Fix flake8 lint errors reported by ruff - take 2 (#99798)

Replaces #99784. This PR is pure autofix.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/99798
Approved by: https://github.com/Skylion007, https://github.com/kit1980

functorch/op_analysis/gen_data.py

@@ -103,7 +103,7 @@ def gen_data(special_op_lists, analysis_name):
     with open(f"{analysis_name}", 'w') as f:
         for op in ops:
             info = [
-                op['full_name'], op['meta'], not (op['full_name'] in noncomposite_ops)
+                op['full_name'], op['meta'], op['full_name'] not in noncomposite_ops
             ] + [check(op) for check in special_op_lists]
             f.write(','.join([str(i) for i in info]) + '\n')
 

test/quantization/core/experimental/test_nonuniform_observer.py

@@ -142,7 +142,7 @@ class TestNonUniformObserver(unittest.TestCase):
         quantlevels_test_list = quantization_levels.tolist()
         negatives_contained = True
         for ele in quantlevels_test_list:
-            if not (-ele) in quantlevels_test_list:
+            if -ele not in quantlevels_test_list:
                 negatives_contained = False
         self.assertTrue(negatives_contained)
 

test/scripts/run_cuda_memcheck.py

@@ -81,7 +81,7 @@ for line in lines:
 # if 'cpu' or 'CPU' is in the name and 'cuda' or 'CUDA' is not in the name, then skip it
 def is_cpu_only(name):
     name = name.lower()
-    return ('cpu' in name) and not ('cuda' in name)
+    return ('cpu' in name) and "cuda" not in name
 
 ALL_TESTS = [x for x in ALL_TESTS if not is_cpu_only(x)]
 

test/test_jit.py

@@ -6125,7 +6125,7 @@ a")
 
         @torch.jit.script
         def test_bool_arith_not(lhs):
-            if not (lhs is None):
+            if lhs is not None:
                 return 1
             else:
                 return 2

test/test_jit_fuser_te.py

@@ -1595,7 +1595,7 @@ class TestTEFuser(JitTestCase):
                 t = torch.jit.trace(fn, (x, y))
                 t(x, y)
                 self.assertEqual(ref, t(x, y))
-                if not str(size) in skip_is_fused_check_sizes:
+                if str(size) not in skip_is_fused_check_sizes:
                     self.assertAllFused(t.graph_for(x, y))
             except Exception as e:
                 raise RuntimeError(

test/test_mps.py

@@ -576,9 +576,7 @@ def mps_ops_modifier(ops):
         'unique': None,
         'vdot': None,
         'view_as_complex': None,
-        'segment_reduce': None,
         'segment_reduce_': None,
-        '_segment_reduce_lengths': None,
         '_upsample_bilinear2d_aa': None,
         'geometric' : None,
         'geometric_': None,

tools/autograd/gen_variable_type.py

@@ -853,12 +853,12 @@ def gen_variable_type_func(
 
         if (
             fn.info is None
-            and not str(f.func.name.name) in RESET_GRAD_ACCUMULATOR
+            and str(f.func.name.name) not in RESET_GRAD_ACCUMULATOR
-            and not get_base_name(f) in DONT_REQUIRE_DERIVATIVE
+            and get_base_name(f) not in DONT_REQUIRE_DERIVATIVE
             and len(gen_differentiable_outputs(fn)) > 0
-            and not cpp.name(f.func) in DONT_ENFORCE_SAME_TENSOR_IMPL_OR_STORAGE
+            and cpp.name(f.func) not in DONT_ENFORCE_SAME_TENSOR_IMPL_OR_STORAGE
-            and not type_wrapper_name(f) in DONT_ENFORCE_STORAGE_IMPL_USE_COUNT
+            and type_wrapper_name(f) not in DONT_ENFORCE_STORAGE_IMPL_USE_COUNT
-            and not type_wrapper_name(f) in DONT_ENFORCE_TENSOR_IMPL_USE_COUNT
+            and type_wrapper_name(f) not in DONT_ENFORCE_TENSOR_IMPL_USE_COUNT
         ):
             # NOTE: [ Registering AutogradNotImplemented boxed kernel ]
             #
@@ -1340,7 +1340,7 @@ def emit_body(
         )
 
         # Check properties of outputs (enforce (2), (3))
-        if not f.func.kind() in (SchemaKind.inplace, SchemaKind.out):
+        if f.func.kind() not in (SchemaKind.inplace, SchemaKind.out):
             base_name = f.func.name.name.base  # TODO: should be str(f.func.name.name)?
             aliased_arg_name = ALL_VIEW_FUNCTIONS.get(base_name, None)
             if aliased_arg_name is not None:

torch/_inductor/codegen/cpp.py

@@ -1801,7 +1801,7 @@ class CppVecKernelChecker(CppVecKernel):
                     if name in VecCheckerProxy.bin_cmp_ops:
                         return VecCheckerProxy._bin_cmp_op(args, kwargs)
 
-                    if not (name in self.fast_vec_list):
+                    if name not in self.fast_vec_list:
                         self.disable_vec(f"op: {name}")
                     return self.simd_vec
 

torch/_tensor.py

@@ -187,7 +187,7 @@ class Tensor(torch._C._TensorBase):
             if self.grad is not None:
                 new_tensor.grad = self.grad.__deepcopy__(memo)
 
-            if not type(self) is Tensor:
+            if type(self) is not Tensor:
                 if type(new_tensor) is not type(self):
                     raise RuntimeError(
                         "Type of deepcopy result does not match the type of the source tensor. "

torch/ao/ns/_numeric_suite_fx.py

@@ -868,7 +868,7 @@ def _prepare_n_shadows_add_loggers_model(
     qconfig_mapping: QConfigMapping,
     backend_config: BackendConfig,
 ) -> torch.nn.Module:
-    """
+    r"""
     Note: this API is not recommended for wide usage, it is only
     provided for customers who need to migrate from the `add_loggers`
     API.

torch/ao/ns/fx/graph_matcher.py

@@ -139,7 +139,7 @@ class _NSGraphMatchableSubgraphsIterator:
 
     def _is_matchable(self, node: Node) -> bool:
         if node.op == 'call_function':
-            return not (node.target in self.non_matchable_functions)
+            return node.target not in self.non_matchable_functions
         elif node.op == 'call_module':
             assert isinstance(node.target, str)
             target_mod = getattr_from_fqn(self.gm, node.target)
@@ -147,7 +147,7 @@ class _NSGraphMatchableSubgraphsIterator:
                 any(isinstance(target_mod, t)  # type: ignore[arg-type]
                     for t in self.non_matchable_modules)
         elif node.op == 'call_method':
-            return not (node.target in self.non_matchable_methods)
+            return node.target not in self.non_matchable_methods
         else:
             return False
 

torch/ao/ns/fx/n_shadows_utils.py

@@ -669,7 +669,7 @@ def create_add_loggers_graph(
     qconfig_mapping: QConfigMapping,
     node_name_to_qconfig: Dict[str, QConfigAny],
 ) -> None:
-    """
+    r"""
     Given a model, a model graph partition (currently a set of matched
     subgraphs) and instructions how to transform each subgraph
     (currently quantizing it according to qconfig_mapping), modifies

torch/ao/pruning/_experimental/data_sparsifier/lightning/callbacks/_data_sparstity_utils.py

@@ -29,7 +29,7 @@ def _get_valid_name(name):
 def _log_sparsified_level(model, data_sparsifier) -> None:
     # Show the level of sparsity AFTER step:
     for name, parameter in model.named_parameters():
-        if not (type(parameter) in SUPPORTED_TYPES):
+        if type(parameter) not in SUPPORTED_TYPES:
             continue
         valid_name = _get_valid_name(name)
         mask = data_sparsifier.get_mask(name=valid_name)

torch/fx/experimental/symbolic_shapes.py

@@ -1672,7 +1672,7 @@ class DimConstraints:
         for s, congruences in reduced_congruences.items():
             for congruence in congruences:
                 # any congruence that cannot be checked becomes a dynamic constraint as well
-                if not (s in self._substitutions) or not sympy.checksol(congruence, {s: self._substitutions[s]}):
+                if s not in self._substitutions or not sympy.checksol(congruence, {s: self._substitutions[s]}):
                     self._dynamic_results.add(self._dcp.doprint(sympy.Eq(congruence, 0)))
 
     def prettify_results(self):

torch/fx/operator_schemas.py

@@ -222,7 +222,7 @@ def type_matches(signature_type : Any, argument_type : Any):
             return issubclass(argument_type.__args__[0], sig_el_type)
 
         def is_homogeneous_tuple(t):
-            if not getattr(t, '__origin__', None) in {tuple, Tuple}:
+            if getattr(t, "__origin__", None) not in {tuple, Tuple}:
                 return False
             contained = t.__args__
             if t.__args__ == ((),):  # Tuple[()].__args__ == ((),) for some reason

torch/signal/windows/windows.py

@@ -71,7 +71,7 @@ def _window_function_checks(function_name: str, M: int, dtype: torch.dtype, layo
         raise ValueError(f'{function_name} requires non-negative window length, got M={M}')
     if layout is not torch.strided:
         raise ValueError(f'{function_name} is implemented for strided tensors only, got: {layout}')
-    if not (dtype in [torch.float32, torch.float64]):
+    if dtype not in [torch.float32, torch.float64]:
         raise ValueError(f'{function_name} expects float32 or float64 dtypes, got: {dtype}')