Create evaluation suite for benchmarking various models

[nikiburggraf]

Related to: #2

• Define our needs for an evaluation suite
• Collect the metrics we're using for benchmarking for various models using this evaluation suite
• Update the evaluation code for deepeval to leverage the modular retrieval / generation code.

[parvatijay2901]

PR #23 addressed part of this.
Remaining work is to update the code to leverage the new modular / class based implementation of the retrieval and generation modules.