Enhance LLM integration and configuration in the backend

[sayedio]

• Updated .env.example to include new LLM provider configurations and parameters.
• Modified config.py to support multiple LLM providers with specific API keys and settings.
• Refactored app.py to utilize a new LLM chain for narrative generation, falling back to the original LLM narrator if necessary.
• Improved narrative handling in alerts.py to include provider information in cached notes.
• Introduced a Draft class in narrative.py to encapsulate LLM responses, including text and provider details.
• Updated documentation in DECISION_POLICY.md to reflect changes in LLM usage and provider prioritization.
• Enhanced frontend components to display the provider of the LLM-generated notes and improved error messaging for rejected drafts.