[Bugfix] Add verbose error if scipy is missing for blocksparse attention (#5695)

vllm/attention/ops/blocksparse_attention/utils.py

@@ -6,7 +6,14 @@ from functools import lru_cache
 
 import torch
 import triton
-from scipy import sparse
+
+try:
+    from scipy import sparse
+except ImportError as err:
+    raise ImportError("Please install scipy via "
+                      "`pip install scipy` to use "
+                      "BlockSparseAttention in "
+                      "models such as Phi-3.") from err
 
 
 def dense_to_crow_col(x: torch.Tensor):
@@ -77,11 +84,11 @@ def _get_sparse_attn_mask_homo_head(
 ):
     """
     :return: a tuple of 3:
-        - tuple of crow_indices, col_indices representation 
+        - tuple of crow_indices, col_indices representation
             of CSR format.
         - block dense mask
-        - all token dense mask (be aware that it can be 
-            OOM if it is too big) if `return_dense==True`, 
+        - all token dense mask (be aware that it can be
+            OOM if it is too big) if `return_dense==True`,
             otherwise, None
     """
     with torch.no_grad():
@@ -148,10 +155,10 @@ def get_sparse_attn_mask(
     :param dense_mask_type: "binary" (0 for skip token, 1 for others)
         or "bias" (-inf for skip token, 0 or others)
     :return: a tuple of 3:
-        - tuple of crow_indices, col_indices representation 
+        - tuple of crow_indices, col_indices representation
             of CSR format.
         - block dense mask
-        - all token dense mask (be aware that it can be OOM if it 
+        - all token dense mask (be aware that it can be OOM if it
             is too big) if `return_dense==True`, otherwise, None
     """
     assert dense_mask_type in ("binary", "bias")