feat: Add ServiceContainer Haystack 2.x component

[vblagoje]

Why:

• To introduce a ServiceContainer component which facilitates managing a service instance and generating JSON descriptions for its methods. ServiceContainer will be used for LLM function calling by proxying LLM functional calls to ServiceContainer while ServiceContainer will in turn facilitate generating JSON descriptions for service's methods to register functions with LLMs.

What:

• Added ServiceContainer component, which encapsulates a service instance and provides functionality for invoking methods on the service instance reflectively based on ChatMessage input and generating JSON descriptions for all methods of the service instance according to the OpenAI functional API specification.

How can it be used:

• By instantiating the ServiceContainer with a service instance and utilizing the run method to process ChatMessage inputs.

How did you test it:

• No tests, yet see section below

Notes for reviewer:

• This component is in draft stage awaiting ChatMessage merging and changes described in Rename GPTGenerator Following LLM Component Naming Scheme #6117

[vblagoje]

The idea is to replace step 5 from LLM function calling https://colab.research.google.com/drive/1hPyK7DKXuMRRm03uE7-h-oG0X4h3Lvpx?usp=sharing It'll be an amazing feature. However we need #6109 and ChatMessage support first

[julian-risch]

@vblagoje Once this PR is ready for review I won't be able to review it so somebody else needs to be assigned. I nominate @anakin87 🎉

[vblagoje]

There is a better way to do this via Pydantic. I'll close this draft and open a new one once it is ready cc @anakin87