Block Transforms is the API that allows a block to be transformed from and to other blocks, as well as from other entities. Existing entities that work with this API include shortcodes, files, regular expressions, and raw DOM nodes.
Transform direction: to and from
A block declares which transformations it supports via the optional transforms
key of the block configuration, whose subkeys to
and from
hold an array of available transforms for every direction. Example:
export const settings = {
title: 'My Block Title',
description: 'My block description',
/* ... */
transforms: {
from: [
/* supported from transforms */
],
to: [
/* supported to transforms */
],
},
};
Transformations types
This section goes through the existing types of transformations blocks support:
- block
- enter
- files
- prefix
- raw
- shortcode
Block
This type of transformations support both from and to directions, allowing blocks to be converted into a different one. It has a corresponding UI control within the block toolbar.
A transformation of type block
is an object that takes the following parameters:
- type (string): the value
block
. - blocks (array): a list of known block types. It also accepts the wildcard value (
"*"
), meaning that the transform is available to all block types (eg: all blocks can transform intocore/group
). - transform (function): a callback that receives the attributes and inner blocks of the block being processed. It should return a block object or an array of block objects.
- isMatch (function, optional): a callback that receives the block attributes as the first argument and the block object as the second argument and should return a boolean. Returning
false
from this function will prevent the transform from being available and displayed as an option to the user. - isMultiBlock (boolean, optional): whether the transformation can be applied when multiple blocks are selected. If true, the
transform
function’s first parameter will be an array containing each selected block’s attributes, and the second an array of each selected block’s inner blocks. False by default. - priority (number, optional): controls the priority with which a transformation is applied, where a lower value will take precedence over higher values. This behaves much like a WordPress hook. Like hooks, the default priority is
10
when not otherwise set.
Example: from Paragraph block to Heading block
To declare this transformation we add the following code into the heading block configuration, which uses the createBlock
function from the wp-blocks
package.
transforms: {
from: [
{
type: 'block',
blocks: [ 'core/paragraph' ],
transform: ( { content } ) => {
return createBlock( 'core/heading', {
content,
} );
},
},
]
},
Example: blocks that have InnerBlocks
A block with InnerBlocks can also be transformed from and to another block with InnerBlocks.
transforms: {
to: [
{
type: 'block',
blocks: [ 'some/block-with-innerblocks' ],
transform: ( attributes, innerBlocks ) => {
return createBlock(
'some/other-block-with-innerblocks',
attributes,
innerBlocks
);
},
},
],
},
Enter
This type of transformations support the from direction, allowing blocks to be created from some content introduced by the user. They’re applied in a new block line after the user has introduced some content and hit the ENTER key.
A transformation of type enter
is an object that takes the following parameters:
- type (string): the value
enter
. - regExp (RegExp): the Regular Expression to use as a matcher. If the value matches, the transformation will be applied.
- transform (function): a callback that receives an object with a
content
field containing the value that has been entered. It should return a block object or an array of block objects. - priority (number, optional): controls the priority with which a transform is applied, where a lower value will take precedence over higher values. This behaves much like a WordPress hook. Like hooks, the default priority is
10
when not otherwise set.
Example: from — to Separator block
To create a separator block when the user types the hyphen three times and then hits the ENTER key we can use the following code:
transforms = {
from: [
{
type: 'enter',
regExp: /^-{3,}$/,
transform: () => createBlock( 'core/separator' ),
},
],
};
Files
This type of transformations support the from direction, allowing blocks to be created from files dropped into the editor.
A transformation of type files
is an object that takes the following parameters:
- type (string): the value
files
. - transform (function): a callback that receives the array of files being processed. It should return a block object or an array of block objects.
- isMatch (function, optional): a callback that receives the array of files being processed and should return a boolean. Returning
false
from this function will prevent the transform from being applied. - priority (number, optional): controls the priority with which a transform is applied, where a lower value will take precedence over higher values. This behaves much like a WordPress hook. Like hooks, the default priority is
10
when not otherwise set.
Example: from file to File block
To create a File block when the user drops a file into the editor we can use the following code:
transforms: {
from: [
{
type: 'files',
isMatch: ( files ) => files.length === 1,
// By defining a lower priority than the default of 10,
// we make that the File block to be created as a fallback,
// if no other transform is found.
priority: 15,
transform: ( files ) => {
const file = files[ 0 ];
const blobURL = createBlobURL( file );
// File will be uploaded in componentDidMount()
return createBlock( 'core/file', {
href: blobURL,
fileName: file.name,
textLinkHref: blobURL,
} );
},
},
];
}
Prefix
This type of transformations support the from direction, allowing blocks to be created from some text typed by the user. They’re applied when, in a new block line, the user types some text and then adds a trailing space.
A transformation of type prefix
is an object that takes the following parameters:
- type (string): the value
prefix
. - prefix (string): the character or sequence of characters that match this transform.
- transform (function): a callback that receives the content introduced. It should return a block object or an array of block objects.
- priority (number, optional): controls the priority with which a transform is applied, where a lower value will take precedence over higher values. This behaves much like a WordPress hook. Like hooks, the default priority is
10
when not otherwise set.
Example: from text to custom block
If we want to create a custom block when the user types the question mark, we could use this code:
transforms: {
from: [
{
type: 'prefix',
prefix: '?',
transform( content ) {
return createBlock( 'my-plugin/question', {
content,
} );
},
},
];
}
Raw
This type of transformations support the from direction, allowing blocks to be created from raw HTML nodes. They’re applied when the user executes the “Convert to Blocks” action from within the block setting UI menu, as well as when some content is pasted or dropped into the editor.
A transformation of type raw
is an object that takes the following parameters:
- type (string): the value
raw
. - transform (function, optional): a callback that receives the node being processed. It should return a block object or an array of block objects.
- schema (object|function, optional): defines an HTML content model used to detect and process pasted contents. See below.
- selector (string, optional): a CSS selector string to determine whether the element matches according to the element.matches method. The transform won’t be executed if the element doesn’t match. This is a shorthand and alternative to using
isMatch
, which, if present, will take precedence. - isMatch (function, optional): a callback that receives the node being processed and should return a boolean. Returning
false
from this function will prevent the transform from being applied. - priority (number, optional): controls the priority with which a transform is applied, where a lower value will take precedence over higher values. This behaves much like a WordPress hook. Like hooks, the default priority is
10
when not otherwise set.
Example: from URLs to Embed block
If we want to create an Embed block when the user pastes some URL in the editor, we could use this code:
transforms: {
from: [
{
type: 'raw',
isMatch: ( node ) =>
node.nodeName === 'P' &&
/^\s*(https?:\/\/\S+)\s*$/i.test( node.textContent ),
transform: ( node ) => {
return createBlock( 'core/embed', {
url: node.textContent.trim(),
} );
},
},
],
}
Schemas and Content Models
When pasting content it’s possible to define a content model that will be used to validate and process pasted content. It’s often the case that HTML pasted into the editor will contain a mixture of elements that should transfer as well as elements that shouldn’t. For example, consider pasting <span class="time">12:04 pm</span>
into the editor. We want to copy 12:04 pm
and omit the <span>
and its class
attribute because those won’t carry the same meaning or structure as they originally did from where they were copied.
When writing raw
transforms you can control this by supplying a schema
which describes allowable content and which will be applied to clean up the pasted content before attempting to match with your block. The schemas are passed into cleanNodeList
from @wordpress/dom
; check there for a complete description of the schema.
schema = { span: { children: { '#text': {} } } };
Example: a custom content model
Suppose we want to match the following HTML snippet and turn it into some kind of custom post preview block.
<div data-post-id="13">
<h2>The Post Title</h2>
<p>Some <em>great</em> content.</p>
</div>
We want to tell the editor to allow the inner h2
and p
elements. We do this by supplying the following schema. In this example we’re using the function form, which accepts an argument supplying phrasingContentSchema
(as well as a boolean isPaste
indicating if the transformation operation started with pasting text). The phrasingContentSchema
is pre-defined to match HTML phrasing elements, such as <strong>
and <sup>
and <kbd>
. Anywhere we expect
a <RichText />
component is a good place to allow phrasing content otherwise we’ll lose all text formatting on conversion.
schema = ({ phrasingContentSchema }) => {
div: {
required: true,
attributes: [ 'data-post-id' ],
children: {
h2: { children: phrasingContentSchema },
p: { children: phrasingContentSchema }
}
}
}
When we successfully match this content every HTML attribute will be stripped away except for data-post-id
and if we have other arrangements of HTML inside of a given div
then it won’t match our transformer. Likewise we’d fail to match if we found an <h3>
in there instead of an <h2>
.
Schemas are most-important when wanting to match HTML snippets containing non-phrasing content, such as <details>
with a <summary>
. Without declaring the custom schema the editor will skip over these other constructions before attempting to run them through any block transforms.
Shortcode
This type of transformations support the from direction, allowing blocks to be created from shortcodes. It’s applied as part of the raw
transformation process.
A transformation of type shortcode
is an object that takes the following parameters:
- type (string): the value
shortcode
. - tag (string|array): the shortcode tag or list of shortcode aliases this transform can work with.
- transform (function, optional): a callback that receives the shortcode attributes as the first argument and the WPShortcodeMatch as the second. It should return a block object or an array of block objects. When this parameter is defined, it will take precedence over the
attributes
parameter. - attributes (object, optional): object representing where the block attributes should be sourced from, according to the attributes shape defined by the block configuration object. If a particular attribute contains a
shortcode
key, it should be a function that receives the shortcode attributes as the first arguments and the WPShortcodeMatch as second, and returns a value for the attribute that will be sourced in the block’s comment. - isMatch (function, optional): a callback that receives the shortcode attributes per the Shortcode API and should return a boolean. Returning
false
from this function will prevent the shortcode to be transformed into this block. - priority (number, optional): controls the priority with which a transform is applied, where a lower value will take precedence over higher values. This behaves much like a WordPress hook. Like hooks, the default priority is
10
when not otherwise set.
Example: from shortcode to block using transform
An existing shortcode can be transformed into its block counterpart using the transform
method.
transforms: {
from: [
{
type: 'shortcode',
tag: 'video',
transform( { named: { src } } ) {
return createBlock( 'core/video', { src } );
},
// Prevent the shortcode to be converted
// into this block when it doesn't
// have the proper ID.
isMatch( { named: { id } } ) {
return id === 'my-id';
},
},
],
},
Example: from shortcode to block using attributes
An existing shortcode can be transformed into its block counterpart using the attributes
parameters.
transforms: {
from: [
{
type: 'shortcode',
tag: 'youtube',
attributes: {
url: {
type: 'string',
source: 'attribute',
attribute: 'src',
selector: 'img',
},
align: {
type: 'string',
// The shortcode function will extract
// the shortcode atts into a value
// to be sourced in the block's comment.
shortcode: ( { named: { align = 'alignnone' } } ) => {
return align.replace( 'align', '' );
},
},
},
// Prevent the shortcode to be converted
// into this block when it doesn't
// have the proper ID.
isMatch( { named: { id } } ) {
return id === 'my-id';
},
},
]
},
ungroup blocks
Via the optional transforms
key of the block configuration, blocks can use the ungroup
subkey to define the blocks that will replace the block being processed. These new blocks will usually be a subset of the existing inner blocks, but could also include new blocks.
If a block has an ungroup
transform, it is eligible for ungrouping, without the requirement of being the default grouping block. The UI used to ungroup a block with this API is the same as the one used for the default grouping block. In order for the Ungroup button to be displayed, we must have a single grouping block selected, which also contains some inner blocks.
ungroup is a callback function that receives the attributes and inner blocks of the block being processed. It should return an array of block objects.
Example:
export const settings = {
title: 'My grouping Block Title',
description: 'My grouping block description',
/* ... */
transforms: {
ungroup: ( attributes, innerBlocks ) =>
innerBlocks.flatMap( ( innerBlock ) => innerBlock.innerBlocks ),
},
};