Merge document_stores

[Rocha-a21906962]

Hi there,
I am using Haystack for FAQ retrieval using 1 sparse retriever with TF-IDF (InMemoryDocumentStore) plus 1 dense retriever with Sentence-Transformers with FAISS (FAISSDocumentStore).
I was wondering if it is possible to create a single DocumentStore able to handle both my retrievers.

[ZanSara]

Hey @Rocha-a21906962, FAISSDocumentStore should be able to support both dense and sparse retrievers.

from haystack import Document
from haystack.document_stores import FAISSDocumentStore
from haystack.nodes import EmbeddingRetriever, TfidfRetriever

document_store = FAISSDocumentStore(sql_url="sqlite:///")
document_store.write_documents([
    Document(content="doc 1"),
    Document(content="doc 2"),
    Document(content="doc 3")
])

retriever_sparse = TfidfRetriever(document_store=document_store)
retriever_dense = EmbeddingRetriever(document_store=document_store, embedding_model="deepset/sentence_bert")
document_store.update_embeddings(retriever=retriever_dense)

print(retriever_sparse.retrieve(query="doc 2"))
print(retriever_dense.retrieve(query="doc 2"))

[Rocha-a21906962]

Thank you! That was what I was looking for.