Fix trivial comparison in model checkpoint test (#3464)

We were comparing keys across the same checkpoint dict instead of ckpt_last vs ckpt_last_epoch

All other changes here are formatting

tests/callbacks/test_model_checkpoint.py

@@ -1,14 +1,13 @@
 import os
-import re
 import pickle
 import platform
+import re
 from pathlib import Path
 
 import cloudpickle
 import pytest
-import torch
-
 import tests.base.develop_utils as tutils
+import torch
 from pytorch_lightning import Trainer, seed_everything
 from pytorch_lightning.callbacks import ModelCheckpoint
 from pytorch_lightning.loggers import TensorBoardLogger
@@ -23,21 +22,31 @@ def test_model_checkpoint_with_non_string_input(tmpdir, save_top_k):
 
     checkpoint = ModelCheckpoint(filepath=None, save_top_k=save_top_k)
 
-    trainer = Trainer(default_root_dir=tmpdir, checkpoint_callback=checkpoint, overfit_batches=0.20, max_epochs=2)
+    trainer = Trainer(
+        default_root_dir=tmpdir,
+        checkpoint_callback=checkpoint,
+        overfit_batches=0.20,
+        max_epochs=2,
+    )
     trainer.fit(model)
-    assert checkpoint.dirpath == tmpdir / trainer.logger.name / "version_0" / "checkpoints"
+    assert (
+        checkpoint.dirpath == tmpdir / trainer.logger.name / "version_0" / "checkpoints"
+    )
 
 
 @pytest.mark.parametrize(
-    "logger_version,expected", [(None, "version_0"), (1, "version_1"), ("awesome", "awesome")],
+    "logger_version,expected",
+    [(None, "version_0"), (1, "version_1"), ("awesome", "awesome")],
 )
 def test_model_checkpoint_path(tmpdir, logger_version, expected):
     """Test that "version_" prefix is only added when logger's version is an integer"""
     tutils.reset_seed()
     model = EvalModelTemplate()
     logger = TensorBoardLogger(str(tmpdir), version=logger_version)
 
-    trainer = Trainer(default_root_dir=tmpdir, overfit_batches=0.2, max_epochs=2, logger=logger)
+    trainer = Trainer(
+        default_root_dir=tmpdir, overfit_batches=0.2, max_epochs=2, logger=logger
+    )
     trainer.fit(model)
 
     ckpt_version = Path(trainer.checkpoint_callback.dirpath).parent.name
@@ -79,12 +88,17 @@ def on_train_end(self, trainer, pl_module):
         )
 
 
-@pytest.mark.skipif(platform.system() == "Windows", reason="Distributed training is not supported on Windows")
+@pytest.mark.skipif(
+    platform.system() == "Windows",
+    reason="Distributed training is not supported on Windows",
+)
 def test_model_checkpoint_no_extraneous_invocations(tmpdir):
     """Test to ensure that the model callback saves the checkpoints only once in distributed mode."""
     model = EvalModelTemplate()
     num_epochs = 4
-    model_checkpoint = ModelCheckpointTestInvocations(expected_count=num_epochs, save_top_k=-1)
+    model_checkpoint = ModelCheckpointTestInvocations(
+        expected_count=num_epochs, save_top_k=-1
+    )
     trainer = Trainer(
         distributed_backend="ddp_cpu",
         num_processes=2,
@@ -102,31 +116,32 @@ def test_model_checkpoint_save_last_checkpoint_contents(tmpdir):
     seed_everything(100)
     model = EvalModelTemplate()
     num_epochs = 3
-    model_checkpoint = ModelCheckpoint(filepath=tmpdir, save_top_k=num_epochs, save_last=True)
+    model_checkpoint = ModelCheckpoint(
+        filepath=tmpdir, save_top_k=num_epochs, save_last=True
+    )
     trainer = Trainer(
-        default_root_dir=tmpdir, early_stop_callback=False, checkpoint_callback=model_checkpoint, max_epochs=num_epochs,
+        default_root_dir=tmpdir,
+        early_stop_callback=False,
+        checkpoint_callback=model_checkpoint,
+        max_epochs=num_epochs,
     )
     trainer.fit(model)
-    path_last_epoch = model_checkpoint.format_checkpoint_name(num_epochs - 1, {})  # epoch=3.ckpt
+    path_last_epoch = model_checkpoint.format_checkpoint_name(
+        num_epochs - 1, {}
+    )  # epoch=3.ckpt
     path_last = str(tmpdir / ModelCheckpoint.CHECKPOINT_NAME_LAST)  # last.ckpt
     assert path_last_epoch != path_last
     ckpt_last_epoch = torch.load(path_last_epoch)
     ckpt_last = torch.load(path_last)
 
-    trainer_keys = (
-        "epoch",
-        "global_step",
-    )
+    trainer_keys = ("epoch", "global_step")
     for key in trainer_keys:
         assert ckpt_last_epoch[key] == ckpt_last[key]
 
-    checkpoint_callback_keys = (
-        "best_model_score",
-        "best_model_path",
-    )
+    checkpoint_callback_keys = ("best_model_score", "best_model_path")
     for key in checkpoint_callback_keys:
         assert (
-            ckpt_last_epoch["callbacks"][type(model_checkpoint)][key]
+            ckpt_last["callbacks"][type(model_checkpoint)][key]
             == ckpt_last_epoch["callbacks"][type(model_checkpoint)][key]
         )
 
@@ -148,16 +163,16 @@ def test_ckpt_metric_names(tmpdir):
         progress_bar_refresh_rate=0,
         limit_train_batches=0.01,
         limit_val_batches=0.01,
-        checkpoint_callback=ModelCheckpoint(filepath=tmpdir + '/{val_loss:.2f}')
+        checkpoint_callback=ModelCheckpoint(filepath=tmpdir + "/{val_loss:.2f}"),
     )
 
     trainer.fit(model)
 
     # make sure the checkpoint we saved has the metric in the name
     ckpts = os.listdir(tmpdir)
-    ckpts = [x for x in ckpts if 'val_loss' in x]
+    ckpts = [x for x in ckpts if "val_loss" in x]
     assert len(ckpts) == 1
-    val = re.sub('[^0-9.]', '', ckpts[0])
+    val = re.sub("[^0-9.]", "", ckpts[0])
     assert len(val) > 3
 
 
@@ -179,14 +194,14 @@ def test_ckpt_metric_names_results(tmpdir):
         progress_bar_refresh_rate=0,
         limit_train_batches=0.01,
         limit_val_batches=0.01,
-        checkpoint_callback=ModelCheckpoint(filepath=tmpdir + '/{val_loss:.2f}')
+        checkpoint_callback=ModelCheckpoint(filepath=tmpdir + "/{val_loss:.2f}"),
     )
 
     trainer.fit(model)
 
     # make sure the checkpoint we saved has the metric in the name
     ckpts = os.listdir(tmpdir)
-    ckpts = [x for x in ckpts if 'val_loss' in x]
+    ckpts = [x for x in ckpts if "val_loss" in x]
     assert len(ckpts) == 1
-    val = re.sub('[^0-9.]', '', ckpts[0])
+    val = re.sub("[^0-9.]", "", ckpts[0])
     assert len(val) > 3