Updated versions for black usort and ufmt and reformatted codebase (#2930)

Author: vfdev-5

.github/workflows/trigger_circle_ci.py

@@ -42,7 +42,6 @@ def assert_pipeline_created(pipeline_id, headers):
 
 
 def get_workflow_id(pipeline_id, headers):
-
     while True:
         result = requests.get(f"https://circleci.com/api/v2/pipeline/{pipeline_id}/workflow", headers=headers)
         assert_result(result, 200)
@@ -59,7 +58,6 @@ def get_workflow_id(pipeline_id, headers):
 
 
 def assert_workflows_successful(pipeline_id, headers):
-
     workflow_id = get_workflow_id(pipeline_id, headers)
 
     base_url = "https://app.circleci.com/pipelines/github/pytorch/ignite"
@@ -84,7 +82,6 @@ def assert_workflows_successful(pipeline_id, headers):
 
 
 if __name__ == "__main__":
-
     print("Trigger new pipeline on Circle-CI")
 
     if "CIRCLE_TOKEN" not in os.environ:

.pre-commit-config.yaml

@@ -15,12 +15,12 @@ repos:
         exclude_types: ["python", "jupyter", "shell", "gitignore"]
 
   - repo: https://github.com/omnilib/ufmt
-    rev: v1.3.1
+    rev: v2.1.0
     hooks:
       - id: ufmt
         additional_dependencies:
-          - black == 21.12b0
-          - usort == 1.0.1
+          - black == 23.3.0
+          - usort == 1.0.6
 
   - repo: https://github.com/pycqa/flake8
     rev: 6.0.0

docker/test_image.py

@@ -21,7 +21,6 @@
 
 
 def run_python_cmd(cmd):
-
     try_except_cmd = f"""
 import warnings
 warnings.filterwarnings("ignore")
@@ -65,7 +64,6 @@ def main():
 
 
 if __name__ == "__main__":
-
     parser = argparse.ArgumentParser("Check docker image script")
     parser.add_argument("image", type=str, help="Docker image to check")
     args = parser.parse_args()

examples/contrib/cifar10/main.py

@@ -20,7 +20,6 @@
 
 
 def training(local_rank, config):
-
     rank = idist.get_rank()
     manual_seed(config["seed"] + rank)
     device = idist.device()
@@ -205,7 +204,6 @@ def run(
         raise RuntimeError("The value of with_amp should be False if backend is xla")
 
     with idist.Parallel(backend=backend, **spawn_kwargs) as parallel:
-
         parallel.run(training, config)
 
 
@@ -283,7 +281,6 @@ def log_basic_info(logger, config):
 
 
 def create_trainer(model, optimizer, criterion, lr_scheduler, train_sampler, config, logger):
-
     device = idist.device()
 
     # Setup Ignite trainer:
@@ -299,7 +296,6 @@ def create_trainer(model, optimizer, criterion, lr_scheduler, train_sampler, con
     scaler = GradScaler(enabled=with_amp)
 
     def train_step(engine, batch):
-
         x, y = batch[0], batch[1]
 
         if x.device != device:

examples/contrib/cifar10_qat/main.py

@@ -19,7 +19,6 @@
 
 
 def training(local_rank, config):
-
     rank = idist.get_rank()
     manual_seed(config["seed"] + rank)
     device = idist.device()
@@ -189,7 +188,6 @@ def run(
     spawn_kwargs["nproc_per_node"] = nproc_per_node
 
     with idist.Parallel(backend=backend, **spawn_kwargs) as parallel:
-
         parallel.run(training, config)
 
 
@@ -267,7 +265,6 @@ def log_basic_info(logger, config):
 
 
 def create_trainer(model, optimizer, criterion, lr_scheduler, train_sampler, config, logger):
-
     device = idist.device()
 
     # Setup Ignite trainer:
@@ -283,7 +280,6 @@ def create_trainer(model, optimizer, criterion, lr_scheduler, train_sampler, con
     scaler = GradScaler(enabled=with_amp)
 
     def train_step(engine, batch):
-
         x, y = batch[0], batch[1]
 
         if x.device != device:

examples/contrib/transformers/main.py

@@ -22,7 +22,6 @@
 
 
 def training(local_rank, config):
-
     rank = idist.get_rank()
     manual_seed(config["seed"] + rank)
     device = idist.device()
@@ -33,7 +32,6 @@ def training(local_rank, config):
 
     output_path = config["output_dir"]
     if rank == 0:
-
         now = datetime.now().strftime("%Y%m%d-%H%M%S")
         folder_name = f"{config['model']}_backend-{idist.backend()}-{idist.get_world_size()}_{now}"
         output_path = Path(output_path) / folder_name
@@ -207,7 +205,6 @@ def run(
     spawn_kwargs["nproc_per_node"] = nproc_per_node
 
     with idist.Parallel(backend=backend, **spawn_kwargs) as parallel:
-
         parallel.run(training, config)
 
 
@@ -293,7 +290,6 @@ def log_basic_info(logger, config):
 
 
 def create_trainer(model, optimizer, criterion, lr_scheduler, train_sampler, config, logger):
-
     device = idist.device()
 
     # Setup Ignite trainer:
@@ -309,7 +305,6 @@ def create_trainer(model, optimizer, criterion, lr_scheduler, train_sampler, con
     scaler = GradScaler(enabled=with_amp)
 
     def train_step(engine, batch):
-
         input_batch = batch[0]
         labels = batch[1].view(-1, 1)
 

examples/fast_neural_style/neural_style.py

@@ -78,7 +78,6 @@ def train(args):
     running_avgs = OrderedDict()
 
     def step(engine, batch):
-
         x, _ = batch
         x = x.to(device)
 

examples/gan/dcgan.py

@@ -207,7 +207,6 @@ def main(
     alpha,
     output_dir,
 ):
-
     # seed
     check_manual_seed(seed)
 
@@ -243,7 +242,6 @@ def get_noise():
 
     # The main function, processing a batch of examples
     def step(engine, batch):
-
         # unpack the batch. It comes from a dataset, so we have <images, labels> pairs. Discard labels.
         real, _ = batch
         real = real.to(device)

examples/references/classification/imagenet/dataflow.py

@@ -19,7 +19,6 @@ def opencv_loader(path):
 
 
 def get_dataloader(dataset, sampler=None, shuffle=False, limit_num_samples=None, **kwargs):
-
     if limit_num_samples is not None:
         g = torch.Generator().manual_seed(limit_num_samples)
         indices = torch.randperm(len(dataset), generator=g)[:limit_num_samples]
@@ -38,7 +37,6 @@ def get_train_val_loaders(
     limit_train_num_samples: Optional[int] = None,
     limit_val_num_samples: Optional[int] = None,
 ) -> Tuple[DataLoader, DataLoader, DataLoader]:
-
     train_ds = ImageFolder(
         Path(root_path) / "train",
         transform=lambda sample: train_transforms(image=sample)["image"],

examples/references/classification/imagenet/main.py

@@ -24,7 +24,6 @@
 
 
 def training(local_rank, config, logger, with_clearml):
-
     rank = idist.get_rank()
     manual_seed(config.seed + local_rank)
 
@@ -305,7 +304,6 @@ def run_training(config_filepath, backend="nccl", with_clearml=True):
     assert config_filepath.exists(), f"File '{config_filepath.as_posix()}' is not found"
 
     with idist.Parallel(backend=backend) as parallel:
-
         logger = setup_logger(name="ImageNet Training", distributed_rank=idist.get_rank())
 
         config = ConfigObject(config_filepath)
@@ -327,7 +325,6 @@ def run_training(config_filepath, backend="nccl", with_clearml=True):
 
 
 def get_model_weights(config, logger, with_clearml):
-
     path = ""
     if with_clearml:
         from clearml import Model
@@ -352,7 +349,6 @@ def get_model_weights(config, logger, with_clearml):
 
 
 def evaluation(local_rank, config, logger, with_clearml):
-
     rank = idist.get_rank()
     device = idist.device()
     manual_seed(config.seed + local_rank)
@@ -428,5 +424,4 @@ def run_evaluation(config_filepath, backend="nccl", with_clearml=True):
 
 
 if __name__ == "__main__":
-
     fire.Fire({"training": run_training, "eval": run_evaluation})