From df873b8b27071be3e6dc166726ae77a8c37f5cfb Mon Sep 17 00:00:00 2001
From: Eric Huang <erichuang@meta.com>
Date: Wed, 3 Sep 2025 21:54:45 -0700
Subject: [PATCH] chore: handle missing finish_reason

# What does this PR do?
Sometimes the stream don't have chunks with finish_reason, e.g. canceled stream, which throws a pydantic error as OpenAIChoice.finish_reason: str

## Test Plan
observe no more such error when benchmarking
---
 llama_stack/core/routers/inference.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/llama_stack/core/routers/inference.py b/llama_stack/core/routers/inference.py
index 4b66601bb..8dcad85e3 100644
--- a/llama_stack/core/routers/inference.py
+++ b/llama_stack/core/routers/inference.py
@@ -755,7 +755,7 @@ class InferenceRouter(Inference):
                             choices_data[idx] = {
                                 "content_parts": [],
                                 "tool_calls_builder": {},
-                                "finish_reason": None,
+                                "finish_reason": "stop",
                                 "logprobs_content_parts": [],
                             }
                         current_choice_data = choices_data[idx]