Query embeddings from /chats/<chat>/converse api to get cosine similarity results from pinecone Index.

[skks1212]

Now that we have upsert and getting the embedding from prompts in place, let us send the embeddings to our Pinecone Index (specified in PINECONE_INDEX env variable) to get the cosine similarity result. Once we get that, send the data to GPT via langchain (I have little understanding of Langchain at the moment, so I will encourage you to use the docs to find out best practices), and then return the result as response.