-
Notifications
You must be signed in to change notification settings - Fork 15.8k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
tests[patch]: populate API reference for chat models #28487
Conversation
The latest updates on your projects. Learn more about Vercel for Git ↗︎ 1 Skipped Deployment
|
@@ -74,7 +74,7 @@ def _validate_tool_call_message_no_args(message: BaseMessage) -> None: | |||
|
|||
class ChatModelIntegrationTests(ChatModelTests): | |||
@property | |||
def standard_chat_model_params(self) -> dict: | |||
def _standard_chat_model_params(self) -> dict: |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
is there a way to configure this via docstring, or does it have to be by name? just to avoid having to update all the implementations
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
From what I can tell the only integration that implements it is AWS. I added :meta private:
to docstrings but it still seems to show up in the attributes table. Going with that for now.
I'm actually unclear on what is intended for this property. It is always combined with chat_model_params
, so my guess was that we were testing a set of standard params (temperature, max_tokens, etc) that we want all models to support, and users should only overwrite chat_model_params
. If that's correct, then we should update AWS to not override standard_chat_model_params
.
Otherwise, we should document when/why this should be overwritten (as opposed to specifying things in chat_model_params
).
Populate API reference for test class properties and test methods for chat models.
Also
pytest.skip
some tests that were previously passed if features are not supported.