ci(datasets:skip) Update pylint version (#3867)

datasets/flwr_datasets/federated_dataset.py

@@ -103,7 +103,7 @@ class FederatedDataset:
     >>> )
     """
 
-    # pylint: disable=too-many-instance-attributes
+    # pylint: disable=too-many-instance-attributes, too-many-arguments
     def __init__(
         self,
         *,

datasets/flwr_datasets/federated_dataset_test.py

@@ -169,7 +169,7 @@ def test_resplit_dataset_into_one(self) -> None:
         if self.test_split is None:
             return
         dataset = datasets.load_dataset(self.dataset_name)
-        dataset_length = sum([len(ds) for ds in dataset.values()])
+        dataset_length = sum(len(ds) for ds in dataset.values())
         fds = FederatedDataset(
             dataset=self.dataset_name,
             partitioners={"train": 100},
@@ -216,7 +216,7 @@ def resplit(dataset: DatasetDict) -> DatasetDict:
         )
         full = fds.load_split("full")
         dataset = datasets.load_dataset(self.dataset_name)
-        dataset_length = sum([len(ds) for ds in dataset.values()])
+        dataset_length = sum(len(ds) for ds in dataset.values())
         self.assertEqual(len(full), dataset_length)
 
     def test_use_load_dataset_kwargs(self) -> None:
@@ -245,7 +245,6 @@ class ShufflingResplittingOnArtificialDatasetTest(unittest.TestCase):
     The load_dataset method is mocked and the artificial dataset is returned.
     """
 
-    # pylint: disable=no-self-use
     def _dummy_setup(self, train_rows: int = 10, test_rows: int = 5) -> DatasetDict:
         """Create a dummy DatasetDict with train, test splits."""
         data_train = {
@@ -436,14 +435,14 @@ def test_if_the_partitions_have_unique_values(self) -> None:
 class IncorrectUsageFederatedDatasets(unittest.TestCase):
     """Test incorrect usages in FederatedDatasets."""
 
-    def test_no_partitioner_for_split(self) -> None:  # pylint: disable=R0201
+    def test_no_partitioner_for_split(self) -> None:
         """Test using load_partition with missing partitioner."""
         dataset_fds = FederatedDataset(dataset="mnist", partitioners={"train": 100})
 
         with pytest.raises(ValueError):
             dataset_fds.load_partition(0, "test")
 
-    def test_no_split_in_the_dataset(self) -> None:  # pylint: disable=R0201
+    def test_no_split_in_the_dataset(self) -> None:
         """Test using load_partition with non-existent split name."""
         dataset_fds = FederatedDataset(
             dataset="mnist", partitioners={"non-existent-split": 100}
@@ -452,15 +451,14 @@ def test_no_split_in_the_dataset(self) -> None:  # pylint: disable=R0201
         with pytest.raises(ValueError):
             dataset_fds.load_partition(0, "non-existent-split")
 
-    def test_unsupported_dataset(self) -> None:  # pylint: disable=R0201
+    def test_unsupported_dataset(self) -> None:
         """Test creating FederatedDataset for unsupported dataset."""
         with pytest.warns(UserWarning):
             FederatedDataset(dataset="food101", partitioners={"train": 100})
 
     def test_cannot_use_the_old_split_names(self) -> None:
         """Test if the initial split names can not be used."""
-        dataset = datasets.load_dataset("mnist")
-        sum([len(ds) for ds in dataset.values()])
+        datasets.load_dataset("mnist")
         fds = FederatedDataset(
             dataset="mnist",
             partitioners={"train": 100},

datasets/flwr_datasets/metrics/utils_test.py

@@ -13,7 +13,6 @@
 # limitations under the License.
 # ==============================================================================
 """Tests for metrics utils."""
-# pylint: disable=no-self-use
 
 
 import unittest

datasets/flwr_datasets/partitioner/natural_id_partitioner.py

@@ -143,7 +143,6 @@ def partition_id_to_natural_id(self) -> Dict[int, str]:
         """
         return self._partition_id_to_natural_id
 
-    # pylint: disable=R0201
     @partition_id_to_natural_id.setter
     def partition_id_to_natural_id(self, value: Dict[int, str]) -> None:
         raise AttributeError(

datasets/flwr_datasets/partitioner/natural_id_partitioner_test.py

@@ -86,7 +86,7 @@ def test_load_partition_max_partition_size(
         print(num_unique_natural_ids)
         _, partitioner = _dummy_setup(num_rows, num_unique_natural_ids)
         max_size = max(
-            [len(partitioner.load_partition(i)) for i in range(num_unique_natural_ids)]
+            len(partitioner.load_partition(i)) for i in range(num_unique_natural_ids)
         )
         self.assertEqual(max_size, math.ceil(num_rows / num_unique_natural_ids))
 

datasets/flwr_datasets/partitioner/partitioner.py

@@ -44,7 +44,7 @@ def dataset(self) -> Dataset:
     @dataset.setter
     def dataset(self, value: Dataset) -> None:
         if self._dataset is not None:
-            raise Exception(
+            raise ValueError(
                 "The dataset should be assigned only once to the partitioner."
                 "This operation might also wipe out the saved references to the "
                 "created partitions (in case the partitioning scheme needs to create "

datasets/flwr_datasets/preprocessor/merger_test.py

@@ -107,7 +107,7 @@ def test_nonexistent_split_in_strategy(self) -> None:
         ):
             merger(self.dataset_dict)
 
-    def test_duplicate_merge_split_name(self) -> None:  # pylint: disable=R0201
+    def test_duplicate_merge_split_name(self) -> None:
         """Test that the new split names are not the same."""
         strategy: Dict[str, Tuple[str, ...]] = {
             "new_train": ("train", "valid"),

datasets/flwr_datasets/utils.py

@@ -183,7 +183,7 @@ def _create_division_indices_ranges(
             ranges.append(range(start_idx, end_idx))
             start_idx = end_idx
     else:
-        TypeError(
+        raise TypeError(
             f"The type of the `division` should be dict, "
             f"tuple or list but is {type(division)} instead. "
         )

datasets/flwr_datasets/visualization/bar_plot.py

@@ -100,7 +100,7 @@ def _plot_bar(
             legend_kwargs["loc"] = "outside center right"
 
         if "bbox_to_anchor" not in legend_kwargs:
-            max_len_label_str = max([len(str(column)) for column in dataframe.columns])
+            max_len_label_str = max(len(str(column)) for column in dataframe.columns)
             shift = min(0.05 + max_len_label_str / 100, 0.15)
             legend_kwargs["bbox_to_anchor"] = (1.0 + shift, 0.5)
 

datasets/pyproject.toml

@@ -68,7 +68,7 @@ isort = "==5.13.2"
 black = { version = "==24.2.0", extras = ["jupyter"] }
 docformatter = "==1.7.5"
 mypy = "==1.4.0"
-pylint = "==2.13.9"
+pylint = "==3.0.3"
 flake8 = "==3.9.2"
 parameterized = "==0.9.0"
 pytest = "==7.1.2"
@@ -95,7 +95,7 @@ line-length = 88
 target-version = ["py38", "py39", "py310", "py311"]
 
 [tool.pylint."MESSAGES CONTROL"]
-disable = "bad-continuation,duplicate-code,too-few-public-methods,useless-import-alias"
+disable = "duplicate-code,too-few-public-methods,useless-import-alias"
 
 [tool.pytest.ini_options]
 minversion = "6.2"