We should include a normalized score value in every vectordb's similarityResponse method. We will not display this at this time but it should be a normalized score of 0-100% in decimal value.

In the future we can show the score next to each snippet and add a default 'relevancy filter' per workspace to determine if a snippet/citation should be included in a response.

Each vectorDB returns a vary idea of similarity scoring. Ref VectorAdmin to see how that is done.