app.py

import streamlit as st
from streamlit_chat import message
from langchain.chains import ConversationalRetrievalChain
from langchain.embeddings import HuggingFaceEmbeddings
from langchain.llms import LlamaCpp
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.vectorstores import FAISS
from langchain.memory import ConversationBufferMemory
from langchain.document_loaders import PyPDFLoader
import os
import tempfile


icon = Image.open("chatbot.png")
icon = icon.resize((64, 64)) # You can adjust the size as per your requirement

# Set the page config with the image icon
st.set_page_config(page_title="Multi-PDF ChatBot", page_icon=icon)

def initialize_session_state():
    if 'history' not in st.session_state:
        st.session_state['history'] = []

    if 'generated' not in st.session_state:
        st.session_state['generated'] = ["Hello! Ask me anything about 🤗"]

    if 'past' not in st.session_state:
        st.session_state['past'] = ["Hey! 👋"]

def conversation_chat(query, chain, history):
    result = chain({"question": query, "chat_history": history})
    history.append((query, result["answer"]))
    return result["answer"]

def display_chat_history(chain):
    reply_container = st.container()
    container = st.container()
    user_input = st.chat_input(placeholder="Please describe your grievance here...", key='input')
    if st.button("What is CPRAM?", key="cpram_button"):
        with st.spinner('Generating response...'):
            output = conversation_chat("What is CPRAM?", chain, st.session_state['history'])
        st.session_state['past'].append("What is CPRAM?")
        st.session_state['generated'].append(output)
    elif st.button("How to fill grievance form?", key="grievance_button"):
        with st.spinner('Generating response...'):
            output = conversation_chat("How to fill grievance form?", chain, st.session_state['history'])
        st.session_state['past'].append("How to fill grievance form?")
        st.session_state['generated'].append(output)
    elif user_input:
        with st.spinner('Generating response...'):
            output = conversation_chat(user_input, chain, st.session_state['history'])
        st.session_state['past'].append(user_input)
        st.session_state['generated'].append(output)

    if st.session_state['generated']:
        with reply_container:
            for i in range(len(st.session_state['generated'])):
                message(st.session_state["past"][i], is_user=True, key=str(i) + '_user')
                message(st.session_state["generated"][i], key=str(i))

def create_conversational_chain(vector_store):
    # Create llm
    llm = LlamaCpp(
    streaming = True,
    model_path="stable-code-3b.Q3_K_S.gguf",  #model_path="mistral-7b-instruct-v0.1.Q2_K.gguf",
    temperature=0.5,
    top_p=1, 
    verbose=True,
    n_ctx=4096
    )
    
    memory = ConversationBufferMemory(memory_key="chat_history", return_messages=True)

    chain = ConversationalRetrievalChain.from_llm(llm=llm, chain_type='stuff',
                                                 retriever=vector_store.as_retriever(search_kwargs={"k": 2}),
                                                 memory=memory)
    return chain

def main():
    # Initialize session state
    initialize_session_state()
    st.title("🤖 CPRAM Grievance Support")
    # Initialize Streamlit
 
    # Create embeddings
    embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2", 
        model_kwargs={'device': 'cpu'})

    # Create vector store
    vector_store = FAISS.load_local("vector_data", embeddings)

        # Create the chain object
    chain = create_conversational_chain(vector_store)

        
    display_chat_history(chain)

if __name__ == "__main__":
    main()