run karma with the --prune-snapshots flag to clean these up.

Ignoring tags and attributes

When working with libraries or custom elements there might be parts of the rendered dom which is random or otherwise outside of your control. In those cases, you might want to ignore certain attributes or tags entirely. This is possible by passing an options object.

it('renders correctly', async () => {
  const el = await fixture(`
    <div my-random-attribute="${Math.random()}">
      Hey
    </div>
  `);
  
  expect(el).dom.to.equal('<div>Hey</div>', {
    ignoreAttributes: ['my-random-attribute']
  });

  expect(el).dom.to.equalSnapshot({
    ignoreAttributes: ['my-random-attribute']
  });
});

Ignoring an attribute only for certain tags

Randomly generated ids are often used, throwing off your diffs. You can ignore attributes on specific tags:

it('renders correctly', async () => {
  const el = await fixture(`
    <input id="customInput${Math.random()}">
  `);

  // ignore id attributes on input elements
  expect(el).dom.to.equal('<div>Hey</div>', {
    ignoreAttributes: [
      { tags: ['input'], attributes: ['id'] }
    ]
  });

  expect(el).dom.to.equalSnapshot({
    ignoreAttributes: [
      { tags: ['input'], attributes: ['id'] }
    ]
  });
});

Ignoring tags

You can tell the diff to ignore certain tags entirely:

it('renders correctly', async () => {
  const el = await fixture(`
    <div>
      <my-custom-element></my-custom-element>
      foo
    </div>
  `);

  // ignore id attributes on input elements
  expect(el).dom.to.equal('<div>Hey</div>', {
    ignoreTags: ['my-custom-element']
  });

  expect(el).dom.to.equalSnapshot({
    ignoreTags: ['my-custom-element']
  });
});

Ignoring children

When working with web components you may find that they sometimes render to their light dom, for example to meet some accessibility requirements. We don't want to ignore the tag completely, as we would then not be able to test if we did render the tag.

We can ignore just it's light dom:

it('renders correctly', async () => {
  const el = await fixture(`
    <div>
      <my-custom-input id="myInput">
        <input id="inputRenderedInLightDom">
        Some text rendered in the light dom
      </my-custom-input>
      foo
    </div>
  `);

  // ignore id attributes on input elements
  expect(el).dom.to.equal(`
    <div>
      <my-custom-input id="myInput"></my-custom-input>
      foo
    </div>
  `, { ignoreChildren: ['my-custom-element'] });

  expect(el).dom.to.equalSnapshot({
    ignoreChildren: ['my-custom-element']
  });
});