huggingface · lhoestq · Feb 4, 2026 · Feb 3, 2026 · Feb 3, 2026 · Feb 3, 2026
diff --git a/DEVELOPER_GUIDE.md b/DEVELOPER_GUIDE.md
@@ -305,6 +305,25 @@ We don't use the package versions (in pyproject.toml files), no need to update t
 
 All the contributions should go through a pull request. The pull requests must be "squashed" (ie: one commit per pull request).
 
+## Update common dependencies
+
+1. Go to `libs/libcommon`
+
+2. Update dependencies in `pyproject.toml`.
+
+3. Run a CLI on all projects that depend on libcommon:
+
+```bash
+>>> poetry run cli_all_projects
+Multi CLI for projects: ...
+```
+
+4. Then you can update all dependencies using
+
+```bash
+>>> poetry lock
+```
+
 ## GitHub Actions
 
 You can use [act](https://github.com/nektos/act) to test the GitHub Actions (see [.github/workflows/](.github/workflows/)) locally. It reduces the retroaction loop when working on the GitHub Actions, avoid polluting the branches with empty pushes only meant to trigger the CI, and allows to only run specific actions.

diff --git a/front/admin_ui/poetry.lock b/front/admin_ui/poetry.lock
diff --git a/jobs/cache_maintenance/poetry.lock b/jobs/cache_maintenance/poetry.lock
diff --git a/jobs/mongodb_migration/poetry.lock b/jobs/mongodb_migration/poetry.lock
diff --git a/libs/libapi/poetry.lock b/libs/libapi/poetry.lock
diff --git a/libs/libcommon/poetry.lock b/libs/libcommon/poetry.lock
diff --git a/libs/libcommon/pyproject.toml b/libs/libcommon/pyproject.toml
@@ -10,7 +10,7 @@ python = "3.12.11"
 anyio = ">=3.4.0,<5"
 appdirs = "^1.4.4"
 cryptography = "^43.0.1"
-datasets = {git="https://github.com/huggingface/datasets.git", rev="f15c59970fef04e277a93b9c262ccdc7bd8038a4", extras = ["vision"]}
+datasets = {git="https://github.com/huggingface/datasets.git", rev="6db5f805a444a767726d58d5dd1cfaa25fc852c0", extras = ["vision"]}
 duckdb = "^1.2.2"
 environs = "^14.3.0"
 filelock = "^3.18.0"
@@ -73,6 +73,9 @@ validators = "^0.22.0"
 build-backend = "poetry.core.masonry.api"
 requires = ["poetry-core>=1.0.0"]
 
+[tool.poetry.scripts]
+cli_all_projects = 'libcommon.scripts:cli_all_projects'
+
 [tool.pytest.ini_options]
 filterwarnings = ["ignore::DeprecationWarning"]
 markers = [

diff --git a/libs/libcommon/src/libcommon/scripts.py b/libs/libcommon/src/libcommon/scripts.py
@@ -0,0 +1,20 @@
+import subprocess  # nosec
+from pathlib import Path
+
+
+def cli_all_projects() -> None:
+    all_projects_dir = Path(__file__).resolve().parents[-5]
+    projects = [
+        p
+        for d in all_projects_dir.glob("*")
+        if not d.name.startswith(".") and d.is_dir()
+        for p in list(d.glob("*")) + [d]
+        if (p / "pyproject.toml").exists()
+    ]
+    projects.sort(key=lambda p: (p.name != "libcommon", "libs" not in p.parts, p))
+
+    print(f"Multi CLI for projects: {', '.join(str(project) for project in projects)}")
+    while (command := input(">>> ")) not in ["exit", "exit()"]:
+        for project in projects:
+            print(f"cd {project}")
+            subprocess.run(command.split(" "), cwd=project)  # nosec
diff --git a/services/admin/poetry.lock b/services/admin/poetry.lock
diff --git a/services/api/poetry.lock b/services/api/poetry.lock
diff --git a/services/rows/poetry.lock b/services/rows/poetry.lock
diff --git a/services/search/poetry.lock b/services/search/poetry.lock
diff --git a/services/sse-api/poetry.lock b/services/sse-api/poetry.lock
diff --git a/services/webhook/poetry.lock b/services/webhook/poetry.lock
diff --git a/services/worker/poetry.lock b/services/worker/poetry.lock
diff --git a/services/worker/tests/fixtures/hub.py b/services/worker/tests/fixtures/hub.py
@@ -566,9 +566,9 @@ def create_dataset_info_response_for_audio(dataset: str, config: str) -> Any:
         "config_name": config,
         "dataset_name": dataset_name,
         "version": {"version_str": "0.0.0", "major": 0, "minor": 0, "patch": 0},
-        "splits": {"train": {"name": "train", "num_bytes": 15448, "num_examples": 1, "dataset_name": None}},
+        "splits": {"train": {"name": "train", "num_bytes": 15414, "num_examples": 1, "dataset_name": None}},
         "download_size": AUDIO_PARQUET_SIZE,
-        "dataset_size": 15448,
+        "dataset_size": 15414,
     }
 
 
@@ -648,7 +648,7 @@ def create_parquet_and_info_response(
 
 CSV_PARQUET_SIZE = 1730
 PARTIAL_CSV_PARQUET_SIZE = 5806
-AUDIO_PARQUET_SIZE = 16546
+AUDIO_PARQUET_SIZE = 16477
 BIG_PARQUET_FILE = 18418
 
 DATA_cols = {