search = vectordb.similarity_search(" ") this returns only 4 docs. is there a way to return all documents from vectordb? isn't this what you want since the purpose is to get a summary of a corpus which comprises of all texts?