Unify embedding model

[almutareb]

In the recent 2 branches adding vector store backed features we are using different embedding models:

1. sentence-transformers/multi-qa-mpnet-base-dot-v1
2. sentence-transformers/all-MiniLM-L6-v2 (which is used also as all-MiniLM-L6-v2 in the same file)

We need to unify the usage of embeddings across all functions.

Suggest to use a config file to handle these sort of things

[almutareb]

config file added in pull #17 @isayahc please make sure, it's used in all instances