Avoid duplicated images when sideloading images to Woo API

Author: gabriel-detassigny

php/class-media.php

@@ -3147,7 +3147,7 @@ public function setup() {
 			// Internal components.
 			$this->global_transformations = new Global_Transformations( $this );
 			$this->gallery                = $this->plugin->get_component( 'gallery' );
-			$this->woocommerce_gallery    = new WooCommerceGallery( $this->gallery );
+			$this->woocommerce_gallery    = new WooCommerceGallery( $this->gallery, $this );
 			$this->filter                 = new Filter( $this );
 			$this->upgrade                = new Upgrade( $this );
 			$this->video                  = new Video( $this );

php/media/class-woocommercegallery.php

@@ -7,6 +7,10 @@
 
 namespace Cloudinary\Media;
 
+use Cloudinary\Media;
+use WP_REST_Request;
+use WP_REST_Response;
+
 /**
  * Class WooCommerceGallery.
  *
@@ -20,16 +24,29 @@ class WooCommerceGallery {
 	 */
 	private $gallery;
 
+	/**
+	 * The media instance.
+	 *
+	 * @var Media
+	 */
+	private $media;
+
 	/**
 	 * Init woo gallery.
 	 *
 	 * @param Gallery $gallery Gallery instance.
+	 * @param Media   $media   Media instance.
 	 */
-	public function __construct( Gallery $gallery ) {
+	public function __construct( Gallery $gallery, Media $media ) {
 		$this->gallery = $gallery;
+		$this->media   = $media;
+
+		if ( self::woocommerce_active() ) {
+			$this->setup_rest_hooks();
 
-		if ( self::woocommerce_active() && $this->enabled() ) {
-			$this->setup_hooks();
+			if ( $this->enabled() ) {
+				$this->setup_hooks();
+			}
 		}
 	}
 
@@ -86,6 +103,13 @@ public function maybe_enqueue_scripts( $can ) {
 		return $can;
 	}
 
+	/**
+	 * Setup hooks for the REST API integration.
+	 */
+	public function setup_rest_hooks() {
+		add_filter( 'rest_request_before_callbacks', array( $this, 'pre_process_product_images' ), 10, 3 );
+	}
+
 	/**
 	 * Setup hooks for the gallery.
 	 */
@@ -105,4 +129,98 @@ static function () {
 
 		add_filter( 'cloudinary_enqueue_gallery_script', array( $this, 'maybe_enqueue_scripts' ) );
 	}
+
+	/**
+	 * Pre-process product images in REST API requests to resolve Cloudinary URLs to existing
+	 * media library attachment IDs, preventing unnecessary sideloads and duplicate assets.
+	 *
+	 * @param WP_REST_Response|null $response The response object or null.
+	 * @param WP_REST_Server        $handler  The request handler.
+	 * @param WP_REST_Request       $request  The request object.
+	 *
+	 * @return WP_REST_Response|null
+	 */
+	public function pre_process_product_images( $response, $handler, $request ) {
+		$route  = $request->get_route();
+		$method = $request->get_method();
+
+		// Ignore requests to other API endpoints.
+		if (
+			false === strpos( $route, '/wc/' )
+			|| false === strpos( $route, '/products' )
+			|| ! in_array( $method, array( 'POST', 'PUT', 'PATCH' ), true )
+		) {
+			return $response;
+		}
+
+		// We only care about requests that include images.
+		$images = $request->get_param( 'images' );
+		if ( empty( $images ) || ! is_array( $images ) ) {
+			return $response;
+		}
+
+		$modified = false;
+
+		foreach ( $images as $index => $image ) {
+			// If the image ID is already passed, WooCommerce will be able to find the corresponding attachment from the Media Library.
+			if ( ! empty( $image['id'] ) ) {
+				continue;
+			}
+
+			$src = isset( $image['src'] ) ? esc_url_raw( $image['src'] ) : '';
+
+			// We only care about images with a cloudinary URL.
+			if ( ! $src || ! $this->media->is_cloudinary_url( $src ) ) {
+				continue;
+			}
+
+			$attachment_id = $this->find_attachment_by_cloudinary_url( $src );
+
+			// Apply the ID so that WooCommerce assigns the existing attachment.
+			if ( ! is_null( $attachment_id ) ) {
+				$images[ $index ]['id'] = $attachment_id;
+				$modified               = true;
+			}
+		}
+
+		if ( $modified ) {
+			$request->set_param( 'images', $images );
+		}
+
+		return $response;
+	}
+
+	/**
+	 * Find an existing media library attachment that corresponds to a Cloudinary URL.
+	 *
+	 * The URL may include transformation segments, so the lookup proceeds in three steps:
+	 * exact sync key match, bare public ID match, then base key match.
+	 *
+	 * @param string $url A Cloudinary asset URL.
+	 *
+	 * @return int|null Attachment ID, or null if not found.
+	 */
+	private function find_attachment_by_cloudinary_url( $url ) {
+		// Step 1: exact sync key — handles URLs that already exist verbatim in the library.
+		$attachment_id = $this->media->get_id_from_url( $url );
+		if ( $attachment_id ) {
+			return $attachment_id;
+		}
+
+		$public_id = $this->media->get_public_id_from_url( $url );
+		if ( ! $public_id ) {
+			return null;
+		}
+
+		// Step 2: bare public ID — matches assets uploaded from WordPress without transformations.
+		$linked = $this->media->get_linked_attachments( $public_id );
+		if ( ! empty( $linked ) ) {
+			return array_shift( $linked );
+		}
+
+		// Step 3: base key — matches assets imported from Cloudinary.
+		$base_id = $this->media->get_id_from_sync_key( 'base_' . $public_id );
+
+		return $base_id ? $base_id : null;
+	}
 }