fix: 🐛 restore the check on the webhook payload (#591)

* fix: 🐛 restore the check on the webhook payload

The upstream issue has been fixed on the Hub. It's important to restore
the test, because it avoids collision between datasets and models
(models don't have a prefix in the webhook v1).

* feat: 🎸 update docker image

chart/docker-images.yaml

@@ -1,7 +1,7 @@
 {
   "dockerImage": {
     "admin": "707930574880.dkr.ecr.us-east-1.amazonaws.com/hub-datasets-server-admin:sha-49a60c5",
-    "api": "707930574880.dkr.ecr.us-east-1.amazonaws.com/hub-datasets-server-api:sha-8d9e37d",
+    "api": "707930574880.dkr.ecr.us-east-1.amazonaws.com/hub-datasets-server-api:sha-5d183a4",
     "reverseProxy": "docker.io/nginx:1.20",
     "worker": {
       "splits": "707930574880.dkr.ecr.us-east-1.amazonaws.com/hub-datasets-server-worker:sha-0dff3bf",

services/api/src/api/routes/webhook.py

@@ -39,11 +39,9 @@ def get_dataset_name(id: Optional[str]) -> Optional[str]:
     if id is None:
         return None
     dataset_name = id.removeprefix("datasets/")
-    # temporarily disabled to fix a bug with the webhook
-    # (see https://github.com/huggingface/datasets-server/issues/380#issuecomment-1254670923)
-    # if id == dataset_name:
-    #     logger.info(f"ignored because a full dataset id must starts with 'datasets/': {id}")
-    #     return None
+    if id == dataset_name:
+        logger.info(f"ignored because a full dataset id must starts with 'datasets/': {id}")
+        return None
     return dataset_name if is_non_empty_string(dataset_name) else None