[Code Format] Remove python version and ignore of F403 & E722

.pre-commit-config.yaml

@@ -6,25 +6,20 @@ repos:
   - id: end-of-file-fixer
   - id: trailing-whitespace
   - id: flake8
-    language_version: python3.11
-    args: ["--ignore=F405,E731,F403,W503,E722,E203", --max-line-length=120]
+    args: ["--ignore=F405,E731,W503,E203", --max-line-length=120]
     # F405 : Name may be undefined, or defined from star imports: module
     # E731 : Do not assign a lambda expression, use a def
-    # F403 : 'from module import *' used; unable to detect undefined names
     # W503 : Line break before binary operator
-    # E722 : Do not use bare 'except'
     # E203 : Whitespace before ':'
 
 - repo: https://github.com/pycqa/isort
   rev: 5.12.0
   hooks:
     - id: isort
-      language_version: python3.11
       args: ["--profile", "black"]
 
 - repo: https://github.com/psf/black.git
   rev: 23.7.0
   hooks:
     - id: black
-      language_version: python3.11
     - id: black-jupyter

benchmark/test_blas_perf.py

@@ -1,7 +1,7 @@
 import pytest
 import torch
 
-from .performance_utils import *
+from .performance_utils import BLAS_BATCH, DEFAULT_BATCH, FLOAT_DTYPES, SIZES, Benchmark
 
 
 @pytest.mark.parametrize("dtype", FLOAT_DTYPES)

benchmark/test_fused_perf.py

@@ -3,7 +3,14 @@
 
 import flag_gems
 
-from .performance_utils import *
+from .performance_utils import (
+    FLOAT_DTYPES,
+    POINTWISE_BATCH,
+    REDUCTION_BATCH,
+    SIZES,
+    Benchmark,
+    binary_args,
+)
 
 
 @pytest.mark.parametrize("dtype", FLOAT_DTYPES)

benchmark/test_pointwise_perf.py

@@ -1,7 +1,18 @@
 import pytest
 import torch
 
-from .performance_utils import *
+from .performance_utils import (
+    FLOAT_DTYPES,
+    INT_DTYPES,
+    POINTWISE_BATCH,
+    SIZES,
+    Benchmark,
+    binary_args,
+    binary_int_args,
+    ternary_args,
+    unary_arg,
+    unary_int_arg,
+)
 
 
 @pytest.mark.parametrize("dtype", FLOAT_DTYPES)

benchmark/test_reduction_perf.py

@@ -1,7 +1,14 @@
 import pytest
 import torch
 
-from .performance_utils import *
+from .performance_utils import (
+    BLAS_BATCH,
+    FLOAT_DTYPES,
+    REDUCTION_BATCH,
+    SIZES,
+    Benchmark,
+    unary_arg,
+)
 
 
 @pytest.mark.parametrize("dtype", FLOAT_DTYPES)

src/flag_gems/__init__.py

@@ -1,7 +1,7 @@
 import torch
 
-from .fused import *
-from .ops import *
+from .fused import *  # noqa: F403
+from .ops import *  # noqa: F403
 
 __version__ = "2.0"
 

src/flag_gems/ops/dropout.py

@@ -17,7 +17,7 @@ def _rand(seed, offset):
     _grid = (1,)
     _seed, _offset = philox_cuda_seed_offset(0)
     _rand[_grid](_seed, _offset)
-except:
+except Exception:
     tl_rand_dtype = tl.int32
 
 del _grid

tests/test_binary_pointwise_ops.py

@@ -3,7 +3,15 @@
 
 import flag_gems
 
-from .accuracy_utils import *
+from .accuracy_utils import (
+    FLOAT_DTYPES,
+    INT_DTYPES,
+    POINTWISE_SHAPES,
+    SCALARS,
+    gems_assert_close,
+    gems_assert_equal,
+    to_reference,
+)
 
 
 @pytest.mark.parametrize("shape", POINTWISE_SHAPES)

tests/test_blas_ops.py

@@ -3,7 +3,13 @@
 
 import flag_gems
 
-from .accuracy_utils import *
+from .accuracy_utils import (
+    FLOAT_DTYPES,
+    MNK_SHAPES,
+    SCALARS,
+    gems_assert_close,
+    to_reference,
+)
 
 
 @pytest.mark.parametrize("M", MNK_SHAPES)

tests/test_reduction_ops.py

@@ -3,7 +3,17 @@
 
 import flag_gems
 
-from .accuracy_utils import *
+from .accuracy_utils import (
+    DIM_LIST,
+    DIMS_LIST,
+    FLOAT_DTYPES,
+    REDUCTION_SHAPES,
+    gems_assert_close,
+    gems_assert_equal,
+    skip_expr,
+    skip_reason,
+    to_reference,
+)
 
 
 @pytest.mark.parametrize("shape", REDUCTION_SHAPES)

tests/test_special_ops.py

@@ -3,7 +3,12 @@
 
 import flag_gems
 
-from .accuracy_utils import *
+from .accuracy_utils import (
+    FLOAT_DTYPES,
+    POINTWISE_SHAPES,
+    gems_assert_close,
+    to_reference,
+)
 
 
 @pytest.mark.parametrize("shape", POINTWISE_SHAPES)
@@ -52,7 +57,7 @@ def get_rope_cos_sin(max_seq_len, dim, dtype, base=10000, device="cuda"):
 def rotate_half(x):
     """Rotates half the hidden dims of the input."""
     x1 = x[..., : x.shape[-1] // 2]
-    x2 = x[..., x.shape[-1] // 2 :]
+    x2 = x[..., x.shape[-1] // 2 :]  # noqa: E203
     return torch.cat((-x2, x1), dim=-1)
 
 

tests/test_unary_pointwise_ops.py

@@ -3,7 +3,14 @@
 
 import flag_gems
 
-from .accuracy_utils import *
+from .accuracy_utils import (
+    FLOAT_DTYPES,
+    INT_DTYPES,
+    POINTWISE_SHAPES,
+    gems_assert_close,
+    gems_assert_equal,
+    to_reference,
+)
 
 
 @pytest.mark.parametrize("shape", POINTWISE_SHAPES)