add provider checking based on model name and provider

[PCSwingle]

No description provided.

[mentatbot]

MENTAT CODE REVIEW IN ACTIVE DEVELOPMENT. Only in use on mentat and internal repos.
Please Reply with feedback.

This pull request introduces important updates to handle provider checking based on model names and providers dynamically. The changes are well thought out, particularly in handling different providers and ensuring that the API calls are guarded against common errors. However, there are a few areas where the implementation could be improved for clarity, error handling, and user guidance. Addressing these points will make the code more robust and user-friendly.

[jakethekoenig]

When I set my model to test I get an exception without an opportunity to set my provider:

Maybe we should wrap count_tokens to return 0 if it would throw an exception?

[jakethekoenig]

When I set my model to "test" and my provider to "openai" or "open_ai" I get the same exception

[jakethekoenig]

This might not be totally connected to this PR but it would be nice if mentat didn't disable the user input box if they get a 404 from the api. For instance if I set my model to test I get the following (reasonable) error but then I can't enter anything into the box again

[PCSwingle]

This might not be totally connected to this PR but it would be nice if mentat didn't disable the user input box if they get a 404 from the api.

Agreed but it might be better to do in a separate PR

[jakethekoenig]

Looks good to me. I guess another thing to do in another PR is upgrade to 0.2 spice