app.py

import gradio as gr
import os

from langchain import OpenAI, ConversationChain
from langchain.prompts import PromptTemplate
from langchain.text_splitter import CharacterTextSplitter
from langchain.vectorstores import Chroma
from langchain.docstore.document import Document
from langchain.embeddings import HuggingFaceInstructEmbeddings
from langchain.chains.conversation.memory import ConversationBufferMemory

from langchain.chains.conversation.memory import ConversationEntityMemory
from langchain.chains.conversation.prompt import ENTITY_MEMORY_CONVERSATION_TEMPLATE

from langchain import LLMChain

memory = ConversationBufferMemory(memory_key="chat_history")

persist_directory="db"
llm=OpenAI(model_name = "text-davinci-003", temperature=0)
vectordb = Chroma(persist_directory=persist_directory, embedding_function=embedding)
model_name = "hkunlp/instructor-large"
embed_instruction = "Represent the text from the BMW website for retrieval"
query_instruction = "Query the most relevant text from the BMW website"
embeddings = HuggingFaceInstructEmbeddings(model_name=model_name, embed_instruction=embed_instruction, query_instruction=query_instruction)
chain = RetrievalQAWithSourcesChain.from_chain_type(llm, chain_type="stuff", retriever=db.as_retriever(), memory=memory)

def chat(message, site,history):
    history = history or []
    response = ""
    try:
        response = chain.run(input=message)
    history.append((message, response))

    return history, history


with gr.Blocks() as demo:
    gr.Markdown("<h3><center>BMW Chat Bot</center></h3>")
    gr.Markdown("<p><center>Ask questions about BMW</center></p>")
    chatbot = gr.Chatbot()
    with gr.Row():
        inp = gr.Textbox(placeholder="Question",label =None)
        btn = gr.Button("Run").style(full_width=False)
        state = gr.State()
        agent_state = gr.State()
        btn.click(chat, [inp,site,state],[chatbot, state])
if __name__ == '__main__':
    demo.launch()