[Fix] Fix typing errors with python3.11

Signed-off-by: Haiyang Shi <[email protected]>

Author: Haiyang Shi

python/aibrix_kvcache/aibrix_kvcache/cache_hashable.py

@@ -78,7 +78,7 @@ def all_tokens(self) -> np.ndarray:
         return self._storage
 
     def all_tokens_memoryview(self) -> memoryview:
-        return memoryview(self._storage)
+        return memoryview(self._storage)  # type: ignore
 
     def __len__(self) -> int:
         return len(self._storage)
@@ -144,7 +144,7 @@ def __init__(self, mr: MemoryRegion):
 
     def all_tokens_memoryview(self) -> memoryview:
         assert self._mr is not None
-        return memoryview(np_array_concat(*self._mr.unpack_tokens()))
+        return memoryview(np_array_concat(*self._mr.unpack_tokens()))  # type: ignore
 
     def __len__(self) -> int:
         assert self._mr is not None

python/aibrix_kvcache/aibrix_kvcache/l2/key_builders/rolling_hash_key_builder.py

@@ -49,7 +49,7 @@ def build(
                     start = 1
                 candidates[1:] = prefix[i : i + self.block_size]
 
-                prev_hash = self.hasher.hash(memoryview(candidates[start:]))
+                prev_hash = self.hasher.hash(memoryview(candidates[start:]))  # type: ignore
 
         all = np_array_concat(prefix, tokens)
         prefix_len = len(prefix) if prefix is not None else 0
@@ -62,7 +62,7 @@ def build(
             candidates[1:] = tokens[i : i + self.block_size]
             keys = all[: prefix_len + i + self.block_size]
 
-            curr_hash = self.hasher.hash(memoryview(candidates[start:]))
+            curr_hash = self.hasher.hash(memoryview(candidates[start:]))  # type: ignore
 
             results.append((keys, curr_hash.to_bytes(16)))
             prev_hash = curr_hash

python/aibrix_kvcache/aibrix_kvcache/l2/key_builders/simple_hash_key_builder.py

@@ -40,7 +40,7 @@ def build(
 
         prefix_len = len(prefix) if prefix is not None else 0
         all = np_array_concat(prefix, tokens)
-        all_bytes = memoryview(all)
+        all_bytes = memoryview(all)  # type: ignore
         for i in range(0, token_size, self.block_size):
             keys = all[: prefix_len + i + self.block_size]