Fix B903 lint: save memory for data classes with slots/namedtuple (#18184)

Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/18184
ghimport-source-id: 2ce860b07c58d06dc10cd7e5b97d4ef7c709a50d

Stack from [ghstack](https://github.com/ezyang/ghstack):
* **#18184 Fix B903 lint: save memory for data classes with slots/namedtuple**
* #18181 Fix B902 lint error: invalid first argument.
* #18178 Fix B006 lint errors: using mutable structure in default argument.
* #18177 Fix lstrip bug revealed by B005 lint

Signed-off-by: Edward Z. Yang <ezyang@fb.com>

Differential Revision: D14530872

fbshipit-source-id: e26cecab3a8545e7638454c28e654e7b82a3c08a

.flake8

@@ -6,5 +6,5 @@ max-line-length = 120
 ignore =
     E203,E305,E402,E501,E721,E741,F401,F403,F405,F821,F841,F999,W503,W504,C408,
     # ignores below are temporary, fix them and remove please!
-    B007,B008,B903
+    B007,B008
 exclude = docs/src,venv,third_party,caffe2,scripts,docs/caffe2,tools/amd_build/pyHIPIFY,torch/lib/include,torch/lib/tmp_install,build,torch/include

aten/src/ATen/function_wrapper.py

@@ -194,6 +194,8 @@ CALL_TEMPLATE = CodeTemplate("${cname}(${actuals})")
 class NYIError(Exception):
     """Indicates we don't support this declaration yet"""
 
+    __slots__ = ['reason']
+
     def __init__(self, reason):
         self.reason = reason
 

test/common_methods_invocations.py

@@ -2,6 +2,7 @@ import torch
 from torch._six import inf, nan, istuple
 from functools import reduce, wraps
 from operator import mul, itemgetter
+import collections
 from torch.autograd import Variable, Function, detect_anomaly
 from torch.testing import make_non_contiguous
 from common_utils import (skipIfNoLapack,
@@ -72,9 +73,7 @@ def prod_zeros(dim_size, dim_select):
     return result
 
 
-class non_differentiable(object):
+non_differentiable = collections.namedtuple('non_differentiable', ['tensor'])
-    def __init__(self, tensor):
-        self.tensor = tensor
 
 
 class dont_convert(tuple):

test/test_jit.py

@@ -13942,7 +13942,7 @@ class TestClassType(JitTestCase):
         self.assertEqual(fn(input), input)
 
     def test_get_attr(self):
-        @torch.jit.script
+        @torch.jit.script  # noqa: B903
         class FooTest:
             def __init__(self, x):
                 self.foo = x
@@ -14005,7 +14005,7 @@ class TestClassType(JitTestCase):
 
     def test_type_annotations(self):
         with self.assertRaisesRegex(RuntimeError, "expected a value of type bool"):
-            @torch.jit.script
+            @torch.jit.script  # noqa: B903
             class FooTest:
                 def __init__(self, x):
                     # type: (bool) -> None
@@ -14026,7 +14026,7 @@ class TestClassType(JitTestCase):
                         self.attr = x
 
     def test_class_type_as_param(self):
-        @torch.jit.script
+        @torch.jit.script  # noqa: B903
         class FooTest:
             def __init__(self, x):
                 self.attr = x
@@ -14094,7 +14094,7 @@ class TestClassType(JitTestCase):
         self.assertEqual(input, output)
 
     def test_save_load_with_classes_nested(self):
-        @torch.jit.script
+        @torch.jit.script  # noqa: B903
         class FooNestedTest:
             def __init__(self, y):
                 self.y = y

torch/autograd/profiler.py

@@ -3,7 +3,7 @@ import re
 import os
 import sys
 import itertools
-from collections import defaultdict
+from collections import defaultdict, namedtuple
 
 import torch
 from torch._six import FileNotFoundError
@@ -366,11 +366,7 @@ class Interval(object):
         return self.end - self.start
 
 
-class Kernel(object):
+Kernel = namedtuple('Kernel', ['name', 'device', 'interval'])
-    def __init__(self, name, device, interval):
-        self.name = name
-        self.device = device
-        self.interval = interval
 
 
 # TODO: record TID too

torch/jit/__init__.py

@@ -1564,10 +1564,7 @@ def annotate(the_type, the_value):
     return the_value
 
 
-class Attribute(object):
+Attribute = collections.namedtuple('Attribute', ['value', 'type'])
-    def __init__(self, value, the_type):
-        self.value = value
-        self.type = the_type
 
 
 if not torch._C._jit_init():