Added Azure AI Chat Completion Client #4723
Conversation
|
@yanivvak can you review this? |
|
@ekzhu @rohanthacker
|
| @@ -50,6 +50,9 @@ video-surfer = [ | |||
| grpc = [ | |||
| "grpcio~=1.62.0", # TODO: update this once we have a stable version. | |||
| ] | |||
| azure-ai-inference = [ | |||
| "azure-ai-inference>=1.0.0b6", | |||
There was a problem hiding this comment.
Is there a specific reason for this version lower bound?
There was a problem hiding this comment.
No this is not intentional, this is what gets added when i run uv add "azure-ai-inference" --optional azure-ai-inference.
Is there another command I should use? I'm not experienced with uv
| usage=usage, | ||
| cached=False, | ||
| ) | ||
| yield result |
There was a problem hiding this comment.
Update object-level usage data.
| usage=usage, | ||
| cached=False, | ||
| ) | ||
| return response |
There was a problem hiding this comment.
Update object-level usage data.
| raise ValueError("Model does not support JSON output") | ||
|
|
||
| if json_output is True: | ||
| create_args["response_format"] = ChatCompletionsResponseFormatJSON() |
There was a problem hiding this comment.
I think we need to check existing response_format value to make sure we aren't overwriting it.
| return name | ||
|
|
||
|
|
||
| class AzureAIChatCompletionClient(ChatCompletionClient): |
There was a problem hiding this comment.
Doc string is required. See examples: https://microsoft.github.io/autogen/dev/reference/python/autogen_ext.agents.openai.html
|
@lspinheiro could you help reviewing this PR? |
|
|
||
| class AzureAIChatCompletionClient(ChatCompletionClient): | ||
| def __init__(self, **kwargs: Unpack[AzureAIChatCompletionClientConfig]): | ||
| if "endpoint" not in kwargs: |
There was a problem hiding this comment.
I think this part could benefit from some better separation of concerns between config validation and instantiation. e.g.
class AzureAIChatCompletionClient(ChatCompletionClient):
def __init__(self, **kwargs: Unpack[AzureAIClientConfiguration]):
config = self._validate_config(kwargs)
self._client = self._create_client(config)
self._create_args = self._prepare_create_args(config)
# ...
@staticmethod
def _validate_config(config: Mapping[str, Any]) -> AzureAIClientConfiguration:
# Validation logic here
return config
Looks quite good and is consistent with the openai client. I have a minor comment about the config validation. @jackgerrits may have more options since a lot of the design decisions here are driven by his original implementation of the openai client. If anything doesn't make since in this context he would be the best person to evaluate. |
* Added: object-level usage data * Added: doc string * Added: check existing response_format value * Added: _validate_config and _create_client
…hacker/autogen into feature/azure-ai-inference-client
Related issue number
#4683 Adds initial support for Azure AI Chat Completion Client
Checks