resolve pre-commit test failure due to E501: line too long

Signed-off-by: qqma <qqma@amazon.com>

Author: qqma

vllm/v1/attention/backends/mla/flashattn_mla.py

@@ -103,7 +103,8 @@ def __init__(self, kv_cache_spec: AttentionSpec, layer_names: list[str],
                     "Getting flash attention max num splits for "
                     "cuda graph from environment variable, value=%s",
                     envs.VLLM_FLASH_ATTN_MAX_NUM_SPLITS_FOR_CUDA_GRAPH)
-                self.max_num_splits = envs.VLLM_FLASH_ATTN_MAX_NUM_SPLITS_FOR_CUDA_GRAPH
+                self.max_num_splits = (
+                    envs.VLLM_FLASH_ATTN_MAX_NUM_SPLITS_FOR_CUDA_GRAPH)
             else:
                 self.max_num_splits = _DEFAULT_MAX_NUM_SPLITS_FOR_CUDA_GRAPH