app.py

import streamlit as st
import fitz  # PyMuPDF
from transformers import AutoTokenizer, AutoModelForCausalLM

# Function to read and extract text from a PDF document
def read_pdf(file_path):
    text = ""
    with fitz.open(file_path) as doc:
        for page in doc:
            text += page.get_text()
    return text

# Load the document text
document_text = read_pdf("jeff_wo.pdf")  # Adjust path to your PDF file

# Streamlit UI
st.title("LLaMA 2-based Q&A System")
st.write("### Enter your query below:")
query = st.text_input("Query")

# Load tokenizer and model
tokenizer = AutoTokenizer.from_pretrained("NousResearch/Llama-2-7b-chat-hf")
model = AutoModelForCausalLM.from_pretrained("NousResearch/Llama-2-7b-chat-hf")

# Function to get answers using the LLaMA 2 model
def get_answer(context, query):
    input_text = f"Context: {context}\nQ: {query}\nA:"
    input_ids = tokenizer.encode(input_text, return_tensors="pt")
    # Generate an answer to the query
    output = model.generate(input_ids, max_length=512, num_return_sequences=1)
    answer = tokenizer.decode(output[0], skip_special_tokens=True)
    return answer

# Button to generate answers
if st.button("Get Answer"):
    with st.spinner("Finding the answer..."):
        answer = get_answer(document_text, query)
    st.write("### Answer:")
    st.write(answer)

# Ensure to replace `path/to/your/document.pdf` with the actual path to the PDF in your repo