Query matching: embed query and match to top-K chunks

[prmsregmi]

Develop functionality to embed a user query into a vector and then compare it against the embedded chunks of text to retrieve the top-K most semantically similar chunks. This will enable targeted information retrieval from large documents.