RePhraseQueryRetriever

简单的检索器，将用户输入和查询传递给检索器之间应用LLM。

它可以用于以任何方式预处理用户输入。

from_llm类方法中使用的默认提示：

DEFAULT_TEMPLATE = """You are an assistant tasked with taking a natural language \
query from a user and converting it into a query for a vectorstore. \
In this process, you strip out information that is not relevant for \
the retrieval task. Here is the user query: {question}"""

创建一个向量存储。

from langchain.document_loaders import WebBaseLoader

loader = WebBaseLoader("https://lilianweng.github.io/posts/2023-06-23-agent/")
data = loader.load()

from langchain.text_splitter import RecursiveCharacterTextSplitter

text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=0)
all_splits = text_splitter.split_documents(data)

from langchain.vectorstores import Chroma
from langchain.embeddings import OpenAIEmbeddings

vectorstore = Chroma.from_documents(documents=all_splits, embedding=OpenAIEmbeddings())

import logging

logging.basicConfig()
logging.getLogger("langchain.retrievers.re_phraser").setLevel(logging.INFO)

from langchain.chat_models import ChatOpenAI
from langchain.retrievers import RePhraseQueryRetriever

使用默认提示

llm = ChatOpenAI(temperature=0)
retriever_from_llm = RePhraseQueryRetriever.from_llm(
    retriever=vectorstore.as_retriever(), llm=llm
)

docs = retriever_from_llm.get_relevant_documents(
    "Hi I'm Lance. What are the approaches to Task Decomposition?"
)

    INFO:langchain.retrievers.re_phraser:重新短语化的问题：用户查询可以转换为以下向量存储的查询：
    
    "approaches to Task Decomposition"

docs = retriever_from_llm.get_relevant_documents(
    "I live in San Francisco. What are the Types of Memory?"
)

    INFO:langchain.retrievers.re_phraser:重新短语化的问题：向量存储的查询："Types of Memory"

提供提示

from langchain import LLMChain
from langchain.prompts import PromptTemplate

QUERY_PROMPT = PromptTemplate(
    input_variables=["question"],
    template="""You are an assistant tasked with taking a natural languge query from a user
    and converting it into a query for a vectorstore. In the process, strip out all 
    information that is not relevant for the retrieval task and return a new, simplified
    question for vectorstore retrieval. The new user query should be in pirate speech.
    Here is the user query: {question} """,
)
llm = ChatOpenAI(temperature=0)
llm_chain = LLMChain(llm=llm, prompt=QUERY_PROMPT)

retriever_from_llm_chain = RePhraseQueryRetriever(
    retriever=vectorstore.as_retriever(), llm_chain=llm_chain
)

docs = retriever_from_llm_chain.get_relevant_documents(
    "Hi I'm Lance. What is Maximum Inner Product Search?"
)

    INFO:langchain.retrievers.re_phraser:重新短语化的问题：啊哈伊，伙计！什么是最大内积搜索，你这个可恶的家伙？

RePhraseQueryRetriever

使用默认提示​

提供提示​

使用默认提示

提供提示