vector_store needs of query embedding.

[HERIUN]

Checked

• I searched existing ideas and did not find a similar one
• I added a very descriptive title
• I've clearly described the feature request and motivation for it

Feature request

google embedding models has task_type. like https://ai.google.dev/api/embeddings?hl=ko#v1beta.TaskType

In RAG system, document embedding and query embedding can be different.

but langchain vector_store. doesn't care about this..

only possible by two line

query_embed = self.query_embedding_model.embed_query(query)
docs = self.vector_store.similarity_search_by_vector(query_embed, k=5)

but as a one line it could be cool

self.vector_store.as_retriever(query_embedding=query_embedding)

Motivation

google embedding models has task_type. like https://ai.google.dev/api/embeddings?hl=ko#v1beta.TaskType

Proposal (If applicable)

No response