diff --git a/pyproject.toml b/pyproject.toml
index 977168c9789..53328fa65cc 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -13,6 +13,9 @@ select = ["C", "E", "F", "I", "W"]
 lines-after-imports = 2
 known-first-party = ["datasets"]
 
+[tool.ruff.lint.per-file-ignores]
+"__init__.py" = ["F401", "F403", "F405"]
+
 [tool.pytest.ini_options]
 # Test fails if a FutureWarning is thrown by `huggingface_hub`
 # Temporarily disabled because transformers 4.41.1 calls deprecated code from `huggingface_hub` that causes FutureWarning
diff --git a/src/datasets/__init__.py b/src/datasets/__init__.py
index f36fc4fd728..653743ec635 100644
--- a/src/datasets/__init__.py
+++ b/src/datasets/__init__.py
@@ -1,4 +1,3 @@
-# ruff: noqa
 # Copyright 2020 The HuggingFace Datasets Authors and the TensorFlow Datasets Authors.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
@@ -54,10 +53,13 @@
 from .utils import logging
 
 
-# deprecated modules
-from datasets import arrow_dataset as _arrow_dataset  # isort:skip
-from datasets import utils as _utils  # isort:skip
-from datasets.utils import download_manager as _deprecated_download_manager  # isort:skip
+# isort: split
+
+# Deprecated modules
+from . import arrow_dataset as _arrow_dataset
+from . import utils as _utils
+from .utils import download_manager as _deprecated_download_manager
+
 
 _arrow_dataset.concatenate_datasets = concatenate_datasets
 _utils.DownloadConfig = DownloadConfig
diff --git a/src/datasets/features/__init__.py b/src/datasets/features/__init__.py
index 65ae879a2bb..b3c03fbfed7 100644
--- a/src/datasets/features/__init__.py
+++ b/src/datasets/features/__init__.py
@@ -1,5 +1,3 @@
-# ruff: noqa
-
 __all__ = [
     "Audio",
     "Array2D",
diff --git a/src/datasets/filesystems/__init__.py b/src/datasets/filesystems/__init__.py
index c2753e3d380..5c48c7069ce 100644
--- a/src/datasets/filesystems/__init__.py
+++ b/src/datasets/filesystems/__init__.py
@@ -14,7 +14,7 @@
 _has_s3fs = importlib.util.find_spec("s3fs") is not None
 
 if _has_s3fs:
-    from .s3filesystem import S3FileSystem  # noqa: F401
+    from .s3filesystem import S3FileSystem
 
 COMPRESSION_FILESYSTEMS: List[compression.BaseCompressedFileFileSystem] = [
     compression.Bz2FileSystem,
diff --git a/src/datasets/formatting/__init__.py b/src/datasets/formatting/__init__.py
index 78f64cfe912..8aa21d37bd2 100644
--- a/src/datasets/formatting/__init__.py
+++ b/src/datasets/formatting/__init__.py
@@ -12,8 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-# ruff: noqa
-
 from typing import Dict, List, Optional, Type
 
 from .. import config
@@ -134,6 +132,4 @@ def get_formatter(format_type: Optional[str], **format_kwargs) -> Formatter:
     if format_type in _FORMAT_TYPES_ALIASES_UNAVAILABLE:
         raise _FORMAT_TYPES_ALIASES_UNAVAILABLE[format_type]
     else:
-        raise ValueError(
-            f"Return type should be None or selected in {list(type for type in _FORMAT_TYPES.keys() if type != None)}, but got '{format_type}'"
-        )
+        raise ValueError(f"Format type should be one of {list(_FORMAT_TYPES.keys())}, but got '{format_type}'")
diff --git a/src/datasets/packaged_modules/__init__.py b/src/datasets/packaged_modules/__init__.py
index bf1408eaf91..984dc0f03a3 100644
--- a/src/datasets/packaged_modules/__init__.py
+++ b/src/datasets/packaged_modules/__init__.py
@@ -6,13 +6,13 @@
 
 from .arrow import arrow
 from .audiofolder import audiofolder
-from .cache import cache  # noqa F401
+from .cache import cache
 from .csv import csv
 from .imagefolder import imagefolder
 from .json import json
 from .pandas import pandas
 from .parquet import parquet
-from .sql import sql  # noqa F401
+from .sql import sql
 from .text import text
 from .webdataset import webdataset
 
diff --git a/src/datasets/parallel/__init__.py b/src/datasets/parallel/__init__.py
index d8309358851..a8d527f4d09 100644
--- a/src/datasets/parallel/__init__.py
+++ b/src/datasets/parallel/__init__.py
@@ -1 +1 @@
-from .parallel import parallel_backend, parallel_map, ParallelBackendConfig  # noqa F401
+from .parallel import ParallelBackendConfig, parallel_backend, parallel_map
diff --git a/src/datasets/utils/__init__.py b/src/datasets/utils/__init__.py
index 5df74ff8cac..a1c3d293599 100644
--- a/src/datasets/utils/__init__.py
+++ b/src/datasets/utils/__init__.py
@@ -12,16 +12,14 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-# ruff: noqa
-
 from . import tqdm as _tqdm  # _tqdm is the module
+from .experimental import experimental
 from .info_utils import VerificationMode
 from .logging import disable_progress_bar, enable_progress_bar, is_progress_bar_enabled
-from .version import Version
-from .experimental import experimental
 from .tqdm import (
+    are_progress_bars_disabled,
     disable_progress_bars,
     enable_progress_bars,
-    are_progress_bars_disabled,
     tqdm,
 )
+from .version import Version
diff --git a/tests/_test_patching.py b/tests/_test_patching.py
index dae9a419ec0..78aa45a2e35 100644
--- a/tests/_test_patching.py
+++ b/tests/_test_patching.py
@@ -1,13 +1,11 @@
-# isort: skip_file
-
+# ruff: noqa: F401
 # This is the module that test_patching.py uses to test patch_submodule()
-
-import os  # noqa: F401 - this is just for tests
-import os as renamed_os  # noqa: F401 - this is just for tests
-from os import path  # noqa: F401 - this is just for tests
-from os import path as renamed_path  # noqa: F401 - this is just for tests
-from os.path import join  # noqa: F401 - this is just for tests
-from os.path import join as renamed_join  # noqa: F401 - this is just for tests
+import os
+import os as renamed_os
+from os import path
+from os import path as renamed_path
+from os.path import join
+from os.path import join as renamed_join
 
 
-open = open  # noqa we just need to have a builtin inside this module to test it properly
+open = open  # we just need to have a builtin inside this module to test it properly