github/AutoGPT
1
1
Fork 0
mirror of https://github.com/Significant-Gravitas/AutoGPT.git synced 2026-02-10 14:55:16 -05:00
Code Issues Packages Projects Releases Wiki Activity

fix(backend/chat): Publish StreamError before StreamFinish on error paths

Browse Source 
When run_ai_generation() or event_generator() encounter errors, they
were only publishing StreamFinish without a preceding StreamError. The
frontend treats finish-without-error as normal completion, leaving the
user with an apparently stuck/empty response requiring a page refresh.
This commit is contained in:
Zamil Majdy
2026-02-10 15:49:23 +04:00
parent 8b8e1df739
commit 3c92a96504
1 changed files with 22 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

27
autogpt_platform/backend/backend/api/features/chat/routes.py
View File

@@ -25,7 +25,7 @@ from .model import (
get_user_sessions,
upsert_chat_session,
)
from .response_model import StreamFinish, StreamHeartbeat, StreamStart
from .response_model import StreamError, StreamFinish, StreamHeartbeat, StreamStart
from .sdk import service as sdk_service
from .tracking import track_user_message
@@ -296,7 +296,7 @@ async def stream_chat_post(
)
session = await _validate_and_get_session(session_id, user_id)
logger.info(
f"[TIMING] session validated in {(time.perf_counter() - stream_start_time)*1000:.1f}ms",
f"[TIMING] session validated in {(time.perf_counter() - stream_start_time) * 1000:.1f}ms",
extra={
"json_fields": {
**log_meta,
@@ -342,7 +342,7 @@ async def stream_chat_post(
operation_id=operation_id,
)
logger.info(
f"[TIMING] create_task completed in {(time.perf_counter() - task_create_start)*1000:.1f}ms",
f"[TIMING] create_task completed in {(time.perf_counter() - task_create_start) * 1000:.1f}ms",
extra={
"json_fields": {
**log_meta,
@@ -367,7 +367,7 @@ async def stream_chat_post(
start_chunk = StreamStart(messageId=task_id, taskId=task_id)
await stream_registry.publish_chunk(task_id, start_chunk)
logger.info(
f"[TIMING] StreamStart published at {(time_module.perf_counter() - gen_start_time)*1000:.1f}ms",
f"[TIMING] StreamStart published at {(time_module.perf_counter() - gen_start_time) * 1000:.1f}ms",
extra={
"json_fields": {
**log_meta,
@@ -417,7 +417,7 @@ async def stream_chat_post(
gen_end_time = time_module.perf_counter()
total_time = (gen_end_time - gen_start_time) * 1000
logger.info(
f"[TIMING] run_ai_generation FINISHED in {total_time/1000:.1f}s; "
f"[TIMING] run_ai_generation FINISHED in {total_time / 1000:.1f}s; "
f"task={task_id}, session={session_id}, "
f"ttfc={ttfc or -1:.2f}s, n_chunks={chunk_count}",
extra={
@@ -445,6 +445,17 @@ async def stream_chat_post(
}
},
)
# Publish a StreamError so the frontend can display an error message
try:
await stream_registry.publish_chunk(
task_id,
StreamError(
errorText="An error occurred. Please try again.",
code="stream_error",
),
)
except Exception:
pass # Best-effort; mark_task_completed will publish StreamFinish
await stream_registry.mark_task_completed(task_id, "failed")
# Start the AI generation in a background task
@@ -593,6 +604,12 @@ async def stream_chat_post(
"json_fields": {**log_meta, "elapsed_ms": elapsed, "error": str(e)}
},
)
# Surface error to frontend so it doesn't appear stuck
yield StreamError(
errorText="An error occurred. Please try again.",
code="stream_error",
).to_sse()
yield StreamFinish().to_sse()
finally:
# Unsubscribe when client disconnects or stream ends
if subscriber_queue is not None: