app.py

import os
import pinecone
import openai

from langchain.embeddings.openai import OpenAIEmbeddings
# from langchain.text_splitter import CharacterTextSplitter
# from langchain.document_loaders import PyPDFLoader
from langchain.vectorstores import Pinecone
from langchain.chains import ConversationalRetrievalChain
from langchain.chat_models import ChatOpenAI
import streamlit as st


def _initialize_env():
    pinecone.init(
        api_key=st.secrets["pinecone_api_key"],
        environment=st.secrets["pinecone_env"]
    )
    openai.api_key = st.secrets["openai_api_key"]
    os.environ['OPENAI_API_KEY'] = st.secrets["openai_api_key"]


def _initialize_indexes():
    embeddings = OpenAIEmbeddings()

    # loader = PyPDFLoader("./etf-book.pdf")
    # documents = loader.load()
    # text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
    # docs = text_splitter.split_documents(documents)
    # db = Pinecone.from_documents(
    #     docs,
    #     embeddings,
    #     index_name=st.secrets["pinecone_table_name"]
    # )

    db_object = Pinecone.from_existing_index(
        st.secrets["pinecone_table_name"],
        embeddings
    )
    return db_object


def _initialize_retriever(db_object):
    retriever = db_object.as_retriever(search_type="similarity", search_kwargs={"k": 2})
    qa = ConversationalRetrievalChain.from_llm(
        ChatOpenAI(temperature=0),
        retriever
    )
    return qa


chat_history = []


def answer(user_input):
    if user_input == '':
        return
    global chat_history
    result = qa({
        "question": user_input,
        "chat_history": chat_history
    })
    chat_history.append((user_input, result["answer"]))
    chat_history = chat_history[-10:]
    st.write(
        "Bot: ",
        result["answer"]
    )


_initialize_env()
db = _initialize_indexes()
qa = _initialize_retriever(db)

question = st.text_input('Question')
answer(question)