diff --git a/litellm/llms/openai/responses/guardrail_translation/handler.py b/litellm/llms/openai/responses/guardrail_translation/handler.py index 559daf767c2..64aca9496eb 100644 --- a/litellm/llms/openai/responses/guardrail_translation/handler.py +++ b/litellm/llms/openai/responses/guardrail_translation/handler.py @@ -1974,7 +1974,7 @@ class OpenAIResponsesHandler(BaseTranslation): "annotations": (), "logprobs": None, } - return ( + events: Final = ( *_open_item_closing_events(responses_so_far), OutputItemAddedEvent( type=ResponsesAPIStreamEvents.OUTPUT_ITEM_ADDED, @@ -2019,6 +2019,10 @@ class OpenAIResponsesHandler(BaseTranslation): response=_blocked_response(exc, response_id=response_id, model=model, output_item=item), ), ) + return tuple( + event.model_copy(update={"sequence_number": number}) + for number, event in enumerate(events, start=_next_stream_sequence_number(responses_so_far)) + ) class _BlockedContentPart(TypedDict): diff --git a/litellm/responses/litellm_completion_transformation/streaming_iterator.py b/litellm/responses/litellm_completion_transformation/streaming_iterator.py index 892da541935..ae58fb0ce79 100644 --- a/litellm/responses/litellm_completion_transformation/streaming_iterator.py +++ b/litellm/responses/litellm_completion_transformation/streaming_iterator.py @@ -26,7 +26,11 @@ from litellm.responses.litellm_completion_transformation.tool_search import ( from litellm.responses.litellm_completion_transformation.transformation import ( LiteLLMCompletionResponsesConfig, ) -from litellm.responses.streaming_iterator import ResponsesAPIStreamingIterator, stamp_responses_usage_cost +from litellm.responses.streaming_iterator import ( + SEQUENCE_NUMBER_STAMPED_ON_EMISSION, + ResponsesAPIStreamingIterator, + stamp_responses_usage_cost, +) from litellm.responses.utils import ResponsesAPIRequestUtils from litellm.types.llms.openai import ( PART_UNION_TYPES, @@ -210,14 +214,11 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): return execution def _next_output_item_added(self, output_index: int, item: BaseLiteLLMOpenAIResponseObject) -> OutputItemAddedEvent: - self._sequence_number += 1 - added: Final = OutputItemAddedEvent( + return OutputItemAddedEvent( type=ResponsesAPIStreamEvents.OUTPUT_ITEM_ADDED, output_index=output_index, item=item, ) - added.__dict__["sequence_number"] = self._sequence_number - return added def _queue_output_item_added(self, output_index: int, item: BaseLiteLLMOpenAIResponseObject) -> None: self._pending_tool_events.append(self._next_output_item_added(output_index, item)) @@ -239,12 +240,10 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): item: Final = build_tool_search_call( call_id, final_args or self._tool_args_by_call_id.get(call_id, ""), "completed", execution ) - self._sequence_number += 1 self._pending_tool_events.append( OutputItemDoneEvent( type=ResponsesAPIStreamEvents.OUTPUT_ITEM_DONE, output_index=output_index, - sequence_number=self._sequence_number, item=BaseLiteLLMOpenAIResponseObject(**item.model_dump()), ) ) @@ -317,12 +316,10 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): if streamed_item is None: return self._sent_compaction_item_done = True - self._sequence_number += 1 self._pending_response_events.append( OutputItemDoneEvent( type=ResponsesAPIStreamEvents.OUTPUT_ITEM_DONE, output_index=self._compaction_output_index, - sequence_number=self._sequence_number, item=streamed_item, ) ) @@ -465,15 +462,12 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): chunk_size = 10 # Match typical OpenAI delta size for i in range(0, len(fn_args_delta), chunk_size): delta_chunk = fn_args_delta[i : i + chunk_size] - self._sequence_number += 1 delta_event: BaseLiteLLMOpenAIResponseObject = FunctionCallArgumentsDeltaEvent( type=ResponsesAPIStreamEvents.FUNCTION_CALL_ARGUMENTS_DELTA, item_id=self._tool_item_id_by_call_id.get(call_id, call_id), output_index=output_index, delta=delta_chunk, ) - # Add sequence_number as extra field (BaseLiteLLMOpenAIResponseObject allows extra fields) - delta_event.__dict__["sequence_number"] = self._sequence_number self._pending_tool_events.append(delta_event) def _queue_final_tool_call_done_events(self, litellm_complete_object: ModelResponse) -> None: @@ -541,33 +535,27 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): chunk_size = 10 # Match typical OpenAI delta size for i in range(0, len(remaining_args), chunk_size): delta_chunk = remaining_args[i : i + chunk_size] - self._sequence_number += 1 delta_event = FunctionCallArgumentsDeltaEvent( type=ResponsesAPIStreamEvents.FUNCTION_CALL_ARGUMENTS_DELTA, item_id=self._tool_item_id_by_call_id.get(call_id, call_id), output_index=output_index, delta=delta_chunk, ) - delta_event.__dict__["sequence_number"] = self._sequence_number self._pending_tool_events.append(delta_event) - self._sequence_number += 1 done_event = FunctionCallArgumentsDoneEvent( type=ResponsesAPIStreamEvents.FUNCTION_CALL_ARGUMENTS_DONE, item_id=self._tool_item_id_by_call_id.get(call_id, call_id), output_index=output_index, arguments=final_args, ) - done_event.__dict__["sequence_number"] = self._sequence_number self._pending_tool_events.append(done_event) - self._sequence_number += 1 item_kwargs = self._tool_call_item_kwargs(call_id, fn_name, final_args, "completed") item_kwargs["id"] = self._tool_item_id_by_call_id.setdefault(call_id, item_kwargs["id"]) item_done_event = OutputItemDoneEvent( type=ResponsesAPIStreamEvents.OUTPUT_ITEM_DONE, output_index=output_index, - sequence_number=self._sequence_number, item=BaseLiteLLMOpenAIResponseObject(**item_kwargs), ) self._pending_tool_events.append(item_done_event) @@ -597,16 +585,11 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): (ResponsesAPIStreamEvents.WEB_SEARCH_CALL_SEARCHING, WebSearchCallSearchingEvent), (ResponsesAPIStreamEvents.WEB_SEARCH_CALL_COMPLETED, WebSearchCallCompletedEvent), ): - self._sequence_number += 1 - event = event_class(type=event_type, output_index=output_index, item_id=item.id) - event.__dict__["sequence_number"] = self._sequence_number - self._pending_tool_events.append(event) - self._sequence_number += 1 + self._pending_tool_events.append(event_class(type=event_type, output_index=output_index, item_id=item.id)) self._pending_tool_events.append( OutputItemDoneEvent( type=ResponsesAPIStreamEvents.OUTPUT_ITEM_DONE, output_index=output_index, - sequence_number=self._sequence_number, item=BaseLiteLLMOpenAIResponseObject(**item.model_dump()), ) ) @@ -742,31 +725,24 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): """ response_created_event_data: Final = self._default_response_created_event_data() - self._sequence_number += 1 - event: Final = ResponseCreatedEvent( + return ResponseCreatedEvent( type=ResponsesAPIStreamEvents.RESPONSE_CREATED, response=self._with_encoded_response_id(ResponsesAPIResponse(**response_created_event_data)), ) - event.__dict__["sequence_number"] = self._sequence_number - return event def create_response_in_progress_event(self) -> ResponseInProgressEvent: response_in_progress_event_data: Final = self._default_response_created_event_data() response_in_progress_event_data["status"] = "in_progress" - self._sequence_number += 1 - event: Final = ResponseInProgressEvent( + return ResponseInProgressEvent( type=ResponsesAPIStreamEvents.RESPONSE_IN_PROGRESS, response=self._with_encoded_response_id(ResponsesAPIResponse(**response_in_progress_event_data)), ) - event.__dict__["sequence_number"] = self._sequence_number - return event def create_output_item_added_event(self) -> OutputItemAddedEvent: if self._cached_item_id is None: self._cached_item_id = f"msg_{uuid.uuid4()}" - self._sequence_number += 1 - event: Final = OutputItemAddedEvent( + return OutputItemAddedEvent( type=ResponsesAPIStreamEvents.OUTPUT_ITEM_ADDED, output_index=self._message_index(), item=BaseLiteLLMOpenAIResponseObject( @@ -779,29 +755,23 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): } ), ) - event.__dict__["sequence_number"] = self._sequence_number - return event def create_content_part_added_event(self) -> ContentPartAddedEvent: if self._cached_item_id is None: self._cached_item_id = f"msg_{uuid.uuid4()}" - self._sequence_number += 1 - event: Final = ContentPartAddedEvent( + return ContentPartAddedEvent( type=ResponsesAPIStreamEvents.CONTENT_PART_ADDED, item_id=self._cached_item_id, output_index=self._message_index(), content_index=0, part=BaseLiteLLMOpenAIResponseObject(**{"type": "output_text", "text": "", "annotations": []}), ) - event.__dict__["sequence_number"] = self._sequence_number - return event def _queue_message_item_added_events(self) -> None: item_id: Final = self._message_item_id() self.sent_message_item_added_event = True self.sent_content_part_added_event = True - self._sequence_number += 1 event: Final = OutputItemAddedEvent( type=ResponsesAPIStreamEvents.OUTPUT_ITEM_ADDED, output_index=self._message_index(), @@ -815,7 +785,6 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): } ), ) - event.__dict__["sequence_number"] = self._sequence_number self._pending_response_events.append(event) self._pending_response_events.append(self.create_content_part_added_event()) @@ -871,7 +840,6 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): self, reasoning_item_id: str, reasoning_content: str, - sequence_number: int, ) -> ReasoningSummaryTextDoneEvent: """ Create response.reasoning_summary_text.done event. @@ -881,7 +849,6 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): "type": "response.reasoning_summary_text.done", "item_id": "rs_0c5dae30e53172980069708ba2f59c8197b71ca9820edad07c", "output_index": 0, - "sequence_number": 97, "summary_index": 0, "text": "**Clarifying the first humans**\n\nThe I'm addressing the user's specific interest." } @@ -890,7 +857,7 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): type=ResponsesAPIStreamEvents.REASONING_SUMMARY_TEXT_DONE, item_id=reasoning_item_id, output_index=self._reasoning_index(), - sequence_number=sequence_number, + sequence_number=SEQUENCE_NUMBER_STAMPED_ON_EMISSION, summary_index=0, text=reasoning_content, ) @@ -899,7 +866,6 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): self, reasoning_item_id: str, reasoning_content: str, - sequence_number: int, ) -> ReasoningSummaryPartDoneEvent: """ Create response.reasoning_summary_part.done event. @@ -913,7 +879,6 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): "type": "summary_text", "text": "**Clarifying the first humans**\n\nThe earlier hominins. It feels important to ensure I'm addressing the user's specific interest." }, - "sequence_number": 98, "summary_index": 0 } """ @@ -921,7 +886,7 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): type=ResponsesAPIStreamEvents.REASONING_SUMMARY_PART_DONE, item_id=reasoning_item_id, output_index=self._reasoning_index(), - sequence_number=sequence_number, + sequence_number=SEQUENCE_NUMBER_STAMPED_ON_EMISSION, summary_index=0, part=BaseLiteLLMOpenAIResponseObject( **{ @@ -985,7 +950,6 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): return OutputItemDoneEvent( type=ResponsesAPIStreamEvents.OUTPUT_ITEM_DONE, output_index=self._message_index(), - sequence_number=1, item=BaseLiteLLMOpenAIResponseObject( **{ "id": self._cached_item_id, @@ -1007,7 +971,6 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): self, reasoning_item_id: str, reasoning_content: str, - sequence_number: int, ) -> OutputItemDoneEvent: """ Create response.output_item.done event for reasoning items. @@ -1016,7 +979,6 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): { "type": "response.output_item.done", "output_index": 0, - "sequence_number": 99, "item": { "id": "rs_0c5dae30e53172980069708ba2f59c8197b71ca9820edad07c", "type": "reasoning", @@ -1032,7 +994,6 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): return OutputItemDoneEvent( type=ResponsesAPIStreamEvents.OUTPUT_ITEM_DONE, output_index=self._reasoning_index(), - sequence_number=sequence_number, item=BaseLiteLLMOpenAIResponseObject( **{ "id": reasoning_item_id, @@ -1137,7 +1098,6 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): if chunk.choices[0].finish_reason is None and not _delta_carries_output(delta): return - self._sequence_number += 1 self.sent_output_item_added_event = True # Reasoning-first @@ -1147,20 +1107,20 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): self._cached_reasoning_item_id = mint_reasoning_item_id() self._reasoning_item_id = self._cached_reasoning_item_id - event = OutputItemAddedEvent( - type=ResponsesAPIStreamEvents.OUTPUT_ITEM_ADDED, - output_index=self._reasoning_index(), - item=BaseLiteLLMOpenAIResponseObject( - **{ - "id": self._cached_reasoning_item_id, - "type": "reasoning", - "status": "in_progress", - "summary": None, - } - ), + self._pending_response_events.append( + OutputItemAddedEvent( + type=ResponsesAPIStreamEvents.OUTPUT_ITEM_ADDED, + output_index=self._reasoning_index(), + item=BaseLiteLLMOpenAIResponseObject( + **{ + "id": self._cached_reasoning_item_id, + "type": "reasoning", + "status": "in_progress", + "summary": None, + } + ), + ) ) - event.__dict__["sequence_number"] = self._sequence_number - self._pending_response_events.append(event) return # Tool-first @@ -1191,30 +1151,34 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): ) reasoning_item_id: Final = self._cached_reasoning_item_id self._queue_compaction_item_done() - self._sequence_number += 1 text_done_event: Final = self.create_reasoning_summary_text_done_event( reasoning_item_id=reasoning_item_id, reasoning_content=reasoning_content, - sequence_number=self._sequence_number, ) - self._sequence_number += 1 part_done_event: Final = self.create_reasoning_summary_part_done_event( reasoning_item_id=reasoning_item_id, reasoning_content=reasoning_content, - sequence_number=self._sequence_number, ) - self._sequence_number += 1 reasoning_output_item_done_event: Final = self.create_reasoning_output_item_done_event( reasoning_item_id=reasoning_item_id, reasoning_content=reasoning_content, - sequence_number=self._sequence_number, ) self._pending_response_events.extend([text_done_event, part_done_event, reasoning_output_item_done_event]) self._reasoning_done_emitted = True self._reasoning_active = False + def _stamp_sequence_number(self, event: BaseLiteLLMOpenAIResponseObject) -> BaseLiteLLMOpenAIResponseObject: + stamped: Final = event.model_copy(update={"sequence_number": self._sequence_number}) + self._sequence_number += 1 + return stamped + async def __anext__( self, + ) -> ResponsesAPIStreamingResponse | ResponseCompletedEvent | BaseLiteLLMOpenAIResponseObject: + return self._stamp_sequence_number(await self._anext_unstamped_event()) + + async def _anext_unstamped_event( + self, ) -> ResponsesAPIStreamingResponse | ResponseCompletedEvent | BaseLiteLLMOpenAIResponseObject: try: while True: @@ -1275,6 +1239,11 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): def __next__( self, + ) -> ResponsesAPIStreamingResponse | ResponseCompletedEvent | BaseLiteLLMOpenAIResponseObject: + return self._stamp_sequence_number(self._next_unstamped_event()) + + def _next_unstamped_event( + self, ) -> ResponsesAPIStreamingResponse | ResponseCompletedEvent | BaseLiteLLMOpenAIResponseObject: try: while True: @@ -1416,16 +1385,13 @@ class LiteLLMCompletionStreamingIterator(ResponsesAPIStreamingIterator): if not self.sent_message_item_added_event: self._queue_message_item_added_events() self._hold_tool_calls_behind_text(getattr(chunk.choices[0].delta, "tool_calls", None)) - self._sequence_number += 1 - text_delta_event: Final = OutputTextDeltaEvent( + return OutputTextDeltaEvent( type=ResponsesAPIStreamEvents.OUTPUT_TEXT_DELTA, item_id=item_id, output_index=self._message_index(), content_index=0, delta=delta_content, ) - text_delta_event.__dict__["sequence_number"] = self._sequence_number - return text_delta_event # Priority 3: Handle tool call deltas (if any) -> queue events and emit them # For each tool call delta, we emit events one at a time to match OpenAI's streaming behavior diff --git a/litellm/responses/streaming_iterator.py b/litellm/responses/streaming_iterator.py index a8e55a5e94d..dc3880e4690 100644 --- a/litellm/responses/streaming_iterator.py +++ b/litellm/responses/streaming_iterator.py @@ -365,6 +365,7 @@ def _mid_stream_fallback_eligible(mapped_exception: Exception) -> bool: PRE_OUTPUT_LIFECYCLE_EVENT_TYPES: Final = frozenset({"response.created", "response.in_progress", "response.queued"}) +SEQUENCE_NUMBER_STAMPED_ON_EMISSION: Final = -1 class BaseResponsesAPIStreamingIterator: @@ -1911,7 +1912,6 @@ def build_synthetic_response_events( _build_response_status_event(openai_types.ResponsesAPIStreamEvents.RESPONSE_IN_PROGRESS, transformed), ] - sequence_number = 0 output_items: Final[Sequence[object]] = getattr(transformed, "output", []) or [] for output_index, output_item in enumerate(output_items): output_item_payload = _dump_response_object(output_item) @@ -2021,35 +2021,31 @@ def build_synthetic_response_events( delta=summary_text[i : i + chunk_size], ) ) - sequence_number += 1 events.append( openai_types.ReasoningSummaryTextDoneEvent( type=openai_types.ResponsesAPIStreamEvents.REASONING_SUMMARY_TEXT_DONE, item_id=item_id, output_index=output_index, - sequence_number=sequence_number, + sequence_number=SEQUENCE_NUMBER_STAMPED_ON_EMISSION, summary_index=summary_index, text=summary_text, ) ) - sequence_number += 1 events.append( openai_types.ReasoningSummaryPartDoneEvent( type=openai_types.ResponsesAPIStreamEvents.REASONING_SUMMARY_PART_DONE, item_id=item_id, output_index=output_index, - sequence_number=sequence_number, + sequence_number=SEQUENCE_NUMBER_STAMPED_ON_EMISSION, summary_index=summary_index, part=openai_types.BaseLiteLLMOpenAIResponseObject(**summary_payload), ) ) - sequence_number += 1 events.append( openai_types.OutputItemDoneEvent( type=openai_types.ResponsesAPIStreamEvents.OUTPUT_ITEM_DONE, output_index=output_index, - sequence_number=sequence_number, item=openai_types.BaseLiteLLMOpenAIResponseObject(**output_item_payload), ) ) @@ -2060,7 +2056,7 @@ def build_synthetic_response_events( response=transformed, ) ) - return events + return [event.model_copy(update={"sequence_number": number}) for number, event in enumerate(events)] # --------------------------------------------------------------------------- diff --git a/litellm/router.py b/litellm/router.py index 535072d5fcf..471c380ebf5 100644 --- a/litellm/router.py +++ b/litellm/router.py @@ -551,6 +551,7 @@ if TYPE_CHECKING: ResponseAPIUsage, ResponseIncompleteEvent, ResponsesAPIResponse, + ResponsesAPIStreamingResponse, ) Span = _Span @@ -4301,12 +4302,13 @@ class Router(DeprecatedRouterMembers): fallback_headers_adopted: bool = False - def __init__(self, async_generator: AsyncGenerator): + def __init__(self, async_generator: "AsyncGenerator[ResponsesAPIStreamingResponse, None]"): import time from datetime import datetime self._async_generator = async_generator self._billing_source: object = source_iterator + self._last_sequence_number = -1 # Mirror every attribute BaseResponsesAPIStreamingIterator.__init__ # would have set. The wrapper bypasses super().__init__ (it has no # httpx.Response of its own and no provider config to drive), so @@ -4413,7 +4415,13 @@ class Router(DeprecatedRouterMembers): if self.completed_response is None and getattr(chunk, "type", None) in _RESPONSES_TERMINAL_EVENT_TYPES: self.completed_response = chunk self._yielded_first_chunk = True - return chunk + sequence_number: Final = getattr(chunk, "sequence_number", None) + if not isinstance(sequence_number, int): + return chunk + self._last_sequence_number = max(self._last_sequence_number + 1, sequence_number) + if self._last_sequence_number == sequence_number: + return chunk + return chunk.model_copy(update={"sequence_number": self._last_sequence_number}) async def aclose(self): # async generators always expose aclose — no defensive check needed. @@ -4475,7 +4483,9 @@ class Router(DeprecatedRouterMembers): exc, ) - wrapper: Final = FallbackResponsesStreamWrapper(stream_with_fallbacks()) + wrapper: Final = FallbackResponsesStreamWrapper( + stream_with_fallbacks() # pyright: ignore[reportArgumentType] # relays response events typed as object + ) return wrapper def _aresponses_background_stream_can_release( diff --git a/litellm/types/llms/base.py b/litellm/types/llms/base.py index 5a2da071cec..37937892e84 100644 --- a/litellm/types/llms/base.py +++ b/litellm/types/llms/base.py @@ -32,16 +32,16 @@ class BaseLiteLLMOpenAIResponseObject(LiteLLMBaseModel): model_config = ConfigDict(extra="allow", protected_namespaces=()) def __getitem__(self, key): - return self.__dict__[key] + return dict(iter(self))[key] def get(self, key, default=None): - return self.__dict__.get(key, default) + return dict(iter(self)).get(key, default) def __contains__(self, key) -> bool: - return key in self.__dict__ + return key in dict(iter(self)) def items(self): - return self.__dict__.items() + return dict(iter(self)).items() class HiddenParams(OpenAIObject): diff --git a/tests/unit/llms/openai/responses/test_openai_responses_guardrail_handler.py b/tests/unit/llms/openai/responses/test_openai_responses_guardrail_handler.py index d732d922375..c0277b59c16 100644 --- a/tests/unit/llms/openai/responses/test_openai_responses_guardrail_handler.py +++ b/tests/unit/llms/openai/responses/test_openai_responses_guardrail_handler.py @@ -4515,6 +4515,34 @@ class TestBuildBlockSseChunks: assert completed["output"][0]["content"][0]["text"] == "Blocked by policy." assert completed["usage"] == {"input_tokens": 7, "output_tokens": 21, "total_tokens": 28} + def test_continuation_numbers_every_event_after_the_last_one_relayed(self): + yielded = [ + {"type": "response.created", "sequence_number": 0, "response": {"id": "resp_live", "output": []}}, + { + "type": "response.output_item.added", + "sequence_number": 1, + "output_index": 0, + "item": {"id": "msg_orig", "type": "message", "role": "assistant"}, + }, + { + "type": "response.output_text.delta", + "sequence_number": 2, + "item_id": "msg_orig", + "output_index": 0, + "content_index": 0, + "delta": "partial", + }, + ] + + payloads = self._payloads( + OpenAIResponsesHandler().build_block_sse_chunks( + self._exc(original_response=yielded), stream_started=True, responses_so_far=yielded + ) + ) + + assert payloads[-1]["type"] == "response.completed" + assert [payload.get("sequence_number") for payload in payloads] == list(range(3, 3 + len(payloads))) + def test_continuation_reads_usage_from_typed_completed_event(self): from litellm.types.llms.openai import ( ResponseCompletedEvent, diff --git a/tests/unit/responses/litellm_completion_transformation/test_streaming_iterator_transformation.py b/tests/unit/responses/litellm_completion_transformation/test_streaming_iterator_transformation.py index 86773dc7ddf..9c9b3fc4809 100644 --- a/tests/unit/responses/litellm_completion_transformation/test_streaming_iterator_transformation.py +++ b/tests/unit/responses/litellm_completion_transformation/test_streaming_iterator_transformation.py @@ -477,7 +477,6 @@ def test_tool_call_arguments_are_chunked_to_match_openai_behavior(): assert evt is not None assert evt.type == ResponsesAPIStreamEvents.OUTPUT_ITEM_ADDED assert evt.output_index == 0 - assert hasattr(evt, "__dict__") and "sequence_number" in evt.__dict__ # Collect all remaining delta events from the pending queue by creating empty chunks delta_events = [] @@ -511,17 +510,11 @@ def test_tool_call_arguments_are_chunked_to_match_openai_behavior(): assert len(evt.delta) <= 10 assert evt.item_id == "fc_call_test" assert evt.output_index == 0 - assert hasattr(evt, "__dict__") and "sequence_number" in evt.__dict__ # Verify all deltas concatenated equal the original arguments concatenated = "".join(evt.delta for evt in delta_events) assert concatenated == large_arguments - # Verify sequence numbers are increasing - sequence_numbers = [evt.__dict__["sequence_number"] for evt in delta_events] - assert sequence_numbers == sorted(sequence_numbers) - assert len(set(sequence_numbers)) == len(sequence_numbers) # All unique - def test_tool_call_delta_without_id_uses_index_mapping(): iterator = LiteLLMCompletionStreamingIterator( @@ -1988,3 +1981,50 @@ def test_the_partial_response_billed_on_disconnect_carries_the_logging_objects_b partial.response.usage.input_tokens, partial.response.usage.output_tokens, ), "the bill is priced on the partial's own usage" + + +def _serialized_sequence_numbers(events: Iterable[BaseLiteLLMOpenAIResponseObject]) -> list[int]: + payloads: Final = tuple(json.loads(event.model_dump_json()) for event in events) + assert all("sequence_number" in payload for payload in payloads), [ + payload["type"] for payload in payloads if "sequence_number" not in payload + ] + return [payload["sequence_number"] for payload in payloads] + + +def _long_text_chunks() -> tuple[ModelResponseStream, ...]: + return (*(_chunk(f"word{i} ") for i in range(300)), _chunk("end", finish_reason="stop")) + + +def _text_then_tool_call_chunks() -> tuple[ModelResponseStream, ...]: + return (_text_opening_a_tool_call_chunk("Checking. "), _tool_arguments_chunk('{"x": 1}', finish_reason="tool_calls")) + + +@pytest.mark.parametrize( + "stream", + [ + lambda: (_chunk("Hello"), _chunk("!", finish_reason="stop")), + _long_text_chunks, + _text_then_tool_call_chunks, + _reasoning_then_text_chunks, + _reasoning_and_text_in_one_chunk, + _reasoning_citation_then_tool_call_finishing_chunks, + lambda: (_tool_call_chunk(), _chunk("", finish_reason="tool_calls")), + ], + ids=["text", "long_text", "text_then_tool", "reasoning_then_text", "reasoning_and_text", "citation_tool", "tool"], +) +@pytest.mark.parametrize("sync_mode", [True, False]) +@pytest.mark.asyncio +async def test_every_event_serializes_its_position_in_the_stream_from_zero(stream: StreamChunks, sync_mode: bool): + iterator: Final = _build_iterator(stream()) + events: Final = list(iterator) if sync_mode else [event async for event in iterator] + + assert _serialized_sequence_numbers(events) == list(range(len(events))) + + +def test_bridged_events_answer_dict_reads_of_their_number(): + events: Final = list(_build_iterator((_chunk("Hello"), _chunk("!", finish_reason="stop")))) + positions: Final = list(range(len(events))) + + assert [event["sequence_number"] for event in events] == positions + assert [event.get("sequence_number") for event in events] == positions + assert all("sequence_number" in event for event in events) diff --git a/tests/unit/responses/test_streaming_iterator.py b/tests/unit/responses/test_streaming_iterator.py index b9454ae6c1a..46389878423 100644 --- a/tests/unit/responses/test_streaming_iterator.py +++ b/tests/unit/responses/test_streaming_iterator.py @@ -2835,3 +2835,52 @@ async def test_websocket_turns_logging_survives_garbage_collection(terminal_even await _run_websocket_turn(logger, terminal_event) assert await _logging_tasks_left_after_garbage_collection(logger) == 1 + + +@pytest.mark.parametrize( + "output", + [ + [ + { + "type": "message", + "id": "msg_seq", + "status": "completed", + "role": "assistant", + "content": [{"type": "output_text", "text": "hello world", "annotations": []}], + } + ], + [ + { + "type": "reasoning", + "id": "rs_seq", + "summary": [{"type": "summary_text", "text": "thinking about it"}], + }, + { + "type": "function_call", + "id": "fc_seq", + "call_id": "call_seq", + "name": "lookup", + "arguments": '{"id":1}', + }, + ], + ], + ids=["message", "reasoning_and_tool"], +) +def test_synthetic_stream_serializes_every_event_position_from_zero(output: list[dict[str, object]]): + transformed: Final = ResponsesAPIResponse( + id="resp_seq", + created_at=int(datetime.now().timestamp()), + status="completed", + model="claude-sonnet-5-5", + object="response", + output=output, + ) + + events: Final = streaming_module.build_synthetic_response_events( + transformed=transformed, + logging_obj=None, + chunk_size=5, + ) + + payloads: Final = tuple(json.loads(event.model_dump_json()) for event in events) + assert [payload.get("sequence_number") for payload in payloads] == list(range(len(payloads))) diff --git a/tests/unit/test_router/test_router.py b/tests/unit/test_router/test_router.py index 7ac27e22a93..aa4b7866eba 100644 --- a/tests/unit/test_router/test_router.py +++ b/tests/unit/test_router/test_router.py @@ -6776,6 +6776,44 @@ async def test_aresponses_streaming_content_policy_error_event_routes_to_content assert fallback_call.await_args.kwargs["model"] == "gemini/gemini-2.5-flash" +@pytest.mark.asyncio +async def test_aresponses_stream_keeps_numbering_after_a_mid_stream_fallback(): + from litellm.types.llms.openai import OutputTextDeltaEvent, ResponsesAPIStreamEvents + + def delta(sequence_number: int, text: str) -> OutputTextDeltaEvent: + return OutputTextDeltaEvent( + type=ResponsesAPIStreamEvents.OUTPUT_TEXT_DELTA, + item_id="msg_1", + output_index=0, + content_index=0, + delta=text, + sequence_number=sequence_number, + ) + + from litellm.exceptions import MidStreamFallbackError + + router = _InjectedFallbackRouter(_AsyncList([delta(0, "c"), delta(1, "d")])) + primary = _make_responses_iterator( + chunks=[delta(0, "a"), delta(1, "b")], + error=MidStreamFallbackError( + message="primary died", + model="primary", + llm_provider="anthropic", + is_pre_first_chunk=False, + generated_content="ab", + ), + model="primary", + ) + + wrapped = await router._aresponses_streaming_iterator( + response=primary, + initial_kwargs={"model": "primary", "stream": True, "input": "Hi", "original_generic_function": litellm.aresponses}, + ) + events = [event async for event in wrapped] + + assert [(event.delta, event.sequence_number) for event in events] == [("a", 0), ("b", 1), ("c", 2), ("d", 3)] + + @pytest.mark.asyncio async def test_aresponses_streaming_iterator_unwraps_content_policy_trigger_for_fallback_dispatch(): """The fallback dispatch matches on the trigger's own type, so the wrapper must hand it the diff --git a/tests/unit/types/llms/test_types_llms_base.py b/tests/unit/types/llms/test_types_llms_base.py index 964689cdeeb..07b641fb392 100644 --- a/tests/unit/types/llms/test_types_llms_base.py +++ b/tests/unit/types/llms/test_types_llms_base.py @@ -9,7 +9,7 @@ from typing import Final import pytest from pydantic import ConfigDict, create_model -from litellm.types.llms.base import LiteLLMBaseModel +from litellm.types.llms.base import BaseLiteLLMOpenAIResponseObject, LiteLLMBaseModel _PROBE: Final = """ from litellm.types.llms.base import LiteLLMBaseModel @@ -142,3 +142,30 @@ def test_concurrent_first_use_of_a_deferred_subclass_always_builds_that_subclass assert _concurrent_first_use_outcomes(child) == {child.__name__}, f"round {round_id}" finally: sys.setswitchinterval(previous_switch_interval) + + +class _ResponseEvent(BaseLiteLLMOpenAIResponseObject): + type: str + + +def test_response_object_answers_dict_reads_for_undeclared_fields(): + event: Final = _ResponseEvent(type="response.created", sequence_number=3) + + assert event["type"] == "response.created" + assert event["sequence_number"] == 3 + assert event.get("sequence_number") == 3 + assert "sequence_number" in event + assert dict(event.items()) == {"type": "response.created", "sequence_number": 3} + + +def test_response_object_answers_dict_reads_for_an_undeclared_field_named_keys(): + assert _ResponseEvent(type="response.created", keys="value")["keys"] == "value" + + +def test_response_object_dict_reads_still_miss_absent_keys(): + event: Final = _ResponseEvent(type="response.created") + + assert event.get("sequence_number", "fallback") == "fallback" + assert "sequence_number" not in event + with pytest.raises(KeyError): + event["sequence_number"]