feat: consensus strategies and examples

[mohulshukla]

Implemented advanced aggregation functions and strategies, and included examples to test the functions.

Example files:

1. majority_consensus.py
2. tournament_consensus.py
3. clustering_consensus.py

For less redundancy, each example file uses sample LLM responses based on an example prompt. This is an extensible approach, since this framework is meant to be used for consensus approaches where a single agent is an individual trained format and consensus strategies are applied on the outputs.

Several advanced aggregation functions have been implemented, providing logging and mathematical approaches for recording Shapley values, clustering similar responses (for text based situations), identifying hallucinations, and more. Certain consensus strategies are more applicable in certain contexts, and the example files identify some appropriate cases where this can be done.

All logic code and strategies are structured under aggregator/, since this is where all of the aggregation logic should exist.