app.py

import streamlit as st
from pypdf import PdfReader
import io
from gemini_kit import get_llm
from langchain_core.messages import HumanMessage

# Initialize session state for the PDF text and messages
if 'pdf' not in st.session_state:
    st.session_state.pdf = ""

if 'messages' not in st.session_state:
    st.session_state.messages = []
if 'extract' not in st.session_state:
    st.session_state.extract = True
def upload_pdf():
    print(st.session_state.extract)
    print(st.session_state.pdf[:10])
    uploaded_file = st.file_uploader("Choose a PDF file", type="pdf")
    if (uploaded_file is not None) & st.session_state.extract:
        st.write("Waiting for pdf to be extracted ...")

        pdf_reader = PdfReader(io.BytesIO(uploaded_file.read()))
        text = ""
        for page_num in range(len(pdf_reader.pages)):
            page = pdf_reader.pages[page_num]
            text += page.extract_text()
        
        # Store the extracted text in session state
        st.session_state.pdf = text
        st.session_state.extract = False
        st.write("PDF Text Extracted. You can chat now!!")
    if uploaded_file is None:
        st.session_state.extract = True
def chatbot_ui():
    user_input = st.text_input("You: ", "")
    if user_input:
        st.session_state.messages.append({"user": user_input})
        if st.session_state.pdf:
            response = generate_response(st.session_state.pdf, user_input)
        else:
            response = "Please upload a PDF to get started."
        st.session_state.messages.append({"Assistant": response})

    chat = st.button("Clear Chat")
    if chat:
        st.session_state.messages = []

    for message in st.session_state.messages:
        if "user" in message:
            st.markdown(f"**You:** {message['user']}")
        else:
            st.markdown(f"**Assistant:** {message['Assistant']}")

def generate_response(pdf, user_input):
    message = f"This is the text extracted from the pdf: {pdf}. The user query is {user_input}."
    llm = get_llm()
    try:
        response = llm.invoke(message).content
    except Exception as e:
        response = "Error occurred. This might be due to exhaustion of LLM quota or your PDF might be much bigger. The exact error: " + str(e)
    return response

def main():
    st.title("NCERT PDF Based AI Assistant")
    
    st.header("Upload a PDF")
    # Call upload_pdf() only when the file uploader is interacted with
    upload_pdf()

    st.header("Chatbot")
    chatbot_ui()

if __name__ == "__main__":
    main()