# rag_chain.py
import os
import weaviate
from weaviate import Client
from weaviate import WeaviateClient
from weaviate.connect import ConnectionParams
from weaviate.auth import AuthApiKey
from weaviate.auth import AuthCredentials
from weaviate.classes.init import Auth
from langchain.vectorstores import Weaviate
from langchain.chains import RetrievalQA
from langchain.chat_models import ChatOpenAI
from langchain.embeddings import OpenAIEmbeddings

def build_rag_chain():
    # 1. Weaviate 클라이언트
    auth = weaviate.AuthApiKey(api_key="01js3q6y7twaxccm5dbh3se9bt")

    OPENAI_API_KEY="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"
    
    # http_config = {
    #     "host": "183.111.96.67",
    #     "port": 32668,
    #     "secure": False
    # }
    # grpc_config = {
    #     "host": "183.111.96.67",
    #     "port": 32619,
    #     "secure": False
    # }

    # connection_params = ConnectionParams(
    #     http=http_config,
    #     grpc=grpc_config,
    #     auth_credentials=auth
    # )

    # client = WeaviateClient(connection_params=connection_params)
  
    # client2 = weaviate.connect_to_custom(
    #     http_host="183.111.96.67",
    #     http_port=32668,
    #     grpc_host="183.111.96.67",
    #     http_secure=False,
    #     grpc_port=32619,
    #     grpc_secure=False,
    #     auth_credentials=auth,  # 인증이 필요 없으면 생략 가능
    #     headers={"X-OpenAI-Api-Key": OPENAI_API_KEY}  # 필요시
    # )

    # if client.is_ready():
    #     print("Weaviate 연결 성공!")
    # else:
    #     print("연결 실패. 서버 상태를 확인하세요.")

    # if client2.is_ready():
    #     print("Weaviate 2 연결 성공!")
    # else:
    #     print("연결 실패 2. 서버 상태를 확인하세요.")

  
    # client3 = weaviate.connect_to_weaviate_cloud(
    #     cluster_url="s4x71b3lt32z3md1ya0yw.c0.asia-southeast1.gcp.weaviate.cloud",  # 예: "http://183.111.96.67:32668"
    #     auth_credentials=Auth.api_key("0ojxypfXy2Y1rKw3KQdBUtwcWSrYPWitE7Qu")  
    # )
  
    # if client3.is_ready():
    #     print("Weaviate 3 연결 성공!")
    # else:
    #     print("연결 실패 3. 서버 상태를 확인하세요.")


    # auth_key = AuthCredentials.api_key("01js3q6y7twaxccm5dbh3se9bt")  # 없으면 None
    # auth_cre= AuthCredentials.api_key("01js3q6y7twaxccm5dbh3se9bt")

    # 커넥션 설정 (http, grpc 둘 다)
    auth_client_secret=Auth.api_key("01js3q6y7twaxccm5dbh3se9bt")
    connection_params = ConnectionParams.from_params(
        http_host="183.111.96.67",
        http_port=32668,
        http_secure=False,
        grpc_host="183.111.96.67",
        grpc_port=32619,  # 예시 포트, 실제 grpc 포트 넣어야 함
        grpc_secure=False
    )

    client4 = WeaviateClient(connection_params=connection_params, auth_client_secret=auth_client_secret)

    if client4.is_ready():
        print("Weaviate 4 연결 성공!")
    else:
        print("연결 실패 4. 서버 상태를 확인하세요.")

    import weaviate as legacy_weaviate
    legacy_client = legacy_weaviate.Client(connection_params=client4.connection_params)  # langchain용


    # 2. 벡터스토어
    vectorstore = Weaviate(
        client=legacy_client,
        index_name="LangDocs",
        text_key="text",
        embedding=OpenAIEmbeddings(openai_api_key=OPENAI_API_KEY)
    )

    # 3. HuggingFace LLM (예: mistralai/Mistral-7B-Instruct-v0.2)
    llm = ChatOpenAI(temperature=0, openai_api_key=OPENAI_API_KEY)

    retriever = vectorstore.as_retriever()

    # 4. RetrievalQA chain 구성

    qa_chain = RetrievalQA.from_chain_type(llm=llm, retriever=retriever)

    client.close()
    return qa_chain