feat: add agent_with_prompt_builder super-component

[marc-mrt]

Related Issues

n/a

Proposed Changes:

As we've seen ChatPromptBuilder be paired with Agent for every use-case on the Platform, a good opportunity for simplification would be a (super-)component that encapsulates the two.
This is an attempts at implementation with a super-component indeed, happy for general feedback on that take.

How did you test it?

Added unit tests, mainly to ensure good behavior of the super-component's input inference based on the template provided.

Notes for the reviewer

I've tried to be exhaustive but there is one thing that I still don't quite get, and would appreciate guidance on: to_dict and from_dict. I saw for example on Haystack's document_preprocessor how they are defined, but not on the multi_file_converter and I'm not sure I understand why on or the other.

Also - as it's my first contribution to Haystack (albeit experimental), any pointers as to unfollowed standards, code quality or misunderstandings of the framework are more than welcome!

Checklist

• I have read the contributors guidelines and the code of conduct
• I have updated the related issue with new insights and changes
• I added unit tests and updated the docstrings
• I've used one of the conventional commit types for my PR title: fix:, feat:, build:, chore:, ci:, docs:, style:, refactor:, perf:, test:.
• I documented my code
• I ran pre-commit hooks and fixed any issue

[coveralls]

Pull Request Test Coverage Report for Build 22105228703

Details

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.9%) to 58.571%

---

Totals
Change from base Build 22096224768:	0.9%
Covered Lines:	697
Relevant Lines:	1190

---

💛 - Coveralls

[sjrl]

For simplicity we could consider dropping variables and just not use it. I don't think it's often used since we often rely on the auto detection of the template

[sjrl]

I think a preferred interface for this new component is to just expose a param called user_prompt: str which we feed into the ChatPromptBuilder internally.

[sjrl]

A few other comments I have:

• Do you know if we can easily reuse this super component to make the LLM one? I don't think we've ever done inheritance with a pre-made super component before
• If we go with the super component approach I think we really need to test the input and output sockets of this component. The Agent component sets many inputs and outputs dynamically that we will still want to be able to use. To test this we should write tests checking the values of the attributes Agent().__haystack_input__ and Agent().__haystack_output__ where we set things like the state_schema={"documents": list[Document]} and we should see documents become a valid input and valid output of the component.