langchain/cookbook/myscale_vector_sql.ipynb

Vector SQL Retriever with MyScale

MyScale is an integrated vector database. You can access your database in SQL and also from here, LangChain. MyScale can make a use of various data types and functions for filters. It will boost up your LLM app no matter if you are scaling up your data or expand your system to broader application.

In [ ]:

!pip3 install clickhouse-sqlalchemy InstructorEmbedding sentence_transformers openai langchain-experimental

In [ ]:

import getpass
from os import environ

from langchain.chains import LLMChain
from langchain.prompts import PromptTemplate
from langchain_community.utilities import SQLDatabase
from langchain_experimental.sql.vector_sql import VectorSQLDatabaseChain
from langchain_openai import OpenAI
from sqlalchemy import MetaData, create_engine

MYSCALE_HOST = "msc-4a9e710a.us-east-1.aws.staging.myscale.cloud"
MYSCALE_PORT = 443
MYSCALE_USER = "chatdata"
MYSCALE_PASSWORD = "myscale_rocks"
OPENAI_API_KEY = getpass.getpass("OpenAI API Key:")

engine = create_engine(
    f"clickhouse://{MYSCALE_USER}:{MYSCALE_PASSWORD}@{MYSCALE_HOST}:{MYSCALE_PORT}/default?protocol=https"
)
metadata = MetaData(bind=engine)
environ["OPENAI_API_KEY"] = OPENAI_API_KEY

In [ ]:

from langchain_community.embeddings import HuggingFaceInstructEmbeddings
from langchain_experimental.sql.vector_sql import VectorSQLOutputParser

output_parser = VectorSQLOutputParser.from_embeddings(
    model=HuggingFaceInstructEmbeddings(
        model_name="hkunlp/instructor-xl", model_kwargs={"device": "cpu"}
    )
)

In [ ]:

from langchain.callbacks import StdOutCallbackHandler
from langchain_community.utilities.sql_database import SQLDatabase
from langchain_experimental.sql.prompt import MYSCALE_PROMPT
from langchain_experimental.sql.vector_sql import VectorSQLDatabaseChain
from langchain_openai import OpenAI

chain = VectorSQLDatabaseChain(
    llm_chain=LLMChain(
        llm=OpenAI(openai_api_key=OPENAI_API_KEY, temperature=0),
        prompt=MYSCALE_PROMPT,
    ),
    top_k=10,
    return_direct=True,
    sql_cmd_parser=output_parser,
    database=SQLDatabase(engine, None, metadata),
)

import pandas as pd

pd.DataFrame(
    chain.run(
        "Please give me 10 papers to ask what is PageRank?",
        callbacks=[StdOutCallbackHandler()],
    )
)

SQL Database as Retriever

In [ ]:

from langchain.chains.qa_with_sources.retrieval import RetrievalQAWithSourcesChain
from langchain_experimental.retrievers.vector_sql_database import (
    VectorSQLDatabaseChainRetriever,
)
from langchain_experimental.sql.prompt import MYSCALE_PROMPT
from langchain_experimental.sql.vector_sql import (
    VectorSQLDatabaseChain,
    VectorSQLRetrieveAllOutputParser,
)
from langchain_openai import ChatOpenAI

output_parser_retrieve_all = VectorSQLRetrieveAllOutputParser.from_embeddings(
    output_parser.model
)

chain = VectorSQLDatabaseChain.from_llm(
    llm=OpenAI(openai_api_key=OPENAI_API_KEY, temperature=0),
    prompt=MYSCALE_PROMPT,
    top_k=10,
    return_direct=True,
    db=SQLDatabase(engine, None, metadata),
    sql_cmd_parser=output_parser_retrieve_all,
    native_format=True,
)

# You need all those keys to get docs
retriever = VectorSQLDatabaseChainRetriever(
    sql_db_chain=chain, page_content_key="abstract"
)

document_with_metadata_prompt = PromptTemplate(
    input_variables=["page_content", "id", "title", "authors", "pubdate", "categories"],
    template="Content:\n\tTitle: {title}\n\tAbstract: {page_content}\n\tAuthors: {authors}\n\tDate of Publication: {pubdate}\n\tCategories: {categories}\nSOURCE: {id}",
)

chain = RetrievalQAWithSourcesChain.from_chain_type(
    ChatOpenAI(
        model_name="gpt-3.5-turbo-16k", openai_api_key=OPENAI_API_KEY, temperature=0.6
    ),
    retriever=retriever,
    chain_type="stuff",
    chain_type_kwargs={
        "document_prompt": document_with_metadata_prompt,
    },
    return_source_documents=True,
)
ans = chain(
    "Please give me 10 papers to ask what is PageRank?",
    callbacks=[StdOutCallbackHandler()],
)
print(ans["answer"])

In [ ]: