OpenAI URL base url change not reflected in model, if model is not changed. (might add double v1)

[d4g]

Which version of integration_openai are you using?

3.2.0

Which version of Nextcloud are you using?

Nextcloud Hub 9 (30.0.2)

Which browser are you using? In case you are using the phone App, specify the Android or iOS version and device please.

Firefox 133

Describe the Bug

I configured litellm with the openai integration.
The model search works, models are found, but when accessing the models, another v1 gets added to the request.

Expected Behavior

Ai works and no additional v1 gets added.

To Reproduce

Configure custom url

[d4g]

Okay, i found out what the error is.
It seems that due to the update, another "v1" was added - to the url and to the model.
As I did not select another model but kept the model the same, the double-v1 was not changed.

Solution was to select another model and then switch back to the original one, then it was correctly set.

I still deem this as a bug, as if you change the base URL, the model should be saved again with the new base-url, even if you don't select a new model.

[FriedVariable]

For me changing the model does not help. Settings saying /v1/models and models are getting loaded, but Chat does not work. Works all well on NC 28, OpenWeb UI etc..

[kyteinsky]

hello,
thank you for the report @d4g, it will be fixed with #207 so accidental or manual migrations would not add /v1 twice.

@FriedVariable this seems like a separate issue. Would you mind opening an issue with the relevant logs from the server?