Uncaught exceptions when determining solvability of AI request

I tested this manually and found that an AiRequestErrorException is thrown when the limit is reached. As a result, it is caught by the generic catch (\Exception $e) block, which is not very helpful in this scenario.

@mglaman Code LGTM. Added a suggestion.

which is not very helpful in this scenario.

How isn't it very helpful?

Re #6 🐛 Uncaught exceptions when determining solvability of AI request Active , I mean to say that we are not getting expected exception when quota is exhausted. We are not getting AiQuotaException instead we get AiRequestErrorException due to which we will get 'Something went wrong' in response which will be not that useful.

@narendrar ah yeah, I see. I was getting that exception. But the problem is that there are so many AI module exceptions without a way of knowing it is an AI module one without tracking each. I opened ✨ Provide an exception interface all exceptions implement Active so it'd be easier to catch this.

I'm guessing we should just always return the exception message regardless. We're not stable yet. ✨ Dispatch AiExceptionEvent when a provider throws an exception Active is a feature request I made that allows third parties to customize exception messages handled by integrations like XB AI.

I'm postponing this issue on the two from #8:

1. ✨ Provide an exception interface all exceptions implement Active
2. ✨ Dispatch AiExceptionEvent when a provider throws an exception Active