[CodeStyle][Ruff][BUAA][G-[181-190]] Fix Ruff RUF005 diagnostic for…

… 10 files in `test/ir/pir/fused_pass/onednn/` and `test/legacy_test/` (#67141)

---------

Co-authored-by: Nyakku Shigure <[email protected]>

test/ir/pir/fused_pass/onednn/test_operator_unsqueeze_onednn_fuse_pass.py

@@ -34,7 +34,7 @@ def build_ir_program(self):
                     name='x', shape=[5, 5, 5, 5], dtype='float32'
                 )
                 transpose = paddle.transpose(
-                    x, [len(x.shape) - 1] + list(range(0, len(x.shape) - 1))
+                    x, [len(x.shape) - 1, *range(0, len(x.shape) - 1)]
                 )
                 out = paddle.unsqueeze(transpose, [1])
                 out = paddle.assign(out)

test/legacy_test/ctr_dataset_reader.py

@@ -77,7 +77,7 @@ def reader():
                             dnn_input = load_dnn_input_record(fs[0])
                             lr_input = load_lr_input_record(fs[1])
                             click = [int(fs[2])]
-                            yield [dnn_input] + [lr_input] + [click]
+                            yield [dnn_input, lr_input, click]
 
         return reader
 

test/legacy_test/dist_allreduce_op.py

@@ -56,7 +56,7 @@ def cnn_model(data):
 
     SIZE = 10
     input_shape = conv_pool_2.shape
-    param_shape = [reduce(lambda a, b: a * b, input_shape[1:], 1)] + [SIZE]
+    param_shape = [reduce(lambda a, b: a * b, input_shape[1:], 1), SIZE]
     scale = (2.0 / (param_shape[0] ** 2 * SIZE)) ** 0.5
 
     predict = paddle.static.nn.fc(

test/legacy_test/dist_fleet_raw_program_optimizer.py

@@ -57,7 +57,7 @@ def cnn_model(data):
 
     SIZE = 10
     input_shape = conv_pool_2.shape
-    param_shape = [reduce(lambda a, b: a * b, input_shape[1:], 1)] + [SIZE]
+    param_shape = [reduce(lambda a, b: a * b, input_shape[1:], 1), SIZE]
     scale = (2.0 / (param_shape[0] ** 2 * SIZE)) ** 0.5
 
     predict = paddle.static.nn.fc(

test/legacy_test/dist_fleet_raw_program_optimizer_fuse_allreduce.py

@@ -57,7 +57,7 @@ def cnn_model(data):
 
     SIZE = 10
     input_shape = conv_pool_2.shape
-    param_shape = [reduce(lambda a, b: a * b, input_shape[1:], 1)] + [SIZE]
+    param_shape = [reduce(lambda a, b: a * b, input_shape[1:], 1), SIZE]
     scale = (2.0 / (param_shape[0] ** 2 * SIZE)) ** 0.5
 
     predict = paddle.static.nn.fc(

test/legacy_test/dist_mnist.py

@@ -57,7 +57,7 @@ def cnn_model(data):
 
     SIZE = 10
     input_shape = conv_pool_2.shape
-    param_shape = [reduce(lambda a, b: a * b, input_shape[1:], 1)] + [SIZE]
+    param_shape = [reduce(lambda a, b: a * b, input_shape[1:], 1), SIZE]
     scale = (2.0 / (param_shape[0] ** 2 * SIZE)) ** 0.5
 
     predict = paddle.static.nn.fc(

test/legacy_test/dist_mnist_dgc.py

@@ -58,7 +58,7 @@ def cnn_model(data):
 
     SIZE = 10
     input_shape = conv_pool_2.shape
-    param_shape = [reduce(lambda a, b: a * b, input_shape[1:], 1)] + [SIZE]
+    param_shape = [reduce(lambda a, b: a * b, input_shape[1:], 1), SIZE]
     scale = (2.0 / (param_shape[0] ** 2 * SIZE)) ** 0.5
 
     predict = paddle.static.nn.fc(

test/legacy_test/nets.py

@@ -577,7 +577,7 @@ def __split_heads(x, num_heads):
         # [batch_size, max_sequence_length, num_heads, hidden_size_per_head].
         reshaped = paddle.reshape(
             x=x,
-            shape=list(x.shape[:-1]) + [num_heads, hidden_size // num_heads],
+            shape=[*x.shape[:-1], num_heads, hidden_size // num_heads],
         )
 
         # permute the dimensions into:

test/legacy_test/seresnext_net.py

@@ -131,7 +131,7 @@ def bottleneck_block(input, num_filters, stride, cardinality, reduction_ratio):
 
 def SE_ResNeXt50Small(use_feed):
     img = paddle.static.data(
-        name='image', shape=[-1] + img_shape, dtype='float32'
+        name='image', shape=[-1, *img_shape], dtype='float32'
     )
     label = paddle.static.data(name='label', shape=[-1, 1], dtype='int64')
 

test/legacy_test/simple_nets.py

@@ -30,8 +30,9 @@ def pir_fc(hidden, size, activation, param_attr, bias_attr):
         input_shape = input.shape
         num_flatten_dims = len(input_shape) - 1
         param_shape = [
-            reduce(lambda a, b: a * b, input_shape[num_flatten_dims:], 1)
-        ] + [size]
+            reduce(lambda a, b: a * b, input_shape[num_flatten_dims:], 1),
+            size,
+        ]
 
         w = helper.create_parameter(
             attr=param_attr, shape=param_shape, dtype=input.dtype, is_bias=False
@@ -216,7 +217,7 @@ def bow_net(
 def init_data(batch_size=32, img_shape=[784], label_range=9):
     np.random.seed(5)
     assert isinstance(img_shape, list)
-    input_shape = [batch_size] + img_shape
+    input_shape = [batch_size, *img_shape]
     img = np.random.random(size=input_shape).astype(np.float32)
     label = (
         np.array([np.random.randint(0, label_range) for _ in range(batch_size)])

test/legacy_test/test_flash_attention.py

@@ -567,10 +567,10 @@ def gen_unpadded_data(self, dtype):
             low=1, high=self.seq_len, size=[self.batch_size]
         )
         cu_seqlen_q = paddle.to_tensor(
-            [0, *np.cumsum(seq_len_q)], dtype=paddle.int32
+            [0, *np.cumsum(seq_len_q).tolist()], dtype=paddle.int32
         )
         cu_seqlen_k = paddle.to_tensor(
-            [0, *np.cumsum(seq_len_k)], dtype=paddle.int32
+            [0, *np.cumsum(seq_len_k).tolist()], dtype=paddle.int32
         )
 
         qs, ks, vs = [], [], []
@@ -966,7 +966,7 @@ def gen_unpadded_data(self, dtype):
         )
         seq_len_k = seq_len_q
         cu_seqlen_q = paddle.to_tensor(
-            [0, *np.cumsum(seq_len_q)], dtype=paddle.int32
+            [0, *np.cumsum(seq_len_q).tolist()], dtype=paddle.int32
         )
         cu_seqlen_k = cu_seqlen_q
 

test/sequence/test_sequence_conv.py

@@ -278,7 +278,9 @@ def init_test_case(self):
         self.input_size = [self.input_row, 25]
         idx = list(range(self.input_size[0]))
         del idx[0]
-        offset_lod = [[0, *np.sort(random.sample(idx, 8)), self.input_size[0]]]
+        offset_lod = [
+            [0, *np.sort(random.sample(idx, 8)).tolist(), self.input_size[0]]
+        ]
         self.lod = [[]]
         # convert from offset-based lod to length-based lod
         for i in range(len(offset_lod[0]) - 1):