feat: standard outputs #31967
feat: standard outputs #31967
Conversation
libs/partners/openai/tests/integration_tests/chat_models/test_responses_api.py
Show resolved
Hide resolved
| return dict(annotation) | ||
|
|
||
|
|
||
| def _implode_reasoning_blocks(blocks: list[dict[str, Any]]) -> Iterable[dict[str, Any]]: |
There was a problem hiding this comment.
Perhaps avoid i and n? It's not looking like conventional python code. i and n are more typical of c code
i -> idx, n -> num_blocks or num_total or total
Co-authored-by: Eugene Yurtsev <eyurtsev@gmail.com>
| for key in ( | ||
| "url", | ||
| "base64", | ||
| "file_id", |
There was a problem hiding this comment.
each block has file_id, should this be removed so that this is a set of only non-required keys? if so we should consider adding title and context
|
|
||
|
|
||
| def convert_to_openai_image_block(content_block: dict[str, Any]) -> dict: | ||
| """Convert image content block to format expected by OpenAI Chat Completions API.""" | ||
| if content_block["source_type"] == "url": | ||
| if "url" in content_block: |
There was a problem hiding this comment.
should this also check the input content_block for type = "image"?
| @@ -117,36 +792,38 @@ def convert_to_openai_data_block(block: dict) -> dict: | |||
| formatted_block = convert_to_openai_image_block(block) | |||
|
|
|||
| elif block["type"] == "file": | |||
| if block["source_type"] == "base64": | |||
| file = {"file_data": f"data:{block['mime_type']};base64,{block['data']}"} | |||
| if "base64" in block or block.get("source_type") == "base64": | |||
There was a problem hiding this comment.
same here, check for type in set(DataContentBlock.type)?
| """The name of the tool to be called.""" | ||
| args: dict[str, Any] | ||
| """The arguments to the tool call.""" | ||
| id: Optional[str] |
There was a problem hiding this comment.
should we be using Optional or NotRequired?
|
|
||
| id: str | ||
| tool_call_id: str | ||
| content: list[dict[str, Any]] |
There was a problem hiding this comment.
why isn't this list[types.ContentBlock]?
| if isinstance(content, str): | ||
| self.content = [{"type": "text", "text": content}] |
There was a problem hiding this comment.
for each case where we have self.content = [{"type": "text", "text": content}], could we use the approach where we construct the TypedDict and then convert to dict? e.g.:
| if isinstance(content, str): | |
| self.content = [{"type": "text", "text": content}] | |
| if isinstance(content, str): | |
| content_typed = TextContentBlock(type="text", text=content) | |
| self.content = [content_typed.asdict()] |
| if text_blocks: | ||
| return "".join(block["text"] for block in text_blocks) | ||
| return None | ||
|
|
There was a problem hiding this comment.
I think a property for reasoning is missing here, no?
| else: | ||
| self.content = content | ||
|
|
||
| def text(self) -> str: |
There was a problem hiding this comment.
doesn't this need @property?
| else: | ||
| self.content = content | ||
|
|
||
| def text(self) -> str: |
…rt_to_openai_image_block()`
| @@ -293,6 +299,45 @@ def _create_message_from_message_type( | |||
| return message | |||
|
|
|||
|
|
|||
| def _convert_from_v1_message(message: MessageV1) -> BaseMessage: | |||
There was a problem hiding this comment.
is this still a WIP? the v0 AIMessage will have content with dicts in the shape of the new content blocks - is there a plan to convert back to the non-std-output (or is that even possible?) don't think so, but wanted to ask for clarity
| if isinstance(message.content, str): | ||
| content: list[types.ContentBlock] = [] | ||
| if message.content: | ||
| content = [{"type": "text", "text": message.content}] |
There was a problem hiding this comment.
perhaps the same asdict conversion here?
Co-authored-by: Nuno Campos <nuno@langchain.dev>
No description provided.