ruff format new test

tests/python_package_test/test_engine.py

@@ -847,8 +847,11 @@ def test_ranking_with_position_information_with_dataset_constructor(tmp_path):
     positions_from_get = lgb_train.get_position()
     np.testing.assert_array_equal(positions_from_get, positions)
 
+
 # Appends queries and positions to the dataset file
-def append_queries_and_positions_to_file(file_dataset_in, file_query_in, positions, out_path):
+def append_queries_and_positions_to_file(
+    file_dataset_in, file_query_in, positions, out_path
+):
     queries = []
     query_id = 0
     with open(file_query_in, "r") as f:
@@ -868,7 +871,8 @@ def append_queries_and_positions_to_file(file_dataset_in, file_query_in, positio
 
 
 @pytest.mark.skipif(
-    getenv("TASK", "") == "cuda", reason="Positions in learning to rank is not supported in CUDA version yet"
+    getenv("TASK", "") == "cuda",
+    reason="Positions in learning to rank is not supported in CUDA version yet",
 )
 def test_ranking_with_position_and_group_information_in_single_file(tmp_path):
     rank_example_dir = Path(__file__).absolute().parents[2] / "examples" / "lambdarank"
@@ -882,8 +886,18 @@ def test_ranking_with_position_and_group_information_in_single_file(tmp_path):
     )
 
     # append queries amd positions to the dataset file. They will have 301 and 302 feature indexes
-    append_queries_and_positions_to_file(str(tmp_path / "rank.intermediate"), str(rank_example_dir / "rank.train.query"), positions, str(tmp_path / "rank.train"))
-    append_queries_and_positions_to_file(str(rank_example_dir / "rank.test"), str(rank_example_dir / "rank.test.query"), None, str(tmp_path / "rank.test"))
+    append_queries_and_positions_to_file(
+        str(tmp_path / "rank.intermediate"),
+        str(rank_example_dir / "rank.train.query"),
+        positions,
+        str(tmp_path / "rank.train"),
+    )
+    append_queries_and_positions_to_file(
+        str(rank_example_dir / "rank.test"),
+        str(rank_example_dir / "rank.test.query"),
+        None,
+        str(tmp_path / "rank.test"),
+    )
 
     # Training with single file
     params = {
@@ -897,12 +911,14 @@ def test_ranking_with_position_and_group_information_in_single_file(tmp_path):
         "min_sum_hessian_in_leaf": 5.0,
         "group_column": 301,
         "position_column": 302,
-        "label_column": 0
+        "label_column": 0,
     }
 
     lgb_train = lgb.Dataset(str(tmp_path / "rank.train"), params=params)
     lgb_valid = [lgb_train.create_valid(str(tmp_path / "rank.test"), params=params)]
-    gbm_unbiased_with_single_file = lgb.train(params, lgb_train, valid_sets=lgb_valid, num_boost_round=50)
+    gbm_unbiased_with_single_file = lgb.train(
+        params, lgb_train, valid_sets=lgb_valid, num_boost_round=50
+    )
 
     # Training with query files and list of positions
     params = {
@@ -915,15 +931,26 @@ def test_ranking_with_position_and_group_information_in_single_file(tmp_path):
         "min_data_in_leaf": 50,
         "min_sum_hessian_in_leaf": 5.0,
         # ignore position and group column
-        "ignore_column": "301,302"
+        "ignore_column": "301,302",
     }
-    copyfile(str(rank_example_dir / "rank.train.query"), str(tmp_path / "rank.train.query"))
-    copyfile(str(rank_example_dir / "rank.test.query"), str(tmp_path / "rank.test.query"))
-    lgb_train = lgb.Dataset(str(tmp_path / "rank.train"), params=params, position=positions)
+    copyfile(
+        str(rank_example_dir / "rank.train.query"), str(tmp_path / "rank.train.query")
+    )
+    copyfile(
+        str(rank_example_dir / "rank.test.query"), str(tmp_path / "rank.test.query")
+    )
+    lgb_train = lgb.Dataset(
+        str(tmp_path / "rank.train"), params=params, position=positions
+    )
     lgb_valid = [lgb_train.create_valid(str(tmp_path / "rank.test"))]
-    gbm_unbiased_with_multiple_files = lgb.train(params, lgb_train, valid_sets=lgb_valid, num_boost_round=50)
+    gbm_unbiased_with_multiple_files = lgb.train(
+        params, lgb_train, valid_sets=lgb_valid, num_boost_round=50
+    )
     # the performance of the unbiased LambdaMART when using query files and list of positions should match the performance of the unbiased LambdaMART when using single file with group and position columns
-    assert gbm_unbiased_with_multiple_files.best_score["valid_0"]["ndcg@3"] == gbm_unbiased_with_single_file.best_score["valid_0"]["ndcg@3"]
+    assert (
+        gbm_unbiased_with_multiple_files.best_score["valid_0"]["ndcg@3"]
+        == gbm_unbiased_with_single_file.best_score["valid_0"]["ndcg@3"]
+    )
 
 
 def test_early_stopping():