import chainlit as cl
import os
from dotenv import load_dotenv
from langchain_openai import OpenAIEmbeddings
from langchain_core.prompts import ChatPromptTemplate
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_community.vectorstores import Pinecone
from operator import itemgetter
from langchain.schema.runnable import RunnablePassthrough
from langchain_openai import ChatOpenAI
from langchain.schema.runnable.config import RunnableConfig
from langchain_core.output_parsers import StrOutputParser
from langchain_community.document_loaders import UnstructuredPDFLoader


load_dotenv()


RAG_PROMPT = """

CONTEXT:
{context}

QUERY:
{question}

You house builder and can only provide your answers from the context. 
You can only provide a response in danish

+++
Please provide sample text from the context next to your response.
+++

Don't tell in your response that you are getting it from the context.

"""


text_splitter = RecursiveCharacterTextSplitter(
    chunk_size = 1800,
    chunk_overlap = 50,
    length_function=len,
    is_separator_regex=True,
    separators=[
        "\n\n",
        "\n",
        " ",
        ".",
        ",",
        "\u200B",
        "\uff0c",
        "\u3001",
        "\uff0e",
        "\u3002",
        "",
    ],
)


loader = UnstructuredPDFLoader("br_femogfirs.pdf",  strategy="fast")
data = loader.load_and_split(text_splitter)

rag_prompt = ChatPromptTemplate.from_template(RAG_PROMPT)


@cl.on_chat_start
async def main():    
    
    user_env = await cl.AskUserMessage(content="Indsæt venligst din api-nøgle før vi kan gå videre:").send()

    if user_env:

            os.environ["OPENAI_API_KEY"] = user_env['output']

            await cl.Message(content=f"Din api nøgle er nu tilføjet for sessionen - nu kan du lave en forespørgsel!").send()

            model = ChatOpenAI(model="gpt-3.5-turbo")
            embedding_model = OpenAIEmbeddings(model="text-embedding-3-small")
            vector_store = Pinecone.from_documents(data, embedding_model, index_name="bygnings-regl-rag-1")
            retriever = vector_store.as_retriever()

            building_qa_chain = (
            {"context": itemgetter("question") | retriever, "question": itemgetter("question")}
            | RunnablePassthrough.assign(context=itemgetter("context"))
            | rag_prompt | model | StrOutputParser())

            cl.user_session.set("runnable", building_qa_chain)

 

@cl.on_message
async def on_message(message: cl.Message):
    runnable = cl.user_session.get("runnable")
    msg = cl.Message(content="")

    async for chunk in runnable.astream(
        {"question":message.content},
        config=RunnableConfig(callbacks=[cl.LangchainCallbackHandler()]),
    ):
        await msg.stream_token(chunk)