From ac67e64ed258b98145dad4203676e72970604a79 Mon Sep 17 00:00:00 2001
From: Nicolas Rabault <rabault.nicolas@gmail.com>
Date: Wed, 24 Jun 2026 11:08:47 +0200
Subject: [PATCH] refactor(jobs): check dataset presence with api.repo_exists

Replace the dataset_info try/except RepositoryNotFoundError dance with a
direct api.repo_exists(repo_id, repo_type="dataset") call, dropping the
httpx/RepositoryNotFoundError test scaffolding.
---
 src/lerobot/jobs/dataset.py |  7 +------
 tests/jobs/test_dataset.py  | 15 ++-------------
 2 files changed, 3 insertions(+), 19 deletions(-)

diff --git a/src/lerobot/jobs/dataset.py b/src/lerobot/jobs/dataset.py
index 21c978f62..efd641ab4 100644
--- a/src/lerobot/jobs/dataset.py
+++ b/src/lerobot/jobs/dataset.py
@@ -24,8 +24,6 @@ from __future__ import annotations
 import os
 from pathlib import Path
 
-from huggingface_hub.errors import RepositoryNotFoundError
-
 
 def ensure_dataset_available(repo_id: str, *, api, tags: list[str] | None = None) -> None:
     """Ensure repo_id resolves on the Hub, pushing a local-only dataset privately first.
@@ -34,11 +32,8 @@ def ensure_dataset_available(repo_id: str, *, api, tags: list[str] | None = None
     dataset is left untouched). Raises RuntimeError if the dataset is neither on
     the Hub nor in the local cache.
     """
-    try:
-        api.dataset_info(repo_id)
+    if api.repo_exists(repo_id, repo_type="dataset"):
         return
-    except RepositoryNotFoundError:
-        pass
 
     cache_root = Path(os.environ.get("HF_LEROBOT_HOME", "~/.cache/huggingface/lerobot")).expanduser()
     local_present = (cache_root / repo_id / "meta" / "info.json").is_file()
diff --git a/tests/jobs/test_dataset.py b/tests/jobs/test_dataset.py
index 56cf640b2..e89cc1805 100644
--- a/tests/jobs/test_dataset.py
+++ b/tests/jobs/test_dataset.py
@@ -15,25 +15,14 @@
 import sys
 from unittest.mock import MagicMock
 
-import httpx
 import pytest
-from huggingface_hub.errors import RepositoryNotFoundError
 
 from lerobot.jobs.dataset import ensure_dataset_available
 
 
-def _repo_not_found() -> RepositoryNotFoundError:
-    req = httpx.Request("GET", "https://huggingface.co/datasets/test")
-    resp = httpx.Response(404, request=req)
-    return RepositoryNotFoundError("nope", response=resp)
-
-
 def _api_with_dataset(exists: bool):
     api = MagicMock()
-    if exists:
-        api.dataset_info.return_value = object()
-    else:
-        api.dataset_info.side_effect = _repo_not_found()
+    api.repo_exists.return_value = exists
     return api
 
 
@@ -48,7 +37,7 @@ def _make_local_cache(tmp_path, repo_id: str) -> None:
 def test_dataset_already_on_hub_is_noop():
     api = _api_with_dataset(True)
     assert ensure_dataset_available("user/ds", api=api) is None
-    api.dataset_info.assert_called_once_with("user/ds")
+    api.repo_exists.assert_called_once_with("user/ds", repo_type="dataset")
 
 
 # Branch 2: not on Hub but present locally → always push privately.