fix: LEAP-860: have GCS generate data URLs if presign is disabled

label_studio/io_storages/gcs/models.py

@@ -87,6 +87,7 @@ def get_data(self, key):
     def generate_http_url(self, url):
         return GCS.generate_http_url(
             url=url,
+            presign=self.presign,
             google_application_credentials=self.google_application_credentials,
             google_project_id=self.google_project_id,
             presign_ttl=self.presign_ttl,

label_studio/io_storages/gcs/utils.py

@@ -176,13 +176,15 @@
     def generate_http_url(
         cls,
         url: str,
+        presign: bool,
         google_application_credentials: Union[str, dict] = None,
         google_project_id: str = None,
         presign_ttl: int = 1,
     ) -> str:
         """
         Gets gs:// like URI string and returns presigned https:// URL
         :param url: input URI
+        :param presign: Whether to generate presigned URL. If false, will generate base64 encoded data URL
         :param google_application_credentials:
         :param google_project_id:
         :param presign_ttl: Presign TTL in minutes
@@ -211,17 +213,24 @@
         if settings.GCS_CLOUD_STORAGE_FORCE_DEFAULT_CREDENTIALS:
             # google_application_credentials has higher priority,
             # use Application Default Credentials (ADC) when google_application_credentials is empty only
-            kwargs = {} if google_application_credentials else cls._get_default_credentials()
+            maybe_credentials = {} if google_application_credentials else cls._get_default_credentials()
+            maybe_client = None if google_application_credentials else cls.get_client()
         else:
-            kwargs = {}
+            maybe_credentials = {}
+            maybe_client = None
+
+        if not presign:
+            blob.reload(client=maybe_client)  # needed to know the content type
+            blob_bytes = blob.download_as_bytes(client=maybe_client)
+            return f'data:{blob.content_type};base64,{base64.b64encode(blob_bytes).decode("utf-8")}'
 
         url = blob.generate_signed_url(
             version='v4',
             # This URL is valid for 15 minutes
             expiration=timedelta(minutes=presign_ttl),
             # Allow GET requests using this URL.
             method='GET',
-            **kwargs,
+            **maybe_credentials,
         )
 
         logger.debug('Generated GCS signed url: ' + url)