diff --git a/.vscode/settings.json b/.vscode/settings.json
index 02e2ab2c..27529ce4 100644
--- a/.vscode/settings.json
+++ b/.vscode/settings.json
@@ -26,7 +26,7 @@
     "black-formatter.args": [
         "--line-length",
         "88",
-        "--preview",
+        "--preview"
     ],
     "isort.args": [
         "--profile",
diff --git a/python/mlcroissant/mlcroissant/_src/core/git.py b/python/mlcroissant/mlcroissant/_src/core/git.py
index 1367baa3..cc9fa589 100644
--- a/python/mlcroissant/mlcroissant/_src/core/git.py
+++ b/python/mlcroissant/mlcroissant/_src/core/git.py
@@ -37,7 +37,9 @@ def download_git_lfs_file(file: Path):
     try:
         repo.execute(["git", "lfs", "pull", "--include", fullpath])
     except deps.git.exc.GitCommandError as ex:
-        raise RuntimeError("Problem when launching `git lfs`. "
-                           "Possible problems: Have you installed git lfs "
-                           f"locally? Is '{fullpath}' a valid `git lfs` "
-                           "repository?") from ex
+        raise RuntimeError(
+            "Problem when launching `git lfs`. "
+            "Possible problems: Have you installed git lfs "
+            f"locally? Is '{fullpath}' a valid `git lfs` "
+            "repository?"
+        ) from ex
diff --git a/python/mlcroissant/mlcroissant/_src/core/git_test.py b/python/mlcroissant/mlcroissant/_src/core/git_test.py
index c1b15b18..611b0a7f 100644
--- a/python/mlcroissant/mlcroissant/_src/core/git_test.py
+++ b/python/mlcroissant/mlcroissant/_src/core/git_test.py
@@ -11,19 +11,15 @@
 from mlcroissant._src.core.optional import deps
 from mlcroissant._src.core.path import Path
 
-_GIT_LFS_CONTENT = (
-    lambda: """version https://git-lfs.github.com/spec/v1
+_GIT_LFS_CONTENT = lambda: """version https://git-lfs.github.com/spec/v1
 oid sha256:5e2785fcd9098567a49d6e62e328923d955b307b6dcd0492f6234e96e670772a
 size 309207547
 """
-)
 
-_NON_GIT_LFS_CONTENT = (
-    lambda: """name,age
+_NON_GIT_LFS_CONTENT = lambda: """name,age
 a,1
 b,2
 c,3"""
-)
 
 _NON_ASCII_CONTENT = lambda: (255).to_bytes(1, byteorder="big")
 
@@ -54,6 +50,6 @@ def test_download_git_lfs_file():
     with mock.patch.object(git, "Git", autospec=True) as git_mock:
         download_git_lfs_file(file)
         git_mock.assert_called_once_with("/tmp/full/")
-        git_mock.return_value.execute.assert_called_once_with(
-            ["git", "lfs", "pull", "--include", "path.json"]
-        )
+        git_mock.return_value.execute.assert_called_once_with([
+            "git", "lfs", "pull", "--include", "path.json"
+        ])
diff --git a/python/mlcroissant/mlcroissant/_src/datasets_test.py b/python/mlcroissant/mlcroissant/_src/datasets_test.py
index 9017387c..a922f3e2 100644
--- a/python/mlcroissant/mlcroissant/_src/datasets_test.py
+++ b/python/mlcroissant/mlcroissant/_src/datasets_test.py
@@ -113,10 +113,16 @@ def test_hermetic_loading(dataset_name, record_set_name, num_records):
 @pytest.mark.parametrize(
     ["dataset_name", "record_set_name", "num_records"],
     [
-        ["flores-200/metadata.json",
-         "language_translations_train_data_with_metadata", 10],
-        ["flores-200/metadata.json",
-         "language_translations_test_data_with_metadata", 10],
+        [
+            "flores-200/metadata.json",
+            "language_translations_train_data_with_metadata",
+            10,
+        ],
+        [
+            "flores-200/metadata.json",
+            "language_translations_test_data_with_metadata",
+            10,
+        ],
         ["gpt-3/metadata.json", "default", 10],
         ["huggingface-c4/metadata.json", "en", 1],
         ["huggingface-mnist/metadata.json", "default", 10],
diff --git a/python/mlcroissant/mlcroissant/_src/nodes.py b/python/mlcroissant/mlcroissant/_src/nodes.py
index 12f0fd40..e567839d 100644
--- a/python/mlcroissant/mlcroissant/_src/nodes.py
+++ b/python/mlcroissant/mlcroissant/_src/nodes.py
@@ -1,4 +1,5 @@
 """Defines the public interface to the `mlcroissant.nodes` package."""
+
 from mlcroissant._src.structure_graph.nodes.field import Field
 from mlcroissant._src.structure_graph.nodes.file_object import FileObject
 from mlcroissant._src.structure_graph.nodes.file_set import FileSet
diff --git a/python/mlcroissant/mlcroissant/_src/operation_graph/graph.py b/python/mlcroissant/mlcroissant/_src/operation_graph/graph.py
index f8c16618..69d347e3 100644
--- a/python/mlcroissant/mlcroissant/_src/operation_graph/graph.py
+++ b/python/mlcroissant/mlcroissant/_src/operation_graph/graph.py
@@ -94,9 +94,9 @@ def _add_operations_for_file_object(
                 >> Concatenate(operations=operations, node=successor)
             ]
         if node.encoding_format and not should_extract(node.encoding_format):
-            fields = tuple(
-                [field for field in node.successors if isinstance(field, Field)]
-            )
+            fields = tuple([
+                field for field in node.successors if isinstance(field, Field)
+            ])
             operation >> Read(
                 operations=operations,
                 node=node,
diff --git a/python/mlcroissant/mlcroissant/_src/operation_graph/operations/concatenate.py b/python/mlcroissant/mlcroissant/_src/operation_graph/operations/concatenate.py
index 37bba6d2..cf4ea8ea 100644
--- a/python/mlcroissant/mlcroissant/_src/operation_graph/operations/concatenate.py
+++ b/python/mlcroissant/mlcroissant/_src/operation_graph/operations/concatenate.py
@@ -21,10 +21,8 @@ def __call__(self, *args: list[Path]) -> pd.DataFrame:
         """See class' docstring."""
         assert len(args) > 0, "No dataframe to merge."
         files = [file for files in args for file in files]
-        return pd.DataFrame(
-            {
-                FileProperty.filepath: [os.fspath(file.filepath) for file in files],
-                FileProperty.filename: [file.filename for file in files],
-                FileProperty.fullpath: [os.fspath(file.fullpath) for file in files],
-            }
-        )
+        return pd.DataFrame({
+            FileProperty.filepath: [os.fspath(file.filepath) for file in files],
+            FileProperty.filename: [file.filename for file in files],
+            FileProperty.fullpath: [os.fspath(file.fullpath) for file in files],
+        })
diff --git a/python/mlcroissant/mlcroissant/_src/operation_graph/operations/field.py b/python/mlcroissant/mlcroissant/_src/operation_graph/operations/field.py
index bfa5acbc..724aa91d 100644
--- a/python/mlcroissant/mlcroissant/_src/operation_graph/operations/field.py
+++ b/python/mlcroissant/mlcroissant/_src/operation_graph/operations/field.py
@@ -64,9 +64,9 @@ def _extract_lines(row: pd.Series) -> pd.Series:
     """Reads a file line-by-line and outputs a named pd.Series of the lines."""
     path = epath.Path(row[FileProperty.filepath])
     lines = path.open("rb").read().splitlines()
-    return pd.Series(
-        {**row, FileProperty.lines: lines, FileProperty.lineNumbers: range(len(lines))}
-    )
+    return pd.Series({
+        **row, FileProperty.lines: lines, FileProperty.lineNumbers: range(len(lines))
+    })
 
 
 def _extract_value(df: pd.DataFrame, field: Field) -> pd.DataFrame:
diff --git a/python/mlcroissant/mlcroissant/_src/operation_graph/operations/read.py b/python/mlcroissant/mlcroissant/_src/operation_graph/operations/read.py
index c2c9e568..914fbd40 100644
--- a/python/mlcroissant/mlcroissant/_src/operation_graph/operations/read.py
+++ b/python/mlcroissant/mlcroissant/_src/operation_graph/operations/read.py
@@ -97,11 +97,9 @@ def _read_file_content(self, encoding_format: str, file: Path) -> pd.DataFrame:
                     return parse_json_content(json_content, self.fields)
                 else:
                     # Raw files are returned as a one-line pd.DataFrame.
-                    return pd.DataFrame(
-                        {
-                            FileProperty.content: [json_content],
-                        }
-                    )
+                    return pd.DataFrame({
+                        FileProperty.content: [json_content],
+                    })
             elif encoding_format == EncodingFormat.JSON_LINES:
                 return pd.read_json(file, lines=True)
             elif encoding_format == EncodingFormat.PARQUET:
@@ -119,11 +117,9 @@ def _read_file_content(self, encoding_format: str, file: Path) -> pd.DataFrame:
                         filepath, header=None, names=[FileProperty.lines]
                     )
                 else:
-                    return pd.DataFrame(
-                        {
-                            FileProperty.content: [file.read()],
-                        }
-                    )
+                    return pd.DataFrame({
+                        FileProperty.content: [file.read()],
+                    })
             else:
                 raise ValueError(
                     f"Unsupported encoding format for file: {encoding_format}"
diff --git a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/field.py b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/field.py
index a573bddd..9945ef0f 100644
--- a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/field.py
+++ b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/field.py
@@ -41,12 +41,10 @@ def from_jsonld(cls, issues: Issues, json_) -> ParentField | None:
 
     def to_json(self) -> Json:
         """Converts the `ParentField` to JSON."""
-        return remove_empty_values(
-            {
-                "references": self.references.to_json() if self.references else None,
-                "source": self.source.to_json() if self.source else None,
-            }
-        )
+        return remove_empty_values({
+            "references": self.references.to_json() if self.references else None,
+            "source": self.source.to_json() if self.source else None,
+        })
 
 
 @dataclasses.dataclass(eq=False, repr=False)
@@ -128,20 +126,18 @@ def to_json(self) -> Json:
             self.rdf.shorten_value(data_type) for data_type in self.data_types
         ]
         parent_field = self.parent_field.to_json() if self.parent_field else None
-        return remove_empty_values(
-            {
-                "@type": "ml:Field",
-                "name": self.name,
-                "description": self.description,
-                "dataType": data_types[0] if len(data_types) == 1 else data_types,
-                "isEnumeration": self.is_enumeration,
-                "parentField": parent_field,
-                "repeated": self.repeated,
-                "references": self.references.to_json() if self.references else None,
-                "source": self.source.to_json() if self.source else None,
-                "subField": [sub_field.to_json() for sub_field in self.sub_fields],
-            }
-        )
+        return remove_empty_values({
+            "@type": "ml:Field",
+            "name": self.name,
+            "description": self.description,
+            "dataType": data_types[0] if len(data_types) == 1 else data_types,
+            "isEnumeration": self.is_enumeration,
+            "parentField": parent_field,
+            "repeated": self.repeated,
+            "references": self.references.to_json() if self.references else None,
+            "source": self.source.to_json() if self.source else None,
+            "subField": [sub_field.to_json() for sub_field in self.sub_fields],
+        })
 
     @classmethod
     def from_jsonld(
diff --git a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/file_object.py b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/file_object.py
index 1565bea2..255083e3 100644
--- a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/file_object.py
+++ b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/file_object.py
@@ -45,20 +45,18 @@ def to_json(self) -> Json:
             contained_in: str | list[str] = self.contained_in[0]
         else:
             contained_in = self.contained_in
-        return remove_empty_values(
-            {
-                "@type": "sc:FileObject",
-                "name": self.name,
-                "description": self.description,
-                "contentSize": self.content_size,
-                "contentUrl": self.content_url,
-                "containedIn": contained_in,
-                "encodingFormat": self.encoding_format,
-                "md5": self.md5,
-                "sha256": self.sha256,
-                "source": self.source.to_json() if self.source else None,
-            }
-        )
+        return remove_empty_values({
+            "@type": "sc:FileObject",
+            "name": self.name,
+            "description": self.description,
+            "contentSize": self.content_size,
+            "contentUrl": self.content_url,
+            "containedIn": contained_in,
+            "encodingFormat": self.encoding_format,
+            "md5": self.md5,
+            "sha256": self.sha256,
+            "source": self.source.to_json() if self.source else None,
+        })
 
     @classmethod
     def from_jsonld(
diff --git a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/file_object_test.py b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/file_object_test.py
index edc58412..5bc919c5 100644
--- a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/file_object_test.py
+++ b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/file_object_test.py
@@ -25,9 +25,9 @@ def test_checks_are_performed():
         Node, "assert_has_exclusive_properties"
     ) as exclusive_mock:
         create_test_node(FileObject)
-        mandatory_mock.assert_has_calls(
-            [mock.call("encoding_format", "name"), mock.call("content_url")]
-        )
+        mandatory_mock.assert_has_calls([
+            mock.call("encoding_format", "name"), mock.call("content_url")
+        ])
         exclusive_mock.assert_called_once_with(["md5", "sha256"])
         validate_name_mock.assert_called_once()
 
diff --git a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/file_set.py b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/file_set.py
index f40899af..46f8aa7c 100644
--- a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/file_set.py
+++ b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/file_set.py
@@ -37,16 +37,14 @@ def to_json(self) -> Json:
             contained_in: str | list[str] = self.contained_in[0]
         else:
             contained_in = self.contained_in
-        return remove_empty_values(
-            {
-                "@type": "sc:FileSet",
-                "name": self.name,
-                "description": self.description,
-                "containedIn": contained_in,
-                "encodingFormat": self.encoding_format,
-                "includes": self.includes,
-            }
-        )
+        return remove_empty_values({
+            "@type": "sc:FileSet",
+            "name": self.name,
+            "description": self.description,
+            "containedIn": contained_in,
+            "encodingFormat": self.encoding_format,
+            "includes": self.includes,
+        })
 
     @classmethod
     def from_jsonld(
diff --git a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/metadata.py b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/metadata.py
index 7c04f44d..02b0ce6a 100644
--- a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/metadata.py
+++ b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/metadata.py
@@ -64,19 +64,17 @@ def __post_init__(self):
 
     def to_json(self) -> Json:
         """Converts the `Metadata` to JSON."""
-        return remove_empty_values(
-            {
-                "@context": self.rdf.context,
-                "@type": "sc:Dataset",
-                "name": self.name,
-                "description": self.description,
-                "citation": self.citation,
-                "license": self.license,
-                "url": self.url,
-                "distribution": [f.to_json() for f in self.distribution],
-                "recordSet": [record_set.to_json() for record_set in self.record_sets],
-            }
-        )
+        return remove_empty_values({
+            "@context": self.rdf.context,
+            "@type": "sc:Dataset",
+            "name": self.name,
+            "description": self.description,
+            "citation": self.citation,
+            "license": self.license,
+            "url": self.url,
+            "distribution": [f.to_json() for f in self.distribution],
+            "recordSet": [record_set.to_json() for record_set in self.record_sets],
+        })
 
     @property
     def file_objects(self) -> list[FileObject]:
diff --git a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/record_set.py b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/record_set.py
index 4b7e7937..09287a67 100644
--- a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/record_set.py
+++ b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/record_set.py
@@ -68,17 +68,15 @@ def __post_init__(self):
 
     def to_json(self) -> Json:
         """Converts the `RecordSet` to JSON."""
-        return remove_empty_values(
-            {
-                "@type": "ml:RecordSet",
-                "name": self.name,
-                "description": self.description,
-                "isEnumeration": self.is_enumeration,
-                "key": self.key,
-                "field": [field.to_json() for field in self.fields],
-                "data": self.data,
-            }
-        )
+        return remove_empty_values({
+            "@type": "ml:RecordSet",
+            "name": self.name,
+            "description": self.description,
+            "isEnumeration": self.is_enumeration,
+            "key": self.key,
+            "field": [field.to_json() for field in self.fields],
+            "data": self.data,
+        })
 
     @classmethod
     def from_jsonld(
diff --git a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/source.py b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/source.py
index 0a5781fc..f7d060b4 100644
--- a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/source.py
+++ b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/source.py
@@ -70,13 +70,11 @@ class Extract:
 
     def to_json(self) -> Json:
         """Converts the `Extract` to JSON."""
-        return remove_empty_values(
-            {
-                "column": self.column,
-                "fileProperty": self.file_property.name if self.file_property else None,
-                "jsonPath": self.json_path,
-            }
-        )
+        return remove_empty_values({
+            "column": self.column,
+            "fileProperty": self.file_property.name if self.file_property else None,
+            "jsonPath": self.json_path,
+        })
 
 
 @dataclasses.dataclass(frozen=True)
@@ -100,15 +98,13 @@ class Transform:
 
     def to_json(self) -> Json:
         """Converts the `Transform` to JSON."""
-        return remove_empty_values(
-            {
-                "format": self.format,
-                "jsonPath": self.json_path,
-                "regex": self.regex,
-                "replace": self.replace,
-                "separator": self.separator,
-            }
-        )
+        return remove_empty_values({
+            "format": self.format,
+            "jsonPath": self.json_path,
+            "regex": self.regex,
+            "replace": self.replace,
+            "separator": self.separator,
+        })
 
     @classmethod
     def from_jsonld(cls, issues: Issues, jsonld: Json | list[Json]) -> list[Transform]:
@@ -196,13 +192,11 @@ def to_json(self) -> Json:
         transforms = [transform.to_json() for transform in self.transforms]
         if self.node_type is None:
             raise ValueError("node_type should be `distribution` or `field`. Got: None")
-        return remove_empty_values(
-            {
-                self.node_type: self.uid,
-                "extract": self.extract.to_json(),
-                "transform": transforms[0] if len(transforms) == 1 else transforms,
-            }
-        )
+        return remove_empty_values({
+            self.node_type: self.uid,
+            "extract": self.extract.to_json(),
+            "transform": transforms[0] if len(transforms) == 1 else transforms,
+        })
 
     @classmethod
     def from_jsonld(cls, issues: Issues, jsonld: Any) -> Source:
diff --git a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/source_test.py b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/source_test.py
index e8b607dc..1538b2a8 100644
--- a/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/source_test.py
+++ b/python/mlcroissant/mlcroissant/_src/structure_graph/nodes/source_test.py
@@ -103,29 +103,25 @@ def test_source_parses_list(json_ld, expected_source):
     [
         [
             "this is not a list",
-            set(
-                [
-                    'Transform "this is not a list" should be a dict with the keys'
-                    " http://mlcommons.org/schema/format,"
-                    " http://mlcommons.org/schema/jsonPath,"
-                    " http://mlcommons.org/schema/regex,"
-                    " http://mlcommons.org/schema/replace,"
-                    " http://mlcommons.org/schema/separator"
-                ]
-            ),
+            set([
+                'Transform "this is not a list" should be a dict with the keys'
+                " http://mlcommons.org/schema/format,"
+                " http://mlcommons.org/schema/jsonPath,"
+                " http://mlcommons.org/schema/regex,"
+                " http://mlcommons.org/schema/replace,"
+                " http://mlcommons.org/schema/separator"
+            ]),
         ],
         [
             [{"not": "the right keys"}],
-            set(
-                [
-                    "Transform \"{'not': 'the right keys'}\" should be a dict with at"
-                    " least one key in http://mlcommons.org/schema/format,"
-                    " http://mlcommons.org/schema/jsonPath,"
-                    " http://mlcommons.org/schema/regex,"
-                    " http://mlcommons.org/schema/replace,"
-                    " http://mlcommons.org/schema/separator"
-                ]
-            ),
+            set([
+                "Transform \"{'not': 'the right keys'}\" should be a dict with at"
+                " least one key in http://mlcommons.org/schema/format,"
+                " http://mlcommons.org/schema/jsonPath,"
+                " http://mlcommons.org/schema/regex,"
+                " http://mlcommons.org/schema/replace,"
+                " http://mlcommons.org/schema/separator"
+            ]),
         ],
     ],
 )
diff --git a/python/mlcroissant/mlcroissant/_src/tests/records_test.py b/python/mlcroissant/mlcroissant/_src/tests/records_test.py
index 1051b1c7..2a3df72e 100644
--- a/python/mlcroissant/mlcroissant/_src/tests/records_test.py
+++ b/python/mlcroissant/mlcroissant/_src/tests/records_test.py
@@ -6,26 +6,22 @@
 
 
 def test_record_to_python():
-    assert record_to_python(
-        {
-            "key1": 1,
-            "key2": {
-                "key3": pd.Timestamp("2017-01-01T12"),
-                "key4": {"key5": b"foo", "key6": float("nan")},
-            },
-        }
-    ) == {
+    assert record_to_python({
+        "key1": 1,
+        "key2": {
+            "key3": pd.Timestamp("2017-01-01T12"),
+            "key4": {"key5": b"foo", "key6": float("nan")},
+        },
+    }) == {
         "key1": 1,
         "key2": {"key3": "2017-01-01 12:00:00", "key4": {"key5": "foo", "key6": None}},
     }
-    assert record_to_python(
-        {
-            "image": (
-                "<PIL.PngImagePlugin.PngImageFile image mode=L size=28x28 at"
-                " 0x7F40B2EB5420>"
-            )
-        }
-    ) == {
+    assert record_to_python({
+        "image": (
+            "<PIL.PngImagePlugin.PngImageFile image mode=L size=28x28 at"
+            " 0x7F40B2EB5420>"
+        )
+    }) == {
         "image": (
             "<PIL.PngImagePlugin.PngImageFile image mode=L size=28x28 at"
             " <MEMORY_ADDRESS>>"
diff --git a/python/mlcroissant/mlcroissant/scripts/cli.py b/python/mlcroissant/mlcroissant/scripts/cli.py
index 7fc74614..afbe051d 100644
--- a/python/mlcroissant/mlcroissant/scripts/cli.py
+++ b/python/mlcroissant/mlcroissant/scripts/cli.py
@@ -14,13 +14,11 @@ class Commands:
     VALIDATE = "validate"
 
 
-choices = set(
-    [
-        Commands.FROM_HUGGINGFACE_TO_CROISSANT,
-        Commands.LOAD,
-        Commands.VALIDATE,
-    ]
-)
+choices = set([
+    Commands.FROM_HUGGINGFACE_TO_CROISSANT,
+    Commands.LOAD,
+    Commands.VALIDATE,
+])
 
 
 def main():