feat: add 'processing' agent state for tool calling

livekit-agents/livekit/agents/voice/agent_activity.py

@@ -2055,7 +2055,7 @@ def _tool_execution_completed_cb(out: ToolExecutionOutput) -> None:
             current_span.set_attribute(trace_types.ATTR_RESPONSE_TEXT, text_out.text)
 
         if len(tool_output.output) > 0:
-            self._session._update_agent_state("thinking")
+            self._session._update_agent_state("processing")
         elif self._session.agent_state == "speaking":
             self._session._update_agent_state("listening")
 
@@ -2545,7 +2545,7 @@ def _create_assistant_message(
 
         speech_handle._mark_generation_done()  # mark the playout done before waiting for the tool execution  # noqa: E501
         tool_output.first_tool_started_fut.add_done_callback(
-            lambda _: self._session._update_agent_state("thinking")
+            lambda _: self._session._update_agent_state("processing")
         )
 
         self._background_speeches.add(speech_handle)

livekit-agents/livekit/agents/voice/events.py

@@ -95,7 +95,7 @@ async def wait_for_playout(self) -> None:
 ]
 
 UserState = Literal["speaking", "listening", "away"]
-AgentState = Literal["initializing", "idle", "listening", "thinking", "speaking"]
+AgentState = Literal["initializing", "idle", "listening", "thinking", "speaking", "processing"]
 
 
 class UserStateChangedEvent(BaseModel):

tests/test_agent_session.py

@@ -177,17 +177,20 @@ async def test_tool_call() -> None:
     check_timestamp(playback_finished_events[0].playback_position, 2.0, speed_factor=speed)
     check_timestamp(playback_finished_events[1].playback_position, 3.0, speed_factor=speed)
 
-    assert len(agent_state_events) == 6
+    assert len(agent_state_events) == 7
     assert agent_state_events[0].old_state == "initializing"
     assert agent_state_events[0].new_state == "listening"
     assert agent_state_events[1].new_state == "thinking"
     assert agent_state_events[2].new_state == "speaking"
     assert (
-        agent_state_events[3].new_state == "thinking"
-    )  # from speaking to thinking when tool call is executed
+        agent_state_events[3].new_state == "processing"
+    )  # from speaking to processing when tool call is executed
     check_timestamp(agent_state_events[3].created_at - t_origin, 5.5, speed_factor=speed)
-    assert agent_state_events[4].new_state == "speaking"
-    assert agent_state_events[5].new_state == "listening"
+    assert (
+        agent_state_events[4].new_state == "thinking"
+    )  # from processing to thinking for LLM inference
+    assert agent_state_events[5].new_state == "speaking"
+    assert agent_state_events[6].new_state == "listening"
 
     assert len(tool_executed_events) == 1
     assert tool_executed_events[0].function_calls[0].name == "get_weather"