chat.py

import os

import openai
import streamlit as st
from dotenv import load_dotenv
from langchain.chains import RetrievalQA
from langchain.chat_models import ChatOpenAI
from langchain.embeddings.openai import OpenAIEmbeddings
from langchain.vectorstores import DeepLake
from streamlit_chat import message

# Load environment variables from a .env file (containing OPENAI_API_KEY)
load_dotenv()
# Set the title for the Streamlit app
st.title(os.environ.get('SITE_TITLE'))
# Set the OpenAI API key from the environment variable
openai.api_key = os.environ.get('OPENAI_API_KEY')
active_loop_data_set_path = os.environ.get('DEEPLAKE_DATASET_PATH')

# Create an instance of OpenAIEmbeddings
embeddings = OpenAIEmbeddings()

# Create an instance of DeepLake with the specified dataset path and embeddings
db = DeepLake(dataset_path=active_loop_data_set_path,
              read_only=True, embedding_function=embeddings)


def generate_response(prompt):
    # Generate a response using OpenAI's ChatCompletion API and the specified prompt
    completion = openai.ChatCompletion.create(
        model="gpt-4-32k",
        messages=[
            {"role": "user", "content": prompt}
        ])
    response = completion.choices[0].message.content
    return response


def get_text():
    # Create a Streamlit input field and return the user's input
    input_text = st.text_input("", key="input")
    return input_text


def search_db(query):
    # Create a retriever from the DeepLake instance
    retriever = db.as_retriever()
    # Set the search parameters for the retriever
    retriever.search_kwargs['distance_metric'] = 'cos'
    retriever.search_kwargs['fetch_k'] = 100
    retriever.search_kwargs['maximal_marginal_relevance'] = True
    retriever.search_kwargs['k'] = 10
    # Create a ChatOpenAI model instance
    model = ChatOpenAI(model='gpt-4-32k')
    # Create a RetrievalQA instance from the model and retriever
    qa = RetrievalQA.from_llm(model, retriever=retriever)
    # Return the result of the query
    return qa.run(query)


# Initialize the session state for generated responses and past inputs
if 'generated' not in st.session_state:
    st.session_state['generated'] = ['i am ready to help you ser']

if 'past' not in st.session_state:
    st.session_state['past'] = ['hello']

# Get the user's input from the text input field
user_input = get_text()

# If there is user input, search for a response using the search_db function
if user_input:
    output = search_db(user_input)
    st.session_state.past.append(user_input)
    st.session_state.generated.append(output)

# If there are generated responses, display the conversation using Streamlit messages
if st.session_state['generated']:
    for i in range(len(st.session_state['generated'])):
        message(st.session_state['past'][i], is_user=True, key=str(i) + '_user')
        # Wrap the bot response with markdown
        st.markdown(st.session_state["generated"][i])