Skip to content

Fix non-stream Responses output aggregation #121

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
EmBista wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Fix non-stream Responses output aggregation #121

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
1657818
Fix non-stream Responses output aggregation
EmBista Aug 20, 2026
83c35f1
Preserve Responses output item ordering
EmBista Aug 20, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
24 changes: 24 additions & 0 deletions chatmock/responses_api.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
from __future__ import annotations

import copy
import json
from dataclasses import dataclass
from typing import Any, Dict, Iterable, Iterator, List
Expand Down Expand Up @@ -171,6 +172,8 @@ def aggregate_response_from_sse(
) -> tuple[Dict[str, Any] | None, Dict[str, Any] | None]:
response_obj: Dict[str, Any] | None = None
error_obj: Dict[str, Any] | None = None
completed_output_items: Dict[int, Dict[str, Any]] = {}
unindexed_output_items = 0
try:
for evt in iter_sse_event_payloads(upstream):
if callable(on_event):
Expand All @@ -182,13 +185,34 @@ def aggregate_response_from_sse(
if isinstance(response, dict):
response_obj = response
kind = evt.get("type")
if kind == "response.output_item.done":
item = evt.get("item")
if isinstance(item, dict):
output_index = evt.get("output_index")
if not isinstance(output_index, int):
# Indexed items are the protocol norm. Keep malformed or
# older unindexed events deterministically after them,
# preserving their arrival order.
output_index = 1_000_000 + unindexed_output_items
unindexed_output_items += 1
completed_output_items[output_index] = copy.deepcopy(item)
if kind == "response.failed":
if isinstance(response, dict) and isinstance(response.get("error"), dict):
error_obj = {"error": response.get("error")}
else:
error_obj = {"error": {"message": "response.failed"}}
break
if kind == "response.completed":
if (
isinstance(response_obj, dict)
and completed_output_items
and not response_obj.get("output")
):
response_obj = dict(response_obj)
response_obj["output"] = [
completed_output_items[index]
for index in sorted(completed_output_items)
]
break
finally:
upstream.close()
Expand Down
102 changes: 102 additions & 0 deletions tests/test_routes.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -260,6 +260,108 @@ def test_responses_route_returns_completed_response_object(self, mock_start) ->
self.assertEqual(outbound_payload["reasoning"]["effort"], "medium")
self.assertIsInstance(outbound_payload["prompt_cache_key"], str)

@patch("chatmock.routes_openai.start_upstream_raw_request")
def test_responses_route_reconstructs_non_stream_output_from_item_events(self, mock_start) -> None:
output = [
{
"type": "reasoning",
"id": "reasoning_1",
"summary": [{"type": "summary_text", "text": "Need the tool."}],
"encrypted_content": "encrypted",
},
{
"type": "function_call",
"id": "fc_1",
"call_id": "call_1",
"name": "get_time",
"arguments": '{"city":"Paris"}',
"status": "completed",
},
{
"type": "message",
"role": "assistant",
"id": "msg_1",
"status": "completed",
"content": [{"type": "output_text", "text": '{"city":"Paris"}'}],
},
]
events = [
{
"type": "response.created",
"response": {"id": "resp_items", "object": "response", "status": "in_progress"},
},
*[
{"type": "response.output_item.done", "output_index": index, "item": output[index]}
# Completion order is not output order; the protocol supplies
# output_index so non-stream aggregation can reconstruct it.
for index in (2, 0, 1)
],
{
"type": "response.completed",
"response": {
"id": "resp_items",
"object": "response",
"status": "completed",
"output": [],
},
},
]
mock_start.return_value = (
FakeUpstream(events, headers={"Content-Type": "text/event-stream"}),
None,
)

response = self.client.post(
"/v1/responses",
json={"model": "gpt-5.6-luna", "input": "Return structured output."},
)

self.assertEqual(response.status_code, 200)
self.assertEqual(response.get_json()["output"], output)

@patch("chatmock.routes_openai.start_upstream_raw_request")
def test_responses_route_keeps_output_from_completed_response(self, mock_start) -> None:
authoritative = {
"type": "message",
"role": "assistant",
"id": "msg_final",
"content": [{"type": "output_text", "text": "final"}],
}
mock_start.return_value = (
FakeUpstream(
[
{
"type": "response.output_item.done",
"item": {
"type": "message",
"role": "assistant",
"id": "msg_event",
"content": [{"type": "output_text", "text": "event"}],
},
},
{
"type": "response.completed",
"response": {
"id": "resp_final",
"object": "response",
"status": "completed",
"output": [authoritative],
},
},
],
headers={"Content-Type": "text/event-stream"},
),
None,
)

response = self.client.post(
"/v1/responses",
json={"model": "gpt-5.6-luna", "input": "hello"},
)

self.assertEqual(response.status_code, 200)
self.assertEqual(response.get_json()["output"], [authoritative])

@patch("chatmock.routes_openai.start_upstream_raw_request")
def test_responses_route_honors_debug_model_override(self, mock_start) -> None:
app = create_app(debug_model="gpt-5.4", model_sync=False)
Expand Down