Allow chatting with different providers

[alberto98fx]

With this PR it's now possible to use also different models and providers included ollama for local execution: You can configure everything you want from .env (check example I supplied)

Providers supported are:

• OpenAI
• Anthropic
• Groq
• OpenRouter
• Ollama

Disclaimer: I am not a professional coder so it might be that I did not apply the best practices here... Any help or suggestion is welcome :)

[alberto98fx]

Hey @ironman5366 I took the chance to also create a PoC Github action (https://github.com/alberto98fx/ai-murder-mystery-hackathon/actions/runs/10093046001/job/27907943930): for this iteration I commented out the health probe because for some reasons it does not want to give back the 200 status code and hence fails within the 30 seconds timeout. I will investigate why this happen when I have a little bit more time to dedicate. I added a small health check endpoint that maybe could be improved a little.

[ironman5366]

Let's leave out the github action, but the rest of this is good to merge! If you want you can open the action as a separate PR, but I think we're fine without it for the moment

[alberto98fx]

@ironman5366 Done!