OpenAI external tools (home-assistant#150599)

Author: Shulyaka

homeassistant/components/openai_conversation/entity.py

@@ -14,13 +14,16 @@
 from openai.types.responses import (
     EasyInputMessageParam,
     FunctionToolParam,
+    ResponseCodeInterpreterToolCall,
     ResponseCompletedEvent,
     ResponseErrorEvent,
     ResponseFailedEvent,
     ResponseFunctionCallArgumentsDeltaEvent,
     ResponseFunctionCallArgumentsDoneEvent,
     ResponseFunctionToolCall,
     ResponseFunctionToolCallParam,
+    ResponseFunctionWebSearch,
+    ResponseFunctionWebSearchParam,
     ResponseIncompleteEvent,
     ResponseInputFileParam,
     ResponseInputImageParam,
@@ -149,16 +152,27 @@ def _convert_content_to_param(
     """Convert any native chat message for this agent to the native format."""
     messages: ResponseInputParam = []
     reasoning_summary: list[str] = []
+    web_search_calls: dict[str, ResponseFunctionWebSearchParam] = {}
 
     for content in chat_content:
         if isinstance(content, conversation.ToolResultContent):
-            messages.append(
-                FunctionCallOutput(
-                    type="function_call_output",
-                    call_id=content.tool_call_id,
-                    output=json.dumps(content.tool_result),
+            if (
+                content.tool_name == "web_search_call"
+                and content.tool_call_id in web_search_calls
+            ):
+                web_search_call = web_search_calls.pop(content.tool_call_id)
+                web_search_call["status"] = content.tool_result.get(  # type: ignore[typeddict-item]
+                    "status", "completed"
+                )
+                messages.append(web_search_call)
+            else:
+                messages.append(
+                    FunctionCallOutput(
+                        type="function_call_output",
+                        call_id=content.tool_call_id,
+                        output=json.dumps(content.tool_result),
+                    )
                 )
-            )
             continue
 
         if content.content:
@@ -173,15 +187,27 @@ def _convert_content_to_param(
 
         if isinstance(content, conversation.AssistantContent):
             if content.tool_calls:
-                messages.extend(
-                    ResponseFunctionToolCallParam(
-                        type="function_call",
-                        name=tool_call.tool_name,
-                        arguments=json.dumps(tool_call.tool_args),
-                        call_id=tool_call.id,
-                    )
-                    for tool_call in content.tool_calls
-                )
+                for tool_call in content.tool_calls:
+                    if (
+                        tool_call.external
+                        and tool_call.tool_name == "web_search_call"
+                        and "action" in tool_call.tool_args
+                    ):
+                        web_search_calls[tool_call.id] = ResponseFunctionWebSearchParam(
+                            type="web_search_call",
+                            id=tool_call.id,
+                            action=tool_call.tool_args["action"],
+                            status="completed",
+                        )
+                    else:
+                        messages.append(
+                            ResponseFunctionToolCallParam(
+                                type="function_call",
+                                name=tool_call.tool_name,
+                                arguments=json.dumps(tool_call.tool_args),
+                                call_id=tool_call.id,
+                            )
+                        )
 
             if content.thinking_content:
                 reasoning_summary.append(content.thinking_content)
@@ -211,25 +237,37 @@ def _convert_content_to_param(
 async def _transform_stream(
     chat_log: conversation.ChatLog,
     stream: AsyncStream[ResponseStreamEvent],
-) -> AsyncGenerator[conversation.AssistantContentDeltaDict]:
+) -> AsyncGenerator[
+    conversation.AssistantContentDeltaDict | conversation.ToolResultContentDeltaDict
+]:
     """Transform an OpenAI delta stream into HA format."""
     last_summary_index = None
+    last_role: Literal["assistant", "tool_result"] | None = None
 
     async for event in stream:
         LOGGER.debug("Received event: %s", event)
 
         if isinstance(event, ResponseOutputItemAddedEvent):
-            if isinstance(event.item, ResponseOutputMessage):
-                yield {"role": event.item.role}
-                last_summary_index = None
-            elif isinstance(event.item, ResponseFunctionToolCall):
+            if isinstance(event.item, ResponseFunctionToolCall):
                 # OpenAI has tool calls as individual events
                 # while HA puts tool calls inside the assistant message.
                 # We turn them into individual assistant content for HA
                 # to ensure that tools are called as soon as possible.
                 yield {"role": "assistant"}
+                last_role = "assistant"
                 last_summary_index = None
                 current_tool_call = event.item
+            elif (
+                isinstance(event.item, ResponseOutputMessage)
+                or (
+                    isinstance(event.item, ResponseReasoningItem)
+                    and last_summary_index is not None
+                )  # Subsequent ResponseReasoningItem
+                or last_role != "assistant"
+            ):
+                yield {"role": "assistant"}
+                last_role = "assistant"
+                last_summary_index = None
         elif isinstance(event, ResponseOutputItemDoneEvent):
             if isinstance(event.item, ResponseReasoningItem):
                 yield {
@@ -240,6 +278,52 @@ async def _transform_stream(
                         encrypted_content=event.item.encrypted_content,
                     )
                 }
+                last_summary_index = len(event.item.summary) - 1
+            elif isinstance(event.item, ResponseCodeInterpreterToolCall):
+                yield {
+                    "tool_calls": [
+                        llm.ToolInput(
+                            id=event.item.id,
+                            tool_name="code_interpreter",
+                            tool_args={
+                                "code": event.item.code,
+                                "container": event.item.container_id,
+                            },
+                            external=True,
+                        )
+                    ]
+                }
+                yield {
+                    "role": "tool_result",
+                    "tool_call_id": event.item.id,
+                    "tool_name": "code_interpreter",
+                    "tool_result": {
+                        "output": [output.to_dict() for output in event.item.outputs]  # type: ignore[misc]
+                        if event.item.outputs is not None
+                        else None
+                    },
+                }
+                last_role = "tool_result"
+            elif isinstance(event.item, ResponseFunctionWebSearch):
+                yield {
+                    "tool_calls": [
+                        llm.ToolInput(
+                            id=event.item.id,
+                            tool_name="web_search_call",
+                            tool_args={
+                                "action": event.item.action.to_dict(),
+                            },
+                            external=True,
+                        )
+                    ]
+                }
+                yield {
+                    "role": "tool_result",
+                    "tool_call_id": event.item.id,
+                    "tool_name": "web_search_call",
+                    "tool_result": {"status": event.item.status},
+                }
+                last_role = "tool_result"
         elif isinstance(event, ResponseTextDeltaEvent):
             yield {"content": event.delta}
         elif isinstance(event, ResponseReasoningSummaryTextDeltaEvent):
@@ -252,6 +336,7 @@ async def _transform_stream(
                 and event.summary_index != last_summary_index
             ):
                 yield {"role": "assistant"}
+                last_role = "assistant"
             last_summary_index = event.summary_index
             yield {"thinking_content": event.delta}
         elif isinstance(event, ResponseFunctionCallArgumentsDeltaEvent):
@@ -348,6 +433,33 @@ async def _async_handle_chat_log(
         """Generate an answer for the chat log."""
         options = self.subentry.data
 
+        messages = _convert_content_to_param(chat_log.content)
+
+        model_args = ResponseCreateParamsStreaming(
+            model=options.get(CONF_CHAT_MODEL, RECOMMENDED_CHAT_MODEL),
+            input=messages,
+            max_output_tokens=options.get(CONF_MAX_TOKENS, RECOMMENDED_MAX_TOKENS),
+            top_p=options.get(CONF_TOP_P, RECOMMENDED_TOP_P),
+            temperature=options.get(CONF_TEMPERATURE, RECOMMENDED_TEMPERATURE),
+            user=chat_log.conversation_id,
+            store=False,
+            stream=True,
+        )
+
+        if model_args["model"].startswith(("o", "gpt-5")):
+            model_args["reasoning"] = {
+                "effort": options.get(
+                    CONF_REASONING_EFFORT, RECOMMENDED_REASONING_EFFORT
+                ),
+                "summary": "auto",
+            }
+            model_args["include"] = ["reasoning.encrypted_content"]
+
+        if model_args["model"].startswith("gpt-5"):
+            model_args["text"] = {
+                "verbosity": options.get(CONF_VERBOSITY, RECOMMENDED_VERBOSITY)
+            }
+
         tools: list[ToolParam] = []
         if chat_log.llm_api:
             tools = [
@@ -381,36 +493,11 @@ async def _async_handle_chat_log(
                     ),
                 )
             )
+            model_args.setdefault("include", []).append("code_interpreter_call.outputs")  # type: ignore[union-attr]
 
-        messages = _convert_content_to_param(chat_log.content)
-
-        model_args = ResponseCreateParamsStreaming(
-            model=options.get(CONF_CHAT_MODEL, RECOMMENDED_CHAT_MODEL),
-            input=messages,
-            max_output_tokens=options.get(CONF_MAX_TOKENS, RECOMMENDED_MAX_TOKENS),
-            top_p=options.get(CONF_TOP_P, RECOMMENDED_TOP_P),
-            temperature=options.get(CONF_TEMPERATURE, RECOMMENDED_TEMPERATURE),
-            user=chat_log.conversation_id,
-            store=False,
-            stream=True,
-        )
         if tools:
             model_args["tools"] = tools
 
-        if model_args["model"].startswith(("o", "gpt-5")):
-            model_args["reasoning"] = {
-                "effort": options.get(
-                    CONF_REASONING_EFFORT, RECOMMENDED_REASONING_EFFORT
-                ),
-                "summary": "auto",
-            }
-            model_args["include"] = ["reasoning.encrypted_content"]
-
-        if model_args["model"].startswith("gpt-5"):
-            model_args["text"] = {
-                "verbosity": options.get(CONF_VERBOSITY, RECOMMENDED_VERBOSITY)
-            }
-
         last_content = chat_log.content[-1]
 
         # Handle attachments by adding them to the last user message

tests/components/openai_conversation/__init__.py

@@ -29,6 +29,7 @@
     ResponseWebSearchCallInProgressEvent,
     ResponseWebSearchCallSearchingEvent,
 )
+from openai.types.responses.response_code_interpreter_tool_call import OutputLogs
 from openai.types.responses.response_function_web_search import ActionSearch
 from openai.types.responses.response_reasoning_item import Summary
 
@@ -320,7 +321,7 @@ def create_web_search_item(id: str, output_index: int) -> list[ResponseStreamEve
 
 
 def create_code_interpreter_item(
-    id: str, code: str | list[str], output_index: int
+    id: str, code: str | list[str], output_index: int, logs: str | None = None
 ) -> list[ResponseStreamEvent]:
     """Create a message item."""
     if isinstance(code, str):
@@ -388,7 +389,7 @@ def create_code_interpreter_item(
                     id=id,
                     code=code,
                     container_id=container_id,
-                    outputs=None,
+                    outputs=[OutputLogs(type="logs", logs=logs)] if logs else None,
                     status="completed",
                     type="code_interpreter_call",
                 ),

tests/components/openai_conversation/snapshots/test_conversation.ambr

@@ -1,4 +1,39 @@
 # serializer version: 1
+# name: test_code_interpreter
+  list([
+    dict({
+      'content': 'Please use the python tool to calculate square root of 55555',
+      'role': 'user',
+      'type': 'message',
+    }),
+    dict({
+      'arguments': '{"code": "import math\\nmath.sqrt(55555)", "container": "cntr_A"}',
+      'call_id': 'ci_A',
+      'name': 'code_interpreter',
+      'type': 'function_call',
+    }),
+    dict({
+      'call_id': 'ci_A',
+      'output': '{"output": [{"logs": "235.70108188126758\\n", "type": "logs"}]}',
+      'type': 'function_call_output',
+    }),
+    dict({
+      'content': 'I’ve calculated it with Python: the square root of 55555 is approximately 235.70108188126758.',
+      'role': 'assistant',
+      'type': 'message',
+    }),
+    dict({
+      'content': 'Thank you!',
+      'role': 'user',
+      'type': 'message',
+    }),
+    dict({
+      'content': 'You are welcome!',
+      'role': 'assistant',
+      'type': 'message',
+    }),
+  ])
+# ---
 # name: test_function_call
   list([
     dict({
@@ -172,3 +207,36 @@
     }),
   ])
 # ---
+# name: test_web_search
+  list([
+    dict({
+      'content': "What's on the latest news?",
+      'role': 'user',
+      'type': 'message',
+    }),
+    dict({
+      'action': dict({
+        'query': 'query',
+        'type': 'search',
+      }),
+      'id': 'ws_A',
+      'status': 'completed',
+      'type': 'web_search_call',
+    }),
+    dict({
+      'content': 'Home Assistant now supports ChatGPT Search in Assist',
+      'role': 'assistant',
+      'type': 'message',
+    }),
+    dict({
+      'content': 'Thank you!',
+      'role': 'user',
+      'type': 'message',
+    }),
+    dict({
+      'content': 'You are welcome!',
+      'role': 'assistant',
+      'type': 'message',
+    }),
+  ])
+# ---