From a393a5aa354369e3b90fbbcf6040668e8e39b24c Mon Sep 17 00:00:00 2001
From: N0bodycan <49983270+N0bodycan@users.noreply.github.com>
Date: Tue, 9 Dec 2025 14:01:30 +0800
Subject: [PATCH] fix: fix infinite retry loop in async_chat_streamly when
 success

The original code continues the retry loop even after a successful response.
---
 rag/llm/chat_model.py | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/rag/llm/chat_model.py b/rag/llm/chat_model.py
index f3f207eb2..8a2743866 100644
--- a/rag/llm/chat_model.py
+++ b/rag/llm/chat_model.py
@@ -187,6 +187,9 @@ class Base(ABC):
                     ans = delta_ans
                     total_tokens += tol
                     yield ans
+
+                yield total_tokens
+                return
             except Exception as e:
                 e = await self._exceptions_async(e, attempt)
                 if e:
@@ -194,8 +197,6 @@ class Base(ABC):
                     yield total_tokens
                     return
 
-        yield total_tokens
-
     def _length_stop(self, ans):
         if is_chinese([ans]):
             return ans + LENGTH_NOTIFICATION_CN