feat: use internal API for more organized implementations #412
Conversation
|
Here's what I've tested locally ( I put a helper script into our ✅ Available Providers (25): 🔑 Missing API Keys (7): |
| @staticmethod | ||
| def _convert_completion_response(response: "Message") -> ChatCompletion: | ||
| """Convert Anthropic Message to OpenAI ChatCompletion format.""" | ||
| return _convert_response(response) |
There was a problem hiding this comment.
I thought about moving all the logic from the utils.py file into this function, but I ended up keeping it like this because it seemed smoother and kept the anthropic file from getting too busy 🤷
There was a problem hiding this comment.
will not complain about smaller files
There was a problem hiding this comment.
I know that there are some provider-specific quirks, but it would be super great if we could make those happen in an inherited call with stuff before and after a super().acompletion(*args, **kwargs)
|
In a sense, I get that this is a bit of a YOLO: I updated all the providers and ran the integration tests, but its definitely still possible that I messed something up. However, this at least migrates us to a working setup, and if the integration tests are all passing I feel like that provides at least a moderate level of confidence that everything's ok. |
| @staticmethod | ||
| def _convert_completion_response(response: "Message") -> ChatCompletion: | ||
| """Convert Anthropic Message to OpenAI ChatCompletion format.""" | ||
| return _convert_response(response) |
There was a problem hiding this comment.
will not complain about smaller files
| yield _create_openai_chunk_from_anthropic_chunk(event, kwargs.get("model", "unknown")) | ||
| yield self._convert_completion_chunk_response(event, model_id=kwargs.get("model", "unknown")) | ||
|
|
||
| async def acompletion( |
There was a problem hiding this comment.
if we're standardizing API by adding more required methods, should we abstract acompletion et al. to provided methods?
There was a problem hiding this comment.
I like this idea. This PR is already a big boy, so I'll split this off into a separate PR, so that this change doesn't keep exploding into a bigger change.
| @staticmethod | ||
| def _convert_completion_response(response: "Message") -> ChatCompletion: | ||
| """Convert Anthropic Message to OpenAI ChatCompletion format.""" | ||
| return _convert_response(response) |
There was a problem hiding this comment.
I know that there are some provider-specific quirks, but it would be super great if we could make those happen in an inherited call with stuff before and after a super().acompletion(*args, **kwargs)
Description
Better structuring for providers where things are organized into shared function definitions
PR Type
💅 Refactor ## Relevant issuesChecklist