Replace referencable_id with response_id (openai#508)

Minor change - naming. So that it doesn't pollute the next PR. --- [//]: # (BEGIN SAPLING FOOTER) * openai#509 * __->__ openai#508

Author: rm-openai

src/agents/items.py

@@ -166,7 +166,7 @@ class ModelResponse:
  usage: Usage
  """The usage information for the response."""
 
- referenceable_id: str | None
+ response_id: str | None
  """An ID for the response which can be used to refer to the response in subsequent calls to the
  model. Not supported by all model providers.
  """

src/agents/models/openai_chatcompletions.py

@@ -156,7 +156,7 @@ async def get_response(
  return ModelResponse(
  output=items,
  usage=usage,
- referenceable_id=None,
+ response_id=None,
  )
 
  async def stream_response(

src/agents/models/openai_responses.py

@@ -120,7 +120,7 @@ async def get_response(
  return ModelResponse(
  output=response.output,
  usage=usage,
- referenceable_id=response.id,
+ response_id=response.id,
  )
 
  async def stream_response(

src/agents/run.py

@@ -677,7 +677,7 @@ async def _run_single_turn_streamed(
  final_response = ModelResponse(
  output=event.response.output,
  usage=usage,
- referenceable_id=event.response.id,
+ response_id=event.response.id,
  )
 
  streamed_result._event_queue.put_nowait(RawResponsesStreamEvent(data=event))

tests/fake_model.py

@@ -81,7 +81,7 @@ async def get_response(
  return ModelResponse(
  output=output,
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
 
  async def stream_response(

tests/test_items_helpers.py

@@ -168,7 +168,7 @@ def test_to_input_items_for_message() -> None:
  message = ResponseOutputMessage(
  id="m1", content=[content], role="assistant", status="completed", type="message"
  )
- resp = ModelResponse(output=[message], usage=Usage(), referenceable_id=None)
+ resp = ModelResponse(output=[message], usage=Usage(), response_id=None)
  input_items = resp.to_input_items()
  assert isinstance(input_items, list) and len(input_items) == 1
  # The dict should contain exactly the primitive values of the message
@@ -193,7 +193,7 @@ def test_to_input_items_for_function_call() -> None:
  tool_call = ResponseFunctionToolCall(
  id="f1", arguments="{}", call_id="c1", name="func", type="function_call"
  )
- resp = ModelResponse(output=[tool_call], usage=Usage(), referenceable_id=None)
+ resp = ModelResponse(output=[tool_call], usage=Usage(), response_id=None)
  input_items = resp.to_input_items()
  assert isinstance(input_items, list) and len(input_items) == 1
  expected: ResponseFunctionToolCallParam = {
@@ -211,7 +211,7 @@ def test_to_input_items_for_file_search_call() -> None:
  fs_call = ResponseFileSearchToolCall(
  id="fs1", queries=["query"], status="completed", type="file_search_call"
  )
- resp = ModelResponse(output=[fs_call], usage=Usage(), referenceable_id=None)
+ resp = ModelResponse(output=[fs_call], usage=Usage(), response_id=None)
  input_items = resp.to_input_items()
  assert isinstance(input_items, list) and len(input_items) == 1
  expected: ResponseFileSearchToolCallParam = {
@@ -226,7 +226,7 @@ def test_to_input_items_for_file_search_call() -> None:
 def test_to_input_items_for_web_search_call() -> None:
  """A web search tool call output should produce the same dict as a web search input."""
  ws_call = ResponseFunctionWebSearch(id="w1", status="completed", type="web_search_call")
- resp = ModelResponse(output=[ws_call], usage=Usage(), referenceable_id=None)
+ resp = ModelResponse(output=[ws_call], usage=Usage(), response_id=None)
  input_items = resp.to_input_items()
  assert isinstance(input_items, list) and len(input_items) == 1
  expected: ResponseFunctionWebSearchParam = {
@@ -248,7 +248,7 @@ def test_to_input_items_for_computer_call_click() -> None:
  pending_safety_checks=[],
  status="completed",
  )
- resp = ModelResponse(output=[comp_call], usage=Usage(), referenceable_id=None)
+ resp = ModelResponse(output=[comp_call], usage=Usage(), response_id=None)
  input_items = resp.to_input_items()
  assert isinstance(input_items, list) and len(input_items) == 1
  converted_dict = input_items[0]
@@ -268,7 +268,7 @@ def test_to_input_items_for_reasoning() -> None:
  """A reasoning output should produce the same dict as a reasoning input item."""
  rc = Summary(text="why", type="summary_text")
  reasoning = ResponseReasoningItem(id="rid1", summary=[rc], type="reasoning")
- resp = ModelResponse(output=[reasoning], usage=Usage(), referenceable_id=None)
+ resp = ModelResponse(output=[reasoning], usage=Usage(), response_id=None)
  input_items = resp.to_input_items()
  assert isinstance(input_items, list) and len(input_items) == 1
  converted_dict = input_items[0]

tests/test_openai_chatcompletions.py

@@ -80,7 +80,7 @@ async def patched_fetch_response(self, *args, **kwargs):
  assert resp.usage.input_tokens == 7
  assert resp.usage.output_tokens == 5
  assert resp.usage.total_tokens == 12
- assert resp.referenceable_id is None
+ assert resp.response_id is None
 
 
 @pytest.mark.allow_call_model_methods

tests/test_run_step_execution.py

@@ -43,7 +43,7 @@ async def test_empty_response_is_final_output():
  response = ModelResponse(
  output=[],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = await get_execute_result(agent, response)
 
@@ -59,7 +59,7 @@ async def test_plaintext_agent_no_tool_calls_is_final_output():
  response = ModelResponse(
  output=[get_text_message("hello_world")],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = await get_execute_result(agent, response)
 
@@ -79,7 +79,7 @@ async def test_plaintext_agent_no_tool_calls_multiple_messages_is_final_output()
  get_text_message("bye"),
  ],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = await get_execute_result(
  agent,
@@ -105,7 +105,7 @@ async def test_plaintext_agent_with_tool_call_is_run_again():
  response = ModelResponse(
  output=[get_text_message("hello_world"), get_function_tool_call("test", "")],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = await get_execute_result(agent, response)
 
@@ -140,7 +140,7 @@ async def test_multiple_tool_calls():
  get_function_tool_call("test_2"),
  ],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
 
  result = await get_execute_result(agent, response)
@@ -166,7 +166,7 @@ async def test_handoff_output_leads_to_handoff_next_step():
  response = ModelResponse(
  output=[get_text_message("Hello, world!"), get_handoff_tool_call(agent_1)],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = await get_execute_result(agent_3, response)
 
@@ -186,7 +186,7 @@ async def test_final_output_without_tool_runs_again():
  response = ModelResponse(
  output=[get_function_tool_call("tool_1")],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = await get_execute_result(agent, response)
 
@@ -203,7 +203,7 @@ async def test_final_output_leads_to_final_output_next_step():
  get_final_output_message(Foo(bar="123").model_dump_json()),
  ],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = await get_execute_result(agent, response)
 
@@ -222,7 +222,7 @@ async def test_handoff_and_final_output_leads_to_handoff_next_step():
  get_handoff_tool_call(agent_1),
  ],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = await get_execute_result(agent_3, response)
 
@@ -241,7 +241,7 @@ async def test_multiple_final_output_leads_to_final_output_next_step():
  get_final_output_message(Foo(bar="456").model_dump_json()),
  ],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = await get_execute_result(agent_3, response)
 

tests/test_run_step_processing.py

@@ -39,7 +39,7 @@ def test_empty_response():
  response = ModelResponse(
  output=[],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
 
  result = RunImpl.process_model_response(
@@ -58,7 +58,7 @@ def test_no_tool_calls():
  response = ModelResponse(
  output=[get_text_message("Hello, world!")],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = RunImpl.process_model_response(
  agent=agent, response=response, output_schema=None, handoffs=[], all_tools=[]
@@ -76,7 +76,7 @@ async def test_single_tool_call():
  get_function_tool_call("test", ""),
  ],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = RunImpl.process_model_response(
  agent=agent,
@@ -102,7 +102,7 @@ async def test_missing_tool_call_raises_error():
  get_function_tool_call("missing", ""),
  ],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
 
  with pytest.raises(ModelBehaviorError):
@@ -132,7 +132,7 @@ async def test_multiple_tool_calls():
  get_function_tool_call("test_2", "xyz"),
  ],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
 
  result = RunImpl.process_model_response(
@@ -162,7 +162,7 @@ async def test_handoffs_parsed_correctly():
  response = ModelResponse(
  output=[get_text_message("Hello, world!")],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = RunImpl.process_model_response(
  agent=agent_3,
@@ -176,7 +176,7 @@ async def test_handoffs_parsed_correctly():
  response = ModelResponse(
  output=[get_text_message("Hello, world!"), get_handoff_tool_call(agent_1)],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = RunImpl.process_model_response(
  agent=agent_3,
@@ -205,7 +205,7 @@ async def test_missing_handoff_fails():
  response = ModelResponse(
  output=[get_text_message("Hello, world!"), get_handoff_tool_call(agent_2)],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  with pytest.raises(ModelBehaviorError):
  RunImpl.process_model_response(
@@ -229,7 +229,7 @@ async def test_multiple_handoffs_doesnt_error():
  get_handoff_tool_call(agent_2),
  ],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = RunImpl.process_model_response(
  agent=agent_3,
@@ -254,7 +254,7 @@ async def test_final_output_parsed_correctly():
  get_final_output_message(Foo(bar="123").model_dump_json()),
  ],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
 
  RunImpl.process_model_response(
@@ -281,7 +281,7 @@ async def test_file_search_tool_call_parsed_correctly():
  response = ModelResponse(
  output=[get_text_message("hello"), file_search_call],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = RunImpl.process_model_response(
  agent=agent,
@@ -306,7 +306,7 @@ async def test_function_web_search_tool_call_parsed_correctly():
  response = ModelResponse(
  output=[get_text_message("hello"), web_search_call],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = RunImpl.process_model_response(
  agent=agent,
@@ -333,7 +333,7 @@ async def test_reasoning_item_parsed_correctly():
  response = ModelResponse(
  output=[reasoning],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = RunImpl.process_model_response(
  agent=Agent(name="test"),
@@ -401,7 +401,7 @@ async def test_computer_tool_call_without_computer_tool_raises_error():
  response = ModelResponse(
  output=[computer_call],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  with pytest.raises(ModelBehaviorError):
  RunImpl.process_model_response(
@@ -430,7 +430,7 @@ async def test_computer_tool_call_with_computer_tool_parsed_correctly():
  response = ModelResponse(
  output=[computer_call],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
  result = RunImpl.process_model_response(
  agent=agent,
@@ -460,7 +460,7 @@ async def test_tool_and_handoff_parsed_correctly():
  get_handoff_tool_call(agent_1),
  ],
  usage=Usage(),
- referenceable_id=None,
+ response_id=None,
  )
 
  result = RunImpl.process_model_response(