Update quality tooling for formatting (#21480)

* Result of black 23.1

* Update target to Python 3.7

* Switch flake8 to ruff

* Configure isort

* Configure isort

* Apply isort with line limit

* Put the right black version

* adapt black in check copies

* Fix copies

tests/sagemaker/scripts/tensorflow/run_tf.py

@@ -10,7 +10,6 @@ from transformers import AutoTokenizer, TFAutoModelForSequenceClassification
 
 
 if __name__ == "__main__":
-
     parser = argparse.ArgumentParser()
 
     # Hyperparameters sent by the client are passed as command-line arguments to the script.

tests/sagemaker/scripts/tensorflow/run_tf_dist.py

@@ -85,7 +85,6 @@ def get_datasets(tokenizer, train_batch_size, eval_batch_size):
 
 
 if __name__ == "__main__":
-
     parser = argparse.ArgumentParser()
 
     # Hyperparameters sent by the client are passed as command-line arguments to the script.
@@ -143,7 +142,6 @@ if __name__ == "__main__":
 
     # Training
     if args.do_train:
-
         # train_results = model.fit(tf_train_dataset, epochs=args.epochs, batch_size=args.train_batch_size)
         start_train_time = time.time()
         train_results = fit(
@@ -171,7 +169,6 @@ if __name__ == "__main__":
 
     # Evaluation
     if args.do_eval and (not SDP_ENABLED or sdp.rank() == 0):
-
         result = model.evaluate(tf_test_dataset, batch_size=args.per_device_eval_batch_size, return_dict=True)
         logger.info("*** Evaluate ***")