Error when retrying a task when the model used for the retried task is not available anymore

Which version of integration_openai are you using?

3.9.1

Which version of Nextcloud are you using?

32.0.1

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

Firefox latest

Describe the Bug

When re-running a task with a LocalAI backend with a model that existed for a previous run of the task but doesn't exist anymore, an immediate error is shown in the UI without indication of what is happening (something with [object Object].

If the new model is selected in the advanced options, then it works fine.

Expected Behavior

If the user tries to re-run a task with a model that is not available anymore, the system should warn him/her. Ideally with a prompt to select one of the currently available modules, or just a message, but I think that the current handling of the error can be improved.

To Reproduce

1. Run a task with a model M1. I'm trying with a summarize task.
2. Delete M in LocalAI, add M2, restart LocalAI
3. Re-run the task at point 1

https://github.com/user-attachments/assets/2a069f74-cb34-4bf0-a954-1e75573be090

The issue names no files or tests. Reproduce the retry flow with a previously used LocalAI model that has been removed, then inspect the UI error handling and the model-selection path; done means the user receives a useful warning or can select an available model instead of seeing [object Object].