Islandora script for triggering new derivatives (tested on isle-dc and production)

README.md

This script adds the ability to use drush to generate missing derivatives. This looks for nodes without a service file (the hint that a derivative wasn't generated) and triggers the derivative actions.

Note: that drush commands that trigger actions should also specify the site url. E.g. drush -l https://islandora.traefik.me scr /var/www/drupal/vendor/drush/drush/src/Utils/generate-image-derivatives.php. Otherwise the messages alpaca receives includes paths like "default/node//media/add/" and it doesn't know what to do with that. I should have mentioned that before.

How to run this script

• Original Script: https://gist.github.com/seth-shaw-unlv/2b6dc74b4b1b988b4224b12ceac06953
• Chat discussion: https://islandora.slack.com/archives/CM6F4C4VA/p1649085505531249

Super Fast method to run

wget https://gist.githubusercontent.com/DonRichards/cc6025eced330c325b4af96a15bf2cb4/raw/1f6f84b61d76ebf63ab9b148ae45c010f13d58ec/generate-image-derivatives.php

chmod +x generate-image-derivatives.php
chmod 755 generate-image-derivatives.php
chown nginx: generate-image-derivatives.php

mv generate-image-derivatives.php /var/www/drupal/vendor/drush/drush/src/Utils/generate-image-derivatives.php

drush -l https://islandora.traefik.me scr /var/www/drupal/vendor/drush/drush/src/Utils/generate-image-derivatives.php

Recap of the modifications made from the original script:

1. Added accessCheck(FALSE) to entity queries:

   • For taxonomy_term entity queries, the accessCheck(FALSE) method was added to explicitly disable access checking.

   $original_tid = reset(
       \Drupal::entityQuery('taxonomy_term')
         ->condition('field_external_uri', 'http://pcdm.org/use#OriginalFile')
         ->accessCheck(FALSE)
         ->execute()
   );
   
   $intermediate_tid = reset(
       \Drupal::entityQuery('taxonomy_term')
         ->condition('field_external_uri', 'http://pcdm.org/use#IntermediateFile')
         ->accessCheck(FALSE)
         ->execute()
   );
   
   $service_tid = reset(
       \Drupal::entityQuery('taxonomy_term')
         ->condition('field_external_uri', 'http://pcdm.org/use#ServiceFile')
         ->accessCheck(FALSE)
         ->execute()
   );

2. Added a check to ensure the node is loaded before calling methods on it:

   • Before executing actions on the node, added a check to ensure that the node is not null.

   foreach ($results as $nid) {
     $node = \Drupal::entityTypeManager()->getStorage('node')->load($nid);
     if ($node) {  // Check if the node is loaded successfully
       $action = (in_array($nid, $nodes_w_intermediates)) ? $intermediate_action : $original_action;
       printf("Performing '%s' on '%s' at %s\n", $action->id(), $node->toUrl()->toString(), date(DATE_ATOM));
       $action->execute([$node]);
     } else {
       printf("Node with ID '%s' could not be loaded\n", $nid);
     }
   }

These changes ensure that the script handles access checks explicitly and prevents errors when attempting to call methods on null objects.

generate-image-derivatives.php

<?php

use Drupal\system\Entity\Action;
use Drupal\Core\Session\UserSession;
use Drupal\user\Entity\User;

$userid = 1;
$account = User::load($userid);
$accountSwitcher = \Drupal::service('account_switcher');
$userSession = new UserSession([
  'uid'   => $account->id(),
  'name'  => $account->getDisplayName(),
  'roles' => $account->getRoles(),
]);
$accountSwitcher->switchTo($userSession);

$original_tid = reset(
    \Drupal::entityQuery('taxonomy_term')
      ->condition('field_external_uri', 'http://pcdm.org/use#OriginalFile')
      ->accessCheck(FALSE)
      ->execute()
  );

$intermediate_tid = reset(
    \Drupal::entityQuery('taxonomy_term')
      ->condition('field_external_uri', 'http://pcdm.org/use#IntermediateFile')
      ->accessCheck(FALSE)
      ->execute()
  );

$service_tid = reset(
    \Drupal::entityQuery('taxonomy_term')
      ->condition('field_external_uri', 'http://pcdm.org/use#ServiceFile')
      ->accessCheck(FALSE)
      ->execute()
  );

$originals_query = \Drupal::database()->select('media', 'm');
$originals_query->join('media__field_media_use', 'u', 'm.mid = u.entity_id');
$originals_query->join('media__field_media_of', 'o', 'm.mid = o.entity_id');
$originals_query->join('media__field_mime_type', 'mime', 'm.mid = mime.entity_id');
$originals_query->fields('o', ['field_media_of_target_id'])
  ->condition('u.field_media_use_target_id', $original_tid)
  ->condition('mime.field_mime_type_value', 'image/%%', 'LIKE');
$nodes_w_originals = $originals_query->execute()->fetchCol();

# We are likely to get more service files than originals because we aren't limiting
# to images. This is because we sometimes use PDF (or perhaps other) service files.
$service_query = \Drupal::database()->select('media', 'm');
$service_query->join('media__field_media_use', 'u', 'm.mid = u.entity_id');
$service_query->join('media__field_media_of', 'o', 'm.mid = o.entity_id');
$service_query->fields('o', ['field_media_of_target_id'])
  ->condition('u.field_media_use_target_id', $service_tid);
$nodes_w_service = $service_query->execute()->fetchCol();

# Similarly, we may have intermediates that aren't images.
$intermediate_query = \Drupal::database()->select('media', 'm');
$intermediate_query->join('media__field_media_use', 'u', 'm.mid = u.entity_id');
$intermediate_query->join('media__field_media_of', 'o', 'm.mid = o.entity_id');
$intermediate_query->fields('o', ['field_media_of_target_id'])
  ->condition('u.field_media_use_target_id', $intermediate_tid);
$nodes_w_intermediates = $intermediate_query->execute()->fetchCol();

$results = array_diff($nodes_w_originals, $nodes_w_service);

$original_action = Action::load('image_generate_a_service_file_from_an_original_file');
$intermediate_action = Action::load('image_generate_a_service_file_from_an_intermediate_file');

foreach ($results as $nid) {
  $node = \Drupal::entityTypeManager()->getStorage('node')->load($nid);
  if ($node) {  // Check if the node is loaded successfully
    $action = (in_array($nid, $nodes_w_intermediates)) ? $intermediate_action : $original_action;
    printf("Performing '%s' on '%s' at %s\n", $action->id(), $node->toUrl()->toString(), date(DATE_ATOM));
    $action->execute([$node]);
  } else {
    printf("Node with ID '%s' could not be loaded\n", $nid);
  }
}
$accountSwitcher->switchBack();