Integrate local LLMs and define a query optimiser to pick the best LLM based on user query

[stationeros]

We have already integrated ollama as a pilot use case #215 , the idea of this issue is to give user the same option of a localLLM but instead of Llama evaluate the best LLM model from a range of models that can best answer the query and return the output from that model