Fix errors

cyyever · cyyever · commit 1e29ee5cd9c3 · 2025-10-10T14:01:14.000+08:00
Signed-off-by: Yuanyuan Chen &lt;cyyever@outlook.com&gt;
diff --git a/.github/scripts/trymerge.py b/.github/scripts/trymerge.py
@@ -1092,7 +1092,7 @@ def _comment_from_node(node: Any) -> GitHubComment:
         editor = node["editor"]
         return GitHubComment(
             body_text=node["bodyText"],
-            created_at=node["createdAt"] if "createdAt" in node else "",
+            created_at=node.get("createdAt", ""),
             author_login=node["author"]["login"],
             author_url=node["author"].get("url", None),
             author_association=node["authorAssociation"],
diff --git a/test/ao/sparsity/test_activation_sparsifier.py b/test/ao/sparsity/test_activation_sparsifier.py
@@ -55,7 +55,7 @@ def _check_constructor(self, activation_sparsifier, model, defaults, sparse_conf
 
         for key, config in sparsifier_defaults.items():
             # all the keys in combined_defaults should be present in sparsifier defaults
-            assert config == combined_defaults.get(key, None)
+            assert config == combined_defaults.get(key)
 
     def _check_register_layer(
         self, activation_sparsifier, defaults, sparse_config, layer_args_list
diff --git a/test/distributed/checkpoint/test_checkpoint.py b/test/distributed/checkpoint/test_checkpoint.py
@@ -152,7 +152,7 @@ def __init__(self, fail_conf):
         self.rank = 0 if not dist.is_initialized() else dist.get_rank()
 
     def _get_ranks(self, name):
-        return self.fail_conf[name] if name in self.fail_conf else None
+        return self.fail_conf.get(name, None)
 
     def _fail_rank(self, name):
         ranks = self._get_ranks(name)
diff --git a/test/distributed/pipelining/test_schedule.py b/test/distributed/pipelining/test_schedule.py
@@ -65,7 +65,7 @@ def __init__(self, *args, **kwargs):
         self.num_stages = kwargs.get("num_stages", 1)
         self.group_size = kwargs.get("group_size", 1)
         self.group_rank = kwargs.get("group_rank", 0)
-        self.group = kwargs.get("group", None)
+        self.group = kwargs.get("group")
 
     def _create_grad_recv_info(self, *args, **kwargs):
         return None
diff --git a/test/distributed/test_c10d_nccl.py b/test/distributed/test_c10d_nccl.py
@@ -2770,11 +2770,7 @@ def hook(work_info: torch._C._distributed_c10d.WorkInfo):
         # from rank0 to other ranks. However, this is DDP's internal implementation,
         # which is subject to change in future versions.
         self.assertTrue(num_hook_fired[OpType.BROADCAST] > 0)
-        ctor_allreduce = (
-            num_hook_fired[OpType.ALLREDUCE]
-            if OpType.ALLREDUCE in num_hook_fired
-            else 0
-        )
+        ctor_allreduce = num_hook_fired.get(OpType.ALLREDUCE, 0)
 
         x = torch.zeros(2, 1000).cuda(self.rank)
         ddp(x).sum().backward()
diff --git a/test/dynamo/test_python_autograd.py b/test/dynamo/test_python_autograd.py
@@ -82,7 +82,7 @@ def grad(L, desired_results: list[Variable]) -> list[Variable]:
     # look up dL_dentries. If a variable is never used to compute the loss,
     # we consider its gradient None, see the note below about zeros for more information.
     def gather_grad(entries: list[str]):
-        return [dL_d[entry] if entry in dL_d else None for entry in entries]
+        return [dL_d.get(entry) for entry in entries]
 
     # propagate the gradient information backward
     for entry in reversed(gradient_tape):
diff --git a/test/dynamo/test_subclasses.py b/test/dynamo/test_subclasses.py
@@ -286,7 +286,7 @@ def __tensor_flatten__(self):
     def __tensor_unflatten__(inner_tensors, metadata, outer_size, outer_stride):
         return OptionalScaledTensor(
             inner_tensors["_data"],
-            inner_tensors["_scale"] if "_scale" in inner_tensors else None,
+            inner_tensors.get("_scale", None),
             constant=metadata["_constant"],
         )
 
diff --git a/test/fx/test_fx_traceback.py b/test/fx/test_fx_traceback.py
@@ -177,9 +177,7 @@ def forward(self, x):
             for node_name_2 in node_name_to_from_node:
                 if node_name_2 in {
                     node_name_1,
-                    same_ancestor_nodes[node_name_1]
-                    if node_name_1 in same_ancestor_nodes
-                    else None,
+                    same_ancestor_nodes.get(node_name_1),
                 }:
                     self.assertEqual(
                         node_name_to_from_node[node_name_1],
diff --git a/test/inductor/test_cudagraph_trees.py b/test/inductor/test_cudagraph_trees.py
@@ -177,7 +177,7 @@ def tearDown(self):
 
         def get_manager(self, device_index=None):
             return torch._inductor.cudagraph_trees.get_container(
-                self.device_idx if not device_index else device_index
+                device_index if device_index else self.device_idx
             ).tree_manager
 
         def get_roots(self):
diff --git a/test/mobile/model_test/update_production_ops.py b/test/mobile/model_test/update_production_ops.py
@@ -17,17 +17,13 @@
     for info in model_infos:
         for op in info["root_operators"]:
             # aggregate occurance per op
-            root_operators[op] = 1 + (root_operators[op] if op in root_operators else 0)
+            root_operators[op] = 1 + (root_operators.get(op, 0))
         for op in info["traced_operators"]:
             # aggregate occurance per op
-            traced_operators[op] = 1 + (
-                traced_operators[op] if op in traced_operators else 0
-            )
+            traced_operators[op] = 1 + (traced_operators.get(op, 0))
         # merge dtypes for each kernel
         for kernal, dtypes in info["kernel_metadata"].items():
-            new_dtypes = dtypes + (
-                kernel_metadata[kernal] if kernal in kernel_metadata else []
-            )
+            new_dtypes = dtypes + (kernel_metadata.get(kernal, []))
             kernel_metadata[kernal] = list(set(new_dtypes))
 
 
diff --git a/test/test_jit.py b/test/test_jit.py
@@ -6723,7 +6723,7 @@ def test_logical_short_circuit(self):
         @torch.jit.script
         def testNoThrows(t):
             c1 = 1
-            if (False and bool(t[1])) or (True or bool(t[1])):
+            if (False and bool(t[1])) or (True or bool(t[1])):  # noqa: SIM222,SIM223
                 c1 = 0
             return c1
 
@@ -15758,7 +15758,7 @@ def test_in_for_and_comp_expr(self):
         def fn(d):
             # type: (Dict[str, int]) -> List[int]
             out = [1]
-            for i in range(d["hi"] if "hi" in d else 6):
+            for i in range(d.get("hi", 6)):
                 out.append(i)  # noqa: PERF402
             return out
 
@@ -16104,7 +16104,7 @@ class TestJitGeneratedFunctional(JitTestCase):
 S = 5
 
 def add_nn_module_test(*args, **kwargs):
-    no_grad = False if 'no_grad' not in kwargs else kwargs['no_grad']
+    no_grad = kwargs.get('no_grad', False)
 
     if 'desc' in kwargs and 'eval' in kwargs['desc']:
         # eval() is not supported, so skip these tests
diff --git a/test/test_jit_autocast.py b/test/test_jit_autocast.py
@@ -111,7 +111,7 @@ def fn(a, b, use_amp: bool):
     def test_runtime_autocast_state_expr(self):
         @torch.jit.script
         def fn(a, b):
-            with autocast(enabled=a[0][0] > 0.5):
+            with autocast(enabled=bool((a[0][0] > 0.5).item())):
                 return torch.mm(a, b)
         # runtime values for autocast enable argument are not supported
         with self.assertRaises(RuntimeError):
diff --git a/test/test_scaled_matmul_cuda.py b/test/test_scaled_matmul_cuda.py
@@ -1083,7 +1083,7 @@ def test_honor_sm_carveout(self) -> None:
                 events = sorted(events, key=lambda x: x['ts'])
                 # ROCm carveout is invisible except for kernels running slower on fewer CUs
                 no_carveout, carveout_0, carveout, no_carveout_again = [float(evt.get("dur", "0.0")) for evt in events]
-                if True or not (no_carveout < carveout and carveout_0 < carveout and no_carveout_again < carveout):
+                if True or not (no_carveout < carveout and carveout_0 < carveout and no_carveout_again < carveout):  # noqa: SIM222
                     # something went wrong, print more info to help debug flaky test
                     print("ROCm debug info for test_honor_sm_carveout")
                     print("cu_count", cu_count)
diff --git a/test/test_serialization.py b/test/test_serialization.py
@@ -4553,7 +4553,7 @@ def test_weights_only_env_variables(self, force_weights_only):
         with TemporaryFileName() as f:
             torch.save(m, f)
             try:
-                old_value = os.environ[env_var] if env_var in os.environ else None
+                old_value = os.environ.get(env_var, None)
                 os.environ[env_var] = "1"
                 # if weights_only is explicitly set, TORCH_FORCE_NO_WEIGHTS_ONLY_LOAD cannot override it
                 with self.assertRaisesRegex(pickle.UnpicklingError, "Weights only load failed"):
diff --git a/test/test_sparse_csr.py b/test/test_sparse_csr.py
@@ -4099,7 +4099,7 @@ def nc_copy(t, axes=(-1,)):
             left_alpha = make_tensor(M, dtype=dtype, device=device, low=0.5, high=high) if has_left_alpha else None
             right_alpha = make_tensor(N, dtype=dtype, device=device, low=0.5, high=high) if has_right_alpha else None
 
-            if 0 and op == "bsr_dense_addmm":
+            if 0 and op == "bsr_dense_addmm":  # noqa: SIM223
                 # Find optimal kernel parameters, the speed-up is
                 # about 10x for running this test.
                 #
diff --git a/test/test_torchfuzz_repros.py b/test/test_torchfuzz_repros.py
@@ -581,7 +581,9 @@ def foo(arg0, arg1, arg2):
             t3 = arg1  # size=(1,), stride=(1,), dtype=int64, device=cuda
             t4 = arg2  # size=(1,), stride=(1,), dtype=int64, device=cuda
             t5 = t3 + t3 + t4  # size=(1,), stride=(1,), dtype=int64, device=cuda
-            t6 = torch.exp(t5)  # size=(1,), stride=(1,), dtype=int64, device=cuda
+            t6 = torch.exp(
+                t5
+            )  # size=(1,), stride=(1,), dtype=int64, device=cuda  # noqa: F841
             t7 = torch.nn.functional.layer_norm(
                 t2, (111,)
             )  # size=(49, 112, 111), stride=(12432, 111, 1), dtype=float32, device=cuda
@@ -658,7 +660,7 @@ def test_fuzzer_issue_163894(self):
         torch.manual_seed(9)
 
         def foo(arg0):
-            var_node_1 = arg0  # size=(1, 2), stride=(2, 1), dtype=int64, device=cuda
+            var_node_1 = arg0  # size=(1, 2), stride=(2, 1), dtype=int64, device=cuda  # noqa: F841
             var_node_5 = torch.full(
                 (1, 2), -66, dtype=torch.int32
             )  # size=(1, 2), stride=(2, 1), dtype=int32, device=cuda
diff --git a/torch/_inductor/analysis/profile_analysis.py b/torch/_inductor/analysis/profile_analysis.py
@@ -416,11 +416,8 @@ def __init__(
             # pyrefly: ignore  # bad-assignment
             self.dtype = dtype
         else:
-            if dtype in _dtype_map:
-                # pyrefly: ignore  # bad-assignment
-                self.dtype = _dtype_map[dtype]
-            else:
-                self.dtype = None
+            # pyrefly: ignore  # bad-assignment
+            self.dtype = _dtype_map.get(dtype)
         self._create_devices()
 
     def convert_dtype(self, event: dict[str, Any]) -> Optional[torch.dtype]:
diff --git a/torch/nn/functional.py b/torch/nn/functional.py
@@ -6686,7 +6686,7 @@ def expand_single_value(v: _Any | list[_Any] | None) -> list[_Any]:
     # So, we need to convert None arguments for lists in python
     # explicitly into empty lists.
     def list_or_empty(l: list[_Any] | None) -> list[_Any]:
-        return [] if not l else l
+        return l if l else []
 
     def enum_list_as_int_list(l: _Any | list[_Any]) -> list[_Any]:
         if not isinstance(l, list):

Original file line number	Diff line number	Diff line change
`@@ -286,7 +286,7 @@ def __tensor_flatten__(self):`
`286`	`286`	`def __tensor_unflatten__(inner_tensors, metadata, outer_size, outer_stride):`
`287`	`287`	`return OptionalScaledTensor(`
`288`	`288`	`inner_tensors["_data"],`
`289`		`- inner_tensors["_scale"] if "_scale" in inner_tensors else None,`
	`289`	`+ inner_tensors.get("_scale", None),`
`290`	`290`	`constant=metadata["_constant"],`
`291`	`291`	`)`
`292`	`292`