Skip to content

AmazeeLabs/silverback_gutenberg

Repository files navigation

Silverback Gutenberg

Helps integrating Drupal's Gutenberg module into Silverback projects.

GraphQL directives

This module provides a set of GraphQL directives that are picked up by the amazeelabs/graphql_directives module. This allows to easily expose Gutenberg blocks through a GraphQL schema.

@resolveEdtiorBlocks

Parse the raw output of a field at a given path and expose its content as structured block data. Allows to define aggregated and ignored blocks:

  • aggregated: All subsequent blocks of these types will be merged into one core/paragraph block. In Gutenberg, standard HTML elements like lists, headings or tables are represented as separate blocks. This directive allows to merge them into one and simplify handling in the frontend.
  • ignored: Blocks of these types will be ignored. This is useful for blocks that are not relevant for the frontend, like the core/group block. The block will simply not part of the result and any children are spread where the block was.
type Page {
  title: String! @resolveProperty(path: "title.value")
  content: [Blocks!]!
    @resolveEditorBlocks(
      path: "body.value"
      aggregated: ["core/paragraph", "core/list"]
      ignored: ["core/group"]
    )
}

@resolveEditorBlockType

Retrieve the type of gutenberg block. Useful for resolving types of a block union.

union Blocks @resolveEditorBlockType = Paragraph | Heading | List

@resolveEditorBlockMarkup

Extract inner markup of a block that was provided by the user via rich HTML.

type Text @type(id: "core/paragraph") {
  content: String @resolveEditorBlockMarkup
}

@resolveEditorBlockAttribute

Retrieve a specific attribute, stored in a block.

type Figure @type(id: "custom/figure") {
  caption: String @resolveEditorBlockAttribute(key: "caption")
}

@resolveEditorBlockMedia

Resolve a media entity referenced in a block.

type Figure @type(id: "custom/figure") {
  image: Image @resolveEditorBlockMedia
}

@resolveEditorBlockChildren

Extract all child blocks of a given block.

type Columns @type(id: "custom/columns") {
  columns: [ColumnBlocks!]! @resolveEditorBlockChildren
}

LinkProcessor

The main idea is that all links added to a Gutenberg page are

  • kept in internal format (e.g. /node/123) when saved to Drupal database
  • processed to language-prefixed aliased form (e.g. /en/my-page) when
    • they are displayed in Gutenberg editor
    • they are sent out via GraphQL

This helps to

  • always display fresh path aliases
  • be sure that the language prefix is correct
  • update link URLs when translating content (e.g. /en/my-page will become /fr/ma-page automatically because it's /node/123 under the hood)
  • keep track of entity usage (TBD)

Implementation

The module does most of the things automatically. Yet there are few things developers should take care of.

First, custom Gutenberg blocks which store links in block attributes should implement hook_silverback_gutenberg_link_processor_block_attrs_alter. See silverback_gutenberg.api.php for an example.

Next, GraphQL resolvers which parse Gutenberg code should call LinkProcessor::processLinks before parsing the blocks. See DataProducer/Gutenberg.php for an example.

Validation

Custom validator plugins can be created in src/Plugin/Validation/GutenbergValidator

Field level validation

Example: to validate that an email is valid and required.

  • the block name is custom/my-block
  • the field attribute is email and the label Email
<?php

namespace Drupal\custom_gutenberg\Plugin\Validation\GutenbergValidator;

use Drupal\silverback_gutenberg\GutenbergValidation\GutenbergValidatorBase;
use Drupal\Core\StringTranslation\StringTranslationTrait;

/**
 * @GutenbergValidator(
 *   id="my_block_validator",
 *   label = @Translation("My block validator")
 * )
 */
class MyBlockValidator extends GutenbergValidatorBase {

  use StringTranslationTrait;

  /**
   * {@inheritDoc}
   */
  public function applies(array $block) {
    return $block['blockName'] === 'custom/my-block';
  }

  /**
   * {@inheritDoc}
   */
  public function validatedFields(array $block = []) {
    return [
      'email' => [
        'field_label' => $this->t('Email'),
        'rules' => ['required', 'email'],
      ],
    ];
  }

}

Block level validation

Perform custom block validation logic then return the result.

public function validateContent(array $block) {
  $isValid = TRUE;

  // Custom validation logic.
  // (...)

  if (!$isValid) {
    return [
      'is_valid' => FALSE,
      'message' => 'Message',
    ];
  }

  // Passes validation.
  return [
    'is_valid' => TRUE,
    'message' => '',
  ];
}

Cardinality validation

Backend

Uses the validateContent() method as a wrapper, with the cardinality validator trait.

use GutenbergCardinalityValidatorTrait;

Validate a given block type for inner blocks.

public function validateContent(array $block) {
  $expectedChildren = [
    [
      'blockName' => 'custom/teaser',
      'blockLabel' => $this->t('Teaser'),
      'min' => 1,
      'max' => 2,
    ],
  ];
  return $this->validateCardinality($block, $expectedChildren);
}

Validate any kind of block type for inner blocks.

public function validateContent(array $block) {
  $expectedChildren = [
    'validationType' => GutenbergCardinalityValidatorInterface::CARDINALITY_ANY,
    'min' => 0,
    'max' => 1,
  ];
  return $this->validateCardinality($block, $expectedChildren);
}

Validate a minimum with no maximum.

public function validateContent(array $block) {
  $expectedChildren = [
    [
      'blockName' => 'custom/teaser',
      'blockLabel' => $this->t('Teaser'),
      'min' => 1,
      'max' => GutenbergCardinalityValidatorInterface::CARDINALITY_UNLIMITED,
    ],
  ];
  return $this->validateCardinality($block, $expectedChildren);
}

Client side alternative

Client side cardinality validation can also be done in custom blocks with this pattern.

  • use getBlockCount
  • remove the InnerBlocks appender when the limit is reached
/* global Drupal */
import { registerBlockType } from 'wordpress__blocks';
import { InnerBlocks } from 'wordpress__block-editor';
import { useSelect } from 'wordpress__data';

// @ts-ignore
const __ = Drupal.t;

const MAX_BLOCKS: number = 1;

registerBlockType('custom/my-block', {
  title: __('My Block'),
  icon: 'location',
  category: 'layout',
  attributes: {},
  edit: (props) => {
    const { blockCount } = useSelect((select) => ({
      blockCount: select('core/block-editor').getBlockCount(props.clientId),
    }));
    return (
      <div>
        <InnerBlocks
          templateLock={false}
          renderAppender={() => {
            if (blockCount >= MAX_BLOCKS) {
              return null;
            } else {
              return <InnerBlocks.ButtonBlockAppender />;
            }
          }}
          allowedBlocks={['core/block']}
          template={[]}
        />
      </div>
    );
  },
  save: () => {
    return <InnerBlocks.Content />;
  },
});

Linkit integration

To enable the integration:

  • Enable the linkit module and create a linkit profile with gutenberg machine name

    This brings
    • Basic linkit integration
    • Improved suggestion labels (e.g. Content: Page, Media: PDF instead of page, pdf)
  • Add Silverback: prefixed matchers to the profile

    How they differ from the default linkit matchers
    • Suggestions order is done by the position of the search string in the label. For example, if you search for "best", the order will be:

      • Best in class
      • The best choice
      • Always choose best
    • Improved display of translated content. By default, linkit searches through all content translations but displays suggestions in the current language. Which can be confusing. The Silverback matchers changes this a bit. If the displayed item does not contain the prompt, a translation containing the prompt will be added in the brackets. For example, if you search for "gift" with the English UI, the suggestions will look like this:

      • Gift for a friend
      • Poison for an enemy (Gift für einen Feind)
  • To use a different profile when using the LinkControl component, add the machine name of the profile to the subtype query parameter in the component prop suggestionsQuery like below, where the custom linkit profile is called customer.

<DrupalLinkControl
  searchInputPlaceholder={__('Target page')}
  value={{
    url: props.attributes.linkUrl,
  }}
  settings={[]}
  suggestionsQuery={{
    // Use the custom linkit profile called customer.
    subtype: 'customer',
  }}
  onChange={(link) => {
    props.setAttributes({
      linkUrl: link.url,
    });
  }}
/>