Skip to content

Commit

Permalink
Add hOCR functionality
Browse files Browse the repository at this point in the history
  • Loading branch information
@alxp @joecorall
alxp authored and joecorall committed Mar 7, 2024
1 parent 3065c87 commit 4d1a95d
Show file tree
Hide file tree
Showing 2 changed files with 105 additions and 12 deletions.
12 changes: 12 additions & 0 deletions modules/islandora_iiif/config/schema/islandora_iiif.schema.yml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -19,3 +19,15 @@ views.style.iiif_manifest:
type: sequence
sequence:
type: string
label: "Tile source field(s)"
iiif_ocr_file_field:
type: sequence
sequence:
type: string:
label: "IIIF hOCR file field"
structured_text_term:
type: string
label: "Structured text term"
search_endpoint:
type: string
label: "Search endpoint path"
105 changes: 93 additions & 12 deletions modules/islandora_iiif/src/Plugin/views/style/IIIFManifest.php
View file
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@
use Drupal\Core\Form\FormStateInterface;
use Drupal\Core\Messenger\MessengerInterface;
use Drupal\Core\Url;
use Drupal\islandora\IslandoraUtils;
use Drupal\views\Plugin\views\style\StylePluginBase;
use Drupal\views\ResultRow;
use GuzzleHttp\Client;
Expand All @@ -35,6 +36,13 @@
*/
class IIIFManifest extends StylePluginBase {

/**
* Islandora utility functions.
*
* @var \Drupal\islandora\IslandoraUtils
*/
protected $utils;

/**
* {@inheritdoc}
*/
Expand Down Expand Up @@ -111,7 +119,7 @@ class IIIFManifest extends StylePluginBase {
/**
* {@inheritdoc}
*/
public function __construct(array $configuration, $plugin_id, $plugin_definition, SerializerInterface $serializer, Request $request, ImmutableConfig $iiif_config, EntityTypeManagerInterface $entity_type_manager, FileSystemInterface $file_system, Client $http_client, MessengerInterface $messenger, ModuleHandlerInterface $moduleHandler) {
public function __construct(array $configuration, $plugin_id, $plugin_definition, SerializerInterface $serializer, Request $request, ImmutableConfig $iiif_config, EntityTypeManagerInterface $entity_type_manager, FileSystemInterface $file_system, Client $http_client, MessengerInterface $messenger, ModuleHandlerInterface $moduleHandler, IslandoraUtils $utils) {
parent::__construct($configuration, $plugin_id, $plugin_definition);

$this->serializer = $serializer;
Expand All @@ -121,6 +129,7 @@ public function __construct(array $configuration, $plugin_id, $plugin_definition
$this->fileSystem = $file_system;
$this->httpClient = $http_client;
$this->messenger = $messenger;
$this->utils = $utils;
$this->moduleHandler = $moduleHandler;
}

Expand All @@ -139,7 +148,8 @@ public static function create(ContainerInterface $container, array $configuratio
$container->get('file_system'),
$container->get('http_client'),
$container->get('messenger'),
$container->get('module_handler')
$container->get('module_handler'),
$container->get('islandora.utils')
);
}

Expand Down Expand Up @@ -188,6 +198,11 @@ public function render() {
$label = $this->t("IIIF Manifest");
}

/**
* @var \Drupal\taxonomy\TermInterface|null
*/
$structured_text_term = $this->utils->getTermForUri($this->options['structured_text_term_uri']);

// @see https://iiif.io/api/presentation/2.1/#manifest
$json += [
'@type' => 'sc:Manifest',
Expand All @@ -207,7 +222,7 @@ public function render() {
// For each row in the View result.
foreach ($this->view->result as $row) {
// Add the IIIF URL to the image to print out as JSON.
$canvases = $this->getTileSourceFromRow($row, $iiif_address, $iiif_base_id);
$canvases = $this->getTileSourceFromRow($row, $iiif_address, $iiif_base_id, $structured_text_term);
foreach ($canvases as $tile_source) {
$json['sequences'][0]['canvases'][] = $tile_source;
}
Expand All @@ -217,6 +232,9 @@ public function render() {

$content_type = 'json';

// Add a search endpoint if one is defined
$this->addSearchEndpoint($json, $url_components);

// Give other modules a chance to alter the manifest.
$this->moduleHandler->alter('islandora_iiif_manifest', $json, $this);

Expand All @@ -237,7 +255,7 @@ public function render() {
* @return array
* List of IIIF URLs to display in the Openseadragon viewer.
*/
protected function getTileSourceFromRow(ResultRow $row, $iiif_address, $iiif_base_id) {
protected function getTileSourceFromRow(ResultRow $row, $iiif_address, $iiif_base_id, $structured_text_term) {
$canvases = [];
foreach (array_filter(array_values($this->options['iiif_tile_field'])) as $iiif_tile_field) {
$viewsField = $this->view->field[$iiif_tile_field];
Expand Down Expand Up @@ -300,7 +318,7 @@ protected function getTileSourceFromRow(ResultRow $row, $iiif_address, $iiif_bas
],
];

if ($ocr_url = $this->getOcrUrl($entity, $row, $i)) {
if ($ocr_url = $this->getOcrUrl($entity, $structured_text_term)) {
$tmp_canvas['seeAlso'] = [
'@id' => $ocr_url,
'format' => 'text/vnd.hocr+html',
Expand Down Expand Up @@ -389,19 +407,29 @@ protected function getCanvasDimensions(string $iiif_url, FieldItemInterface $ima
* The absolute URL of the current row's structured text,
* or FALSE if none.
*/
protected function getOcrUrl(EntityInterface $entity, ResultRow $row, $delta) {
protected function getOcrUrl(EntityInterface $entity, $structured_text_term) {
$ocr_url = FALSE;
$iiif_ocr_file_field = !empty($this->options['iiif_ocr_file_field']) ? array_filter(array_values($this->options['iiif_ocr_file_field'])) : [];
$ocrField = count($iiif_ocr_file_field) > 0 ? $this->view->field[$iiif_ocr_file_field[0]] : NULL;
if ($ocrField) {
$ocr_entity = $ocrField->getEntity($row);
$ocr_entity = $entity;
$ocr_field_name = $ocrField->definition['field_name'];
if (!is_null($ocr_field_name)) {
$ocrs = $ocr_entity->{$ocr_field_name};
$ocr = isset($ocrs[$delta]) ? $ocrs[$delta] : FALSE;
if ($ocr) {
$ocr_url = $ocr->entity->createFileUrl(FALSE);
}
$ocr = $ocrs[0] ?? FALSE;
$ocr_url = $ocr->entity->createFileUrl(FALSE);
}
}
elseif ($structured_text_term) {
$parent_node = $this->utils->getParentNode($entity);
$ocr_entity_array = $this->utils->getMediaReferencingNodeAndTerm($parent_node, $structured_text_term);
$ocr_entity_id = is_array($ocr_entity_array) ? array_shift($ocr_entity_array) : NULL;
$ocr_entity = $ocr_entity_id ? $this->entityTypeManager->getStorage('media')->load($ocr_entity_id) : NULL;
if ($ocr_entity) {
$ocr_file_source = $ocr_entity->getSource();
$ocr_fid = $ocr_file_source->getSourceFieldValue($ocr_entity);
$ocr_file = $this->entityTypeManager->getStorage('file')->load($ocr_fid);
$ocr_url = $ocr_file->createFileUrl(FALSE);
}
}

Expand Down Expand Up @@ -448,6 +476,21 @@ protected function defineOptions() {
return $options;
}

protected function addSearchEndpoint(array &$json, array $url_components) {
$url_base = $this->getRequest()->getSchemeAndHttpHost();
$hocr_search_path = $this->options['search_endpoint'];
$hocr_search_url = $url_base . '/' . ltrim($hocr_search_path, '/');

$hocr_search_url = str_replace('%node', $url_components[1], $hocr_search_url);

$json['service'][] = [
"@context" => "http://iiif.io/api/search/0/context.json",
"@id" => $hocr_search_url,
"profile" => "http://iiif.io/api/search/0/search",
"label" => t("Search inside this work"),
];
}

/**
* {@inheritdoc}
*/
Expand Down Expand Up @@ -504,10 +547,27 @@ public function buildOptionsForm(&$form, FormStateInterface $form_state) {
'#title' => $this->t('Structured OCR data file field'),
'#type' => 'checkboxes',
'#default_value' => $this->options['iiif_ocr_file_field'],
'#description' => $this->t('The source of structured OCR text for each entity.'),
'#description' => $this->t('The source of structured OCR text for each entity. If the term setting below is left blank, it will be the same entity as the source image'),
'#options' => $field_options,
'#required' => FALSE,
];

$form['structured_text_term'] = [
'#type' => 'entity_autocomplete',
'#target_type' => 'taxonomy_term',
'#title' => $this->t('Structured OCR text term'),
'#default_value' => $this->utils->getTermForUri($this->options['structured_text_term_uri']),
'#required' => FALSE,
'#description' => $this->t('Term indicating the media that holds structured text, such as hOCR, for the given object. Use this if the text is on a separate media from the tile source.'),
];

$form['search_endpoint'] = [
'#type' => 'textfield',
'#title' => $this->t("Search endpoint path."),
'#description' => $this->t("If there is a search endpoint to search within the book that returns IIIF annotations, put it here. Use substitutions %node and %keywords.<br>E.g., paged-content-search/%node?search-in-pages=%keywords"),
'#default_value' => $this->options['search_endpoint'],
'#required' => FALSE,
];
}

/**
Expand All @@ -520,4 +580,25 @@ public function getFormats() {
return ['json' => 'json'];
}

/**
* Submit handler for options form.
*
* Used to store the structured text media term by URL instead of Ttid.
*
* @param array $form
* The form.
* @param \Drupal\Core\Form\FormStateInterface $form_state
* The form state object.
*/
// @codingStandardsIgnoreStart
public function submitOptionsForm(&$form, FormStateInterface $form_state) {
// @codingStandardsIgnoreEnd
$style_options = $form_state->getValue('style_options');
$tid = $style_options['structured_text_term'];
$term = $this->entityTypeManager->getStorage('taxonomy_term')->load($tid);
$style_options['structured_text_term_uri'] = $this->utils->getUriForTerm($term);
$form_state->setValue('style_options', $style_options);
parent::submitOptionsForm($form, $form_state);
}

}

0 comments on commit 4d1a95d

Please sign in to comment.