fix: change KVBM Connector build_connector_meta due to vLLM ToT change (#2887)

Signed-off-by: Ziqi Fan <ziqif@nvidia.com>

lib/bindings/python/src/dynamo/llm/vllm_integration/connector_leader.py

@@ -142,6 +142,7 @@ class KvConnectorLeader:
             scheduler_output.scheduled_cached_reqs.new_block_ids,
             scheduler_output.scheduled_cached_reqs.num_computed_tokens,
         ):
+            if new_block_ids is not None:
                 output.add_cached_request(
                     request_id=req_id,
                     resumed_from_preemption=resumed_from_preemption,
@@ -149,6 +150,14 @@ class KvConnectorLeader:
                     new_block_ids=new_block_ids[0],
                     num_computed_tokens=num_computed_tokens,
                 )
+            else:
+                output.add_cached_request(
+                    request_id=req_id,
+                    resumed_from_preemption=resumed_from_preemption,
+                    new_token_ids=new_token_ids,
+                    new_block_ids=[],
+                    num_computed_tokens=num_computed_tokens,
+                )
 
         output.add_num_scheduled_tokens(scheduler_output.num_scheduled_tokens)