[Bugfix] Missing tokens in `return_token_ids` when tool parsers is enabled in...

[Bugfix] Missing tokens in `return_token_ids` when tool parsers is enabled in streaming mode (#29074)
Signed-off-by: Peng-YM <1048217874pengym@gmail.com>

vllm/entrypoints/openai/serving_chat.py

@@ -1072,9 +1072,14 @@ class OpenAIServingChat(OpenAIServing):
                     # wasn't ready to send a token, then
                     #   get the next token without streaming a chunk
                     if delta_message is None:
-                        if output.finish_reason is None:
+                        # NOTE: If return_token_ids is enabled, we still need to
+                        # send a chunk with token_ids even if delta_message is None
+                        # to ensure all tokens are included in the response
+                        if (
+                            output.finish_reason is None
+                            and not request.return_token_ids
+                        ):
                             continue
-                        else:
                         delta_message = DeltaMessage()
 
                     # Log streaming delta if output logging is enabled