fix(live): keep streaming tool yields open

Author: he-yufeng

src/google/adk/agents/live_request_queue.py

@@ -55,6 +55,9 @@ class LiveRequest(BaseModel):
   close: bool = False
   """If set, close the queue. queue.shutdown() is only supported in Python 3.13+."""
 
+  turn_complete: bool = True
+  """If set, content messages complete the current model turn."""
+
 
 class LiveRequestQueue:
   """Queue used to send LiveRequest in a live(bidirectional streaming) way."""
@@ -65,8 +68,10 @@ def __init__(self):
   def close(self):
     self._queue.put_nowait(LiveRequest(close=True))
 
-  def send_content(self, content: types.Content):
-    self._queue.put_nowait(LiveRequest(content=content))
+  def send_content(self, content: types.Content, turn_complete: bool = True):
+    self._queue.put_nowait(
+        LiveRequest(content=content, turn_complete=turn_complete)
+    )
 
   def send_realtime(self, blob: types.Blob):
     self._queue.put_nowait(LiveRequest(blob=blob))

src/google/adk/flows/llm_flows/base_llm_flow.py

@@ -770,9 +770,9 @@ async def _send_to_model(
         is_function_response = content.parts and any(
             part.function_response for part in content.parts
         )
-        if not is_function_response:
-          if not content.role:
-            content.role = 'user'
+        if not is_function_response and not content.role:
+          content.role = 'user'
+        if not is_function_response and live_request.turn_complete:
           user_content_event = Event(
               id=Event.new_id(),
               invocation_id=invocation_context.invocation_id,
@@ -783,7 +783,9 @@ async def _send_to_model(
               session=invocation_context.session,
               event=user_content_event,
           )
-        await llm_connection.send_content(live_request.content)
+        await llm_connection.send_content(
+            live_request.content, turn_complete=live_request.turn_complete
+        )
 
   async def _receive_from_model(
       self,

src/google/adk/flows/llm_flows/functions.py

@@ -985,7 +985,9 @@ async def run_tool_and_update_queue(tool, function_args, tool_context):
                     )
                 ],
             )
-            invocation_context.live_request_queue.send_content(updated_content)
+            invocation_context.live_request_queue.send_content(
+                updated_content, turn_complete=False
+            )
       except asyncio.CancelledError:
         raise  # Re-raise to properly propagate the cancellation
 

src/google/adk/models/base_llm_connection.py

@@ -39,15 +39,18 @@ async def send_history(self, history: list[types.Content]):
     pass
 
   @abstractmethod
-  async def send_content(self, content: types.Content):
+  async def send_content(
+      self, content: types.Content, turn_complete: bool = True
+  ):
     """Sends a user content to the model.
 
-    The model will respond immediately upon receiving the content.
+    By default, the model will respond upon receiving the content.
     If you send function responses, all parts in the content should be function
     responses.
 
     Args:
       content: The content to send to the model.
+      turn_complete: Whether this content completes the model turn.
     """
     pass
 

src/google/adk/models/gemini_llm_connection.py

@@ -117,15 +117,18 @@ async def send_history(self, history: list[types.Content]):
     else:
       logger.info('no content is sent')
 
-  async def send_content(self, content: types.Content):
+  async def send_content(
+      self, content: types.Content, turn_complete: bool = True
+  ):
     """Sends a user content to the gemini model.
 
-    The model will respond immediately upon receiving the content.
+    By default, the model will respond upon receiving the content.
     If you send function responses, all parts in the content should be function
     responses.
 
     Args:
       content: The content to send to the model.
+      turn_complete: Whether this content completes the model turn.
     """
     assert content.parts
     if content.parts[0].function_response:
@@ -138,7 +141,8 @@ async def send_content(self, content: types.Content):
     else:
       logger.debug('Sending LLM new content %s', content)
       if (
-          self._is_gemini_3_1_flash_live
+          turn_complete
+          and self._is_gemini_3_1_flash_live
           and len(content.parts) == 1
           and content.parts[0].text
       ):
@@ -150,7 +154,7 @@ async def send_content(self, content: types.Content):
         await self._gemini_session.send(
             input=types.LiveClientContent(
                 turns=[content],
-                turn_complete=True,
+                turn_complete=turn_complete,
             )
         )
 

tests/unittests/agents/test_live_request_queue.py

@@ -40,6 +40,17 @@ def test_send_content():
     mock_put_nowait.assert_called_once_with(LiveRequest(content=content))
 
 
+def test_send_content_sets_turn_complete():
+  queue = LiveRequestQueue()
+  content = MagicMock(spec=types.Content)
+
+  with patch.object(queue._queue, "put_nowait") as mock_put_nowait:
+    queue.send_content(content, turn_complete=False)
+    mock_put_nowait.assert_called_once_with(
+        LiveRequest(content=content, turn_complete=False)
+    )
+
+
 def test_send_realtime():
   queue = LiveRequestQueue()
   blob = MagicMock(spec=types.Blob)

tests/unittests/flows/llm_flows/test_base_llm_flow_realtime.py

@@ -197,5 +197,36 @@ async def test_send_to_model_with_text_content(mock_llm_connection):
   await flow._send_to_model(mock_llm_connection, invocation_context)
 
   # Verify send_content was called instead of send_realtime
-  mock_llm_connection.send_content.assert_called_once_with(content)
+  mock_llm_connection.send_content.assert_called_once_with(
+      content, turn_complete=True
+  )
   mock_llm_connection.send_realtime.assert_not_called()
+
+
+@pytest.mark.asyncio
+async def test_send_to_model_with_intermediate_text_content(
+    mock_llm_connection,
+):
+  agent = Agent(name='test_agent', model='mock')
+  invocation_context = await testing_utils.create_invocation_context(
+      agent=agent, user_content=''
+  )
+  invocation_context.live_request_queue = LiveRequestQueue()
+  invocation_context.session_service.append_event = mock.AsyncMock()
+
+  flow = TestBaseLlmFlow()
+
+  content = types.Content(
+      role='user', parts=[types.Part.from_text(text='progress')]
+  )
+  invocation_context.live_request_queue.send(
+      LiveRequest(content=content, turn_complete=False)
+  )
+  invocation_context.live_request_queue.close()
+
+  await flow._send_to_model(mock_llm_connection, invocation_context)
+
+  mock_llm_connection.send_content.assert_called_once_with(
+      content, turn_complete=False
+  )
+  invocation_context.session_service.append_event.assert_not_called()

tests/unittests/models/test_gemini_llm_connection.py

@@ -124,6 +124,22 @@ async def test_send_content_text(gemini_connection, mock_gemini_session):
   assert call_args['input'].turn_complete is True
 
 
+@pytest.mark.asyncio
+async def test_send_content_text_can_keep_turn_open(
+    gemini_connection, mock_gemini_session
+):
+  content = types.Content(
+      role='user', parts=[types.Part.from_text(text='progress')]
+  )
+
+  await gemini_connection.send_content(content, turn_complete=False)
+
+  mock_gemini_session.send.assert_called_once()
+  call_args = mock_gemini_session.send.call_args[1]
+  assert call_args['input'].turns == [content]
+  assert call_args['input'].turn_complete is False
+
+
 @pytest.mark.asyncio
 async def test_send_content_function_response(
     gemini_connection, mock_gemini_session

tests/unittests/testing_utils.py

@@ -420,7 +420,9 @@ def __init__(self, llm_responses: list[LlmResponse]):
   async def send_history(self, history: list[types.Content]):
     pass
 
-  async def send_content(self, content: types.Content):
+  async def send_content(
+      self, content: types.Content, turn_complete: bool = True
+  ):
     pass
 
   async def send(self, data):

tests/unittests/workflow/testing_utils.py

@@ -484,7 +484,9 @@ def __init__(
   async def send_history(self, history: list[types.Content]):
     pass
 
-  async def send_content(self, content: types.Content):
+  async def send_content(
+      self, content: types.Content, turn_complete: bool = True
+  ):
     self.mock_model.live_contents.append(content)
     self._input_event.set()