feat: add abstraction layer to leapfrogai_evals to allow custom RAG pipelines

[jalling97]

User Story

As a developer leveraging LeapfrogAI
I want to be able to evaluation my custom RAG pipelines
So that I can compare my custom RAG pipeline to the LeapfrogAI baseline

Acceptance Criteria

An abstract class RAGPipeline is added to replace the current method of running each evaluation. a baseline LFAI RAG pipeline class implementation will be added so that the current baseline still functions as expected. If custom pipelines are required, one can follow the implementation details using the abstract class to run LFAI evals.