Skip to content

A programming framework for agentic AI

FEAT: SelectorGroupChat could using stream inner select_prompt (#6286)

EeS9b0a0bd6b867
diff --git a/python/packages/autogen-agentchat/src/autogen_agentchat/messages.py b/python/packages/autogen-agentchat/src/autogen_agentchat/messages.py
index 206bcc6e4..ac21fbaa2 100644
--- a/python/packages/autogen-agentchat/src/autogen_agentchat/messages.py
+++ b/python/packages/autogen-agentchat/src/autogen_agentchat/messages.py
@@ -531,6 +531,18 @@ class SelectSpeakerEvent(BaseAgentEvent):
return str(self.content)
+class SelectorEvent(BaseAgentEvent):
+ """An event emitted from the `SelectorGroupChat`."""
+
+ content: str
+ """The content of the event."""
+
+ type: Literal["SelectorEvent"] = "SelectorEvent"
+
+ def to_text(self) -> str:
+ return str(self.content)
+
+
class MessageFactory:
""":meta private:
diff --git a/python/packages/autogen-agentchat/src/autogen_agentchat/teams/_group_chat/_selector_group_chat.py b/python/packages/autogen-agentchat/src/autogen_agentchat/teams/_group_chat/_selector_group_chat.py
index 2a7b15889..2af568ac4 100644
--- a/python/packages/autogen-agentchat/src/autogen_agentchat/teams/_group_chat/_selector_group_chat.py
+++ b/python/packages/autogen-agentchat/src/autogen_agentchat/teams/_group_chat/_selector_group_chat.py
@@ -5,7 +5,14 @@ from inspect import iscoroutinefunction
from typing import Any, Awaitable, Callable, Dict, List, Mapping, Optional, Sequence, Union, cast
from autogen_core import AgentRuntime, Component, ComponentModel
-from autogen_core.models import AssistantMessage, ChatCompletionClient, ModelFamily, SystemMessage, UserMessage
+from autogen_core.models import (
+ AssistantMessage,
+ ChatCompletionClient,
+ CreateResult,
+ ModelFamily,
+ SystemMessage,
+ UserMessage,
+)
from pydantic import BaseModel
from typing_extensions import Self
@@ -16,6 +23,8 @@ from ...messages import (
BaseAgentEvent,
BaseChatMessage,
MessageFactory,
+ ModelClientStreamingChunkEvent,
+ SelectorEvent,
)
from ...state import SelectorManagerState
from ._base_group_chat import BaseGroupChat
@@ -56,6 +65,7 @@ class SelectorGroupChatManager(BaseGroupChatManager):
max_selector_attempts: int,
candidate_func: Optional[CandidateFuncType],
emit_team_events: bool,
+ model_client_streaming: bool = False,
) -> None:
super().__init__(
name,
@@ -79,6 +89,7 @@ class SelectorGroupChatManager(BaseGroupChatManager):
self._max_selector_attempts = max_selector_attempts
self._candidate_func = candidate_func
self._is_candidate_func_async = iscoroutinefunction(self._candidate_func)
+ self._model_client_streaming = model_client_streaming
async def validate_group_state(self, messages: List[BaseChatMessage] | None) -> None:
pass
@@ -194,7 +205,26 @@ class SelectorGroupChatManager(BaseGroupChatManager):
num_attempts = 0
while num_attempts < max_attempts:
num_attempts += 1
- response = await self._model_client.create(messages=select_speaker_messages)
+ if self._model_client_streaming:
+ chunk: CreateResult | str = ""
+ async for _chunk in self._model_client.create_stream(messages=select_speaker_messages):
+ chunk = _chunk
+ if self._emit_team_events:
+ if isinstance(chunk, str):
+ await self._output_message_queue.put(
+ ModelClientStreamingChunkEvent(content=cast(str, _chunk), source=self._name)
+ )
+ else:
+ assert isinstance(chunk, CreateResult)
+ assert isinstance(chunk.content, str)
+ await self._output_message_queue.put(
+ SelectorEvent(content=chunk.content, source=self._name)
+ )
+ # The last chunk must be CreateResult.
+ assert isinstance(chunk, CreateResult)
+ response = chunk
+ else:
+ response = await self._model_client.create(messages=select_speaker_messages)
assert isinstance(response.content, str)
select_speaker_messages.append(AssistantMessage(content=response.content, source="selector"))
# NOTE: we use all participant names to check for mentions, even if the previous speaker is not allowed.
@@ -281,6 +311,7 @@ class SelectorGroupChatConfig(BaseModel):
# selector_func: ComponentModel | None
max_selector_attempts: int = 3
emit_team_events: bool = False
+ model_client_streaming: bool = False
class SelectorGroupChat(BaseGroupChat, Component[SelectorGroupChatConfig]):
@@ -311,6 +342,7 @@ class SelectorGroupChat(BaseGroupChat, Component[SelectorGroupChatConfig]):
selection using model. If the function returns an empty list or `None`, `SelectorGroupChat` will raise a `ValueError`.
This function is only used if `selector_func` is not set. The `allow_repeated_speaker` will be ignored if set.
emit_team_events (bool, optional): Whether to emit team events through :meth:`BaseGroupChat.run_stream`. Defaults to False.
+ model_client_streaming (bool, optional): Whether to use streaming for the model client. (This is useful for reasoning models like QwQ). Defaults to False.
Raises:
ValueError: If the number of participants is less than two or if the selector prompt is invalid.
@@ -453,6 +485,7 @@ Read the above conversation. Then select the next role from {participants} to pl
candidate_func: Optional[CandidateFuncType] = None,
custom_message_types: List[type[BaseAgentEvent | BaseChatMessage]] | None = None,
emit_team_events: bool = False,
+ model_client_streaming: bool = False,
):
super().__init__(
participants,
@@ -473,6 +506,7 @@ Read the above conversation. Then select the next role from {participants} to pl
self._selector_func = selector_func
self._max_selector_attempts = max_selector_attempts
self._candidate_func = candidate_func
+ self._model_client_streaming = model_client_streaming
def _create_group_chat_manager_factory(
self,
@@ -505,6 +539,7 @@ Read the above conversation. Then select the next role from {participants} to pl
self._max_selector_attempts,
self._candidate_func,
self._emit_team_events,
+ self._model_client_streaming,
)
def _to_config(self) -> SelectorGroupChatConfig:
@@ -518,6 +553,7 @@ Read the above conversation. Then select the next role from {participants} to pl
max_selector_attempts=self._max_selector_attempts,
# selector_func=self._selector_func.dump_component() if self._selector_func else None,
emit_team_events=self._emit_team_events,
+ model_client_streaming=self._model_client_streaming,
)
@classmethod
@@ -536,4 +572,5 @@ Read the above conversation. Then select the next role from {participants} to pl
# if config.selector_func
# else None,
emit_team_events=config.emit_team_events,
+ model_client_streaming=config.model_client_streaming,
)
diff --git a/python/packages/autogen-agentchat/tests/test_group_chat.py b/python/packages/autogen-agentchat/tests/test_group_chat.py
index 947d9595b..971dec1ce 100644
--- a/python/packages/autogen-agentchat/tests/test_group_chat.py
+++ b/python/packages/autogen-agentchat/tests/test_group_chat.py
@@ -13,12 +13,19 @@ from autogen_agentchat.agents import (
CodeExecutorAgent,
)
from autogen_agentchat.base import Handoff, Response, TaskResult, TerminationCondition
-from autogen_agentchat.conditions import HandoffTermination, MaxMessageTermination, TextMentionTermination
+from autogen_agentchat.conditions import (
+ HandoffTermination,
+ MaxMessageTermination,
+ StopMessageTermination,
+ TextMentionTermination,
+)
from autogen_agentchat.messages import (
BaseAgentEvent,
BaseChatMessage,
HandoffMessage,
+ ModelClientStreamingChunkEvent,
MultiModalMessage,
+ SelectorEvent,
SelectSpeakerEvent,
StopMessage,
StructuredMessage,
@@ -1698,3 +1705,54 @@ async def test_structured_message_state_roundtrip(runtime: AgentRuntime | None)
)
assert manager1._message_thread == manager2._message_thread # pyright: ignore
+
+
+async def test_selector_group_chat_streaming(runtime: AgentRuntime | None) -> None:
+ model_client = ReplayChatCompletionClient(
+ ["the agent should be agent2"],
+ )
+ agent2 = _StopAgent("agent2", description="stop agent 2", stop_at=0)
+ agent3 = _EchoAgent("agent3", description="echo agent 3")
+ termination = StopMessageTermination()
+ team = SelectorGroupChat(
+ participants=[agent2, agent3],
+ model_client=model_client,
+ termination_condition=termination,
+ runtime=runtime,
+ emit_team_events=True,
+ model_client_streaming=True,
+ )
+ result = await team.run(
+ task="Write a program that prints 'Hello, world!'",
+ )
+
+ assert len(result.messages) == 4
+ assert isinstance(result.messages[0], TextMessage)
+ assert isinstance(result.messages[1], SelectorEvent)
+ assert isinstance(result.messages[2], SelectSpeakerEvent)
+ assert isinstance(result.messages[3], StopMessage)
+
+ assert result.messages[0].content == "Write a program that prints 'Hello, world!'"
+ assert result.messages[1].content == "the agent should be agent2"
+ assert result.messages[2].content == ["agent2"]
+ assert result.messages[3].source == "agent2"
+ assert result.stop_reason is not None and result.stop_reason == "Stop message received"
+
+ # Test streaming
+ await team.reset()
+ model_client.reset()
+ index = 0
+ streaming: List[str] = []
+ async for message in team.run_stream(task="Write a program that prints 'Hello, world!'"):
+ if isinstance(message, TaskResult):
+ assert message == result
+ elif isinstance(message, ModelClientStreamingChunkEvent):
+ streaming.append(message.content)
+ else:
+ if streaming:
+ assert isinstance(message, SelectorEvent)
+ assert message.content == "".join([chunk for chunk in streaming])
+ streaming = []
+ assert message == result.messages[index]
+ index += 1