proposal: rag evaluation results presentation

[davidsbatista]

Proposal for a presentation of RAG pipeline evaluation results

[julian-risch]

We're almost there. 👍 I am not sure about the idea of passing other incomparative_summary(self, other: "RAGPipelineEvaluation"): but we can iterate on it later. I wouldn't block the proposal because of that. Better to start with the implementation.
One decision that should be made when implementation of this proposal starts is whether we want to rely on pandas or not. We should check whether we want the Document class to rely on pandas or not for that. We could definitely implement the new eval features without pandas and have one data export option that makes it easy for advanced users to use pandas if they want to.
One change request: could you map the user stories directly to the methods please? For example, that mapping should explain when the user uses find_thresholds for one of the stories from the issue.
I left some minor comments in the proposal too.

[davidsbatista]

We could definitely implement the new eval features without pandas and have one data export option that makes it easy for advanced users to use pandas if they want to
@julian-risch I agree with you, I would make clear that we don't want pandas, and have JSON that can easily be transformed into a pandas

[shadeMe]

Just a couple of minor changes before it's good to merge (from my side) 🎉

[davidsbatista]

@julian-risch @mrm1001 do you want to add, suggest anything else? if not I will merge it