Process Plugin to capture Scald Atom embeds in Wysiwyg Fields to Media Entity Embeds

ScaldWysiwyg.php

<?php
namespace Drupal\your_module\Plugin\migrate\process;

use \Drupal\Core\Database\Database;
use \Drupal\file\Entity\File;
use \Drupal\media\Entity\Media;
use Drupal\migrate\MigrateExecutableInterface;
use Drupal\migrate\ProcessPluginBase;
use Drupal\migrate\Row;
use Drupal\paragraphs\Entity\Paragraph;

/**
 * This processes text fields that contain scald embed codes in the wysiwyg.
 * It converts them to drupal-media embed strings
 *
 * @MigrateProcessPlugin(
 *   id = "scald_wysiwyg_replace"
 * )
 */
class ScaldWysiwyg extends ProcessPluginBase {
  /**
   * {@inheritdoc}
   */
  public function transform($value, MigrateExecutableInterface $migrate_executable, Row $row, $destination_property) {
    return $this->query($value);
  }

  /**
   * Checks for the existence of some value.
   *
   * @param mixed $value
   *   The body or text field value. 
   *
   * @return mixed|null
   *   Entity id if the queried entity exists. Otherwise NULL.
   */
  protected function query($value) {
    $sids = $this->getSIDs($value);
    $media = $file = NULL;

    foreach ($sids as $sid) {
      $media = $this->getMedia($sid);

      if ($media && $media->uuid->value) {
        //Scaled Media Embed:
        // <div class="dnd-atom-wrapper" data-scald-align="left" data-scald-context="thumbnail" data-scald-options="%7B%22hideCaption%22%3Afalse%2C%22usesSlideshow%22%3A%220%22%2C%22additionalClasses%22%3A%22%22%7D" data-scald-sid="1234" data-scald-type="image"><!-- scald embed --></div>

        // Media Library Embed:
        // <drupal-media data-entity-type="media" data-entity-uuid="0000000-0000-0000-0000-0000000000000" data-view-mode="thumbnail"></drupal-media>

        $value = str_replace('<div class="dnd-atom-wrapper"', '<drupal-media data-entity-type="media" ', $value);
        $value = str_replace('<!-- scald embed --></div>', '</drupal-media>', $value);
        $value = str_replace('data-scald-align="', 'data-align="', $value);
        $value = str_replace('data-scald-sid="' . $sid . '"', ' data-entity-uuid="' . $media->uuid->value . '"', $value);

        // Replace scald-context (old view mode) with a specific new view-mode
        $value = str_replace('data-scald-context="original_size_wysiwyg"', 'data-view-mode="wysiwyg_original"', $value);
      }
    }

    return $value;
  }

  function getSIDs($value) {
    $sids = [];
    // Find matching markup for scald ids. Match them to our media field_sid
    preg_match_all('/data-scald-sid="([0-9])*"/', $value, $matches);

    if (isset($matches[0])) {
      foreach ($matches[0] as $v) {
        if (strpos($v, 'data-scald-sid') !== false) {
          $sid = (int) str_replace(['data-scald-sid="', '"'], '', $v);
          $sids[] = $sid;
        }
      }
    }

    return $sids;
  }

  function getMedia($sid) {
    $media = $file = $fieldName = NULL;

    // See if we already imported media with matching sid.
    // We're using a field we created on migration but you could also query the migration_map_* table
    $query = \Drupal::entityQuery('media');
    $query->condition('field_sid', $sid);
    $entity_ids = $query->execute();

    if ($entity_ids && count($entity_ids) > 0) {
      $media = Media::load(array_values($entity_ids)[0]);
    }

    return $media;
  }

  function getSidData($sid) {
    // Find our media from the migrate db.
    Database::setActiveConnection('migrate');
    $migrateDB = Database::getConnection();
    $query = $migrateDB->select('scald_atoms', 's')
    ->fields('s', array(
      'sid',
      'provider',
      'type',
      'base_id',
      'language',
      'publisher',
      'actions',
      'title',
      'data',
      'created',
      'changed',
      'publisher',
      'data',
    ))
    ->fields('st', array(
      'scald_thumbnail_alt'
    ))
    ->fields('cred', array(
      'field_credit_value'
    ))
    ->fields('url', array(
      'field_photo_credit_url_url'
    ))
    ->fields('cap', array(
      'field_caption_long_value'
    ))
    ->fields('cop', array(
      'field_copyright_value'
    ))
    ->fields('fm', array(
      'uri'
    ));

    // Left Join to your own fields. Here's a few examples of what fields we are using.
    $query->leftJoin('field_data_scald_thumbnail', 'st', 'st.entity_id = s.sid');
    $query->leftJoin('field_data_field_credit', 'cred', 'cred.entity_id = s.sid');
    $query->leftJoin('field_data_field_photo_credit_url', 'url', 'url.entity_id = s.sid');
    $query->leftJoin('field_data_field_caption_long', 'cap', 'cap.entity_id = s.sid');
    $query->leftJoin('field_data_field_copyright', 'cop', 'cop.entity_id = s.sid');
    $query->leftJoin('file_managed', 'fm', 'fm.fid = s.base_id');

    $query->condition('s.sid', $sid);
    $scalds = $query->execute()->fetchAll();

    // Reset the connection to our default db.
    Database::setActiveConnection();

    return $scalds && count($scalds) > 0 ? $scalds[0] : null;
  }
}