From 0a5e971bf7cd5347b42bc58ae5d56662fd36fc18 Mon Sep 17 00:00:00 2001
From: Vadym Samoilenko <vadymsamoilenko@oliver.agency>
Date: Fri, 27 Feb 2026 22:34:26 +0000
Subject: [PATCH] Add timeout protection to outline generation

Prevents socket hang up when LLM is slow or unavailable.
- Timeout: 120 seconds (configurable via OUTLINE_TIMEOUT_SECONDS)
- Returns 504 Gateway Timeout error if exceeded
- Fixes: socket hang up errors in Next.js proxy

Co-Authored-By: Claude Sonnet 4.5 (1M context) <noreply@anthropic.com>
---
 .../generate_presentation_outlines.py         | 65 ++++++++++++-------
 1 file changed, 42 insertions(+), 23 deletions(-)

diff --git a/backend/utils/llm_calls/generate_presentation_outlines.py b/backend/utils/llm_calls/generate_presentation_outlines.py
index f76f6dd..4649fa2 100644
--- a/backend/utils/llm_calls/generate_presentation_outlines.py
+++ b/backend/utils/llm_calls/generate_presentation_outlines.py
@@ -1,6 +1,9 @@
+import asyncio
+import os
 from datetime import datetime
 from typing import Optional
 
+from fastapi import HTTPException
 from models.llm_message import LLMSystemMessage, LLMUserMessage
 from models.llm_tools import SearchWebTool
 from services.llm_client import LLMClient
@@ -137,30 +140,46 @@ async def generate_ppt_outline(
 
     client = LLMClient()
 
+    # Timeout for outline generation (prevents socket hang up)
+    OUTLINE_TIMEOUT = int(os.getenv("OUTLINE_TIMEOUT_SECONDS", "120"))
+
     try:
-        async for chunk in client.stream_structured(
-            model,
-            get_messages(
-                content,
-                n_slides,
-                language,
-                additional_context,
-                tone,
-                verbosity,
-                instructions,
-                include_title_slide,
-                brand_context,
-                available_layouts,
-                content_summary,
-            ),
-            response_model.model_json_schema(),
-            strict=True,
-            tools=(
-                [SearchWebTool]
-                if (client.enable_web_grounding() and web_search)
-                else None
-            ),
-        ):
+        # Wrap streaming in timeout to prevent indefinite hangs
+        async def stream_with_timeout():
+            try:
+                async for chunk in client.stream_structured(
+                    model,
+                    get_messages(
+                        content,
+                        n_slides,
+                        language,
+                        additional_context,
+                        tone,
+                        verbosity,
+                        instructions,
+                        include_title_slide,
+                        brand_context,
+                        available_layouts,
+                        content_summary,
+                    ),
+                    response_model.model_json_schema(),
+                    strict=True,
+                    tools=(
+                        [SearchWebTool]
+                        if (client.enable_web_grounding() and web_search)
+                        else None
+                    ),
+                ):
+                    yield chunk
+            except asyncio.TimeoutError:
+                yield HTTPException(
+                    status_code=504,
+                    detail=f"Outline generation timed out after {OUTLINE_TIMEOUT}s. Please try again or reduce the number of slides."
+                )
+            except Exception as e:
+                yield handle_llm_client_exceptions(e)
+
+        async for chunk in stream_with_timeout():
             yield chunk
     except Exception as e:
         yield handle_llm_client_exceptions(e)