From f518afc88b3c0467a8da187be882dc0e17ef6293 Mon Sep 17 00:00:00 2001
From: sauravniraula <developmentsaurav@gmail.com>
Date: Thu, 14 Aug 2025 12:08:55 +0545
Subject: [PATCH] fix(fastapi): adds check in llm client for available openai
 event choices

---
 servers/fastapi/services/llm_client.py | 30 +++++++++-----------------
 1 file changed, 10 insertions(+), 20 deletions(-)

diff --git a/servers/fastapi/services/llm_client.py b/servers/fastapi/services/llm_client.py
index ce86f6db..487300ef 100644
--- a/servers/fastapi/services/llm_client.py
+++ b/servers/fastapi/services/llm_client.py
@@ -387,11 +387,7 @@ class LLMClient:
         max_tokens: Optional[int] = None,
         depth: int = 0,
     ):
-        extra_body = (
-            {"enable_thinking": False}
-            if self.disable_thinking()
-            else None
-        )
+        extra_body = {"enable_thinking": False} if self.disable_thinking() else None
         return await self._generate_openai(
             model=model,
             messages=messages,
@@ -758,11 +754,7 @@ class LLMClient:
         max_tokens: Optional[int] = None,
         depth: int = 0,
     ):
-        extra_body = (
-            {"enable_thinking": False}
-            if self.disable_thinking()
-            else None
-        )
+        extra_body = {"enable_thinking": False} if self.disable_thinking() else None
         return await self._generate_openai_structured(
             model=model,
             messages=messages,
@@ -860,6 +852,9 @@ class LLMClient:
             stream=True,
         ):
             event: OpenAIChatCompletionChunk = event
+            if not event.choices:
+                continue
+
             content_chunk = event.choices[0].delta.content
             if content_chunk:
                 yield content_chunk
@@ -1079,11 +1074,7 @@ class LLMClient:
         max_tokens: Optional[int] = None,
         depth: int = 0,
     ):
-        extra_body = (
-            {"enable_thinking": False}
-            if self.disable_thinking()
-            else None
-        )
+        extra_body = {"enable_thinking": False} if self.disable_thinking() else None
         return self._stream_openai(
             model=model,
             messages=messages,
@@ -1204,6 +1195,9 @@ class LLMClient:
             stream=True,
         ):
             event: OpenAIChatCompletionChunk = event
+            if not event.choices:
+                continue
+
             content_chunk = event.choices[0].delta.content
             if content_chunk:
                 yield content_chunk
@@ -1496,11 +1490,7 @@ class LLMClient:
         max_tokens: Optional[int] = None,
         depth: int = 0,
     ):
-        extra_body = (
-            {"enable_thinking": False}
-            if self.disable_thinking()
-            else None
-        )
+        extra_body = {"enable_thinking": False} if self.disable_thinking() else None
         return self._stream_openai_structured(
             model=model,
             messages=messages,