app.py

import streamlit as st
from transformers import AutoTokenizer, AutoModelForCausalLM, TextIteratorStreamer
from huggingface_hub import login
from threading import Thread
import PyPDF2
import pandas as pd
import torch
import time

# Set page configuration
st.set_page_config(
    page_title="WizNerd Insp",
    page_icon="🚀",
    layout="centered"
)

MODEL_NAME = "amiguel/optimizedModelListing6.1"

# Title with rocket emojis
st.title("🚀 WizNerd Insp 🚀")

# Sidebar configuration
with st.sidebar:
    st.header("Authentication 🔒")
    hf_token = st.text_input("Hugging Face Token", type="password", 
                           help="Get your token from https://huggingface.co/settings/tokens")
    
    st.header("Upload Documents 📂")
    uploaded_file = st.file_uploader(
        "Choose a PDF or XLSX file",
        type=["pdf", "xlsx"],
        label_visibility="collapsed"
    )

# Initialize chat history
if "messages" not in st.session_state:
    st.session_state.messages = []

@st.cache_data
def process_file(uploaded_file):
    # Existing file processing logic
    pass

@st.cache_resource
def load_model(hf_token):
    try:
        if not hf_token:
            st.error("🔐 Authentication required! Please provide a Hugging Face token.")
            return None
        
        # Login to Hugging Face Hub
        login(token=hf_token)
        
        # Load tokenizer
        tokenizer = AutoTokenizer.from_pretrained(
            MODEL_NAME,
            token=hf_token
        )
        
        # Load model with KV caching support
        model = AutoModelForCausalLM.from_pretrained(
            MODEL_NAME,
            device_map="auto",
            torch_dtype=torch.float16,
            token=hf_token
        )
        
        return model, tokenizer
        
    except Exception as e:
        st.error(f"🤖 Model loading failed: {str(e)}")
        return None

# In the main chat handling section:
if prompt := st.chat_input("Ask your inspection question..."):
    if not hf_token:
        st.error("🔑 Authentication required!")
        st.stop()

    # Load model if not already loaded
    if "model" not in st.session_state:
        model_data = load_model(hf_token)
        if model_data is None:
            st.error("Failed to load model. Please check your token and try again.")
            st.stop()
            
        st.session_state.model, st.session_state.tokenizer = model_data
    
    model = st.session_state.model
    tokenizer = st.session_state.tokenizer
    
    
    # Add user message
    with st.chat_message("user", avatar="👤"):
        st.markdown(prompt)
    st.session_state.messages.append({"role": "user", "content": prompt})

    # Process file
    file_context = process_file(uploaded_file)
    
    # Generate response with KV caching
    if model and tokenizer:
        try:
            with st.chat_message("assistant", avatar="🤖"):
                start_time = time.time()
                streamer = generate_with_kv_cache(prompt, file_context, use_cache=True)
                
                response_container = st.empty()
                full_response = ""
                
                for chunk in streamer:
                    cleaned_chunk = chunk.replace("<think>", "").replace("</think>", "").strip()
                    full_response += cleaned_chunk + " "
                    response_container.markdown(full_response + "▌", unsafe_allow_html=True)
                
                # Display metrics
                end_time = time.time()
                st.caption(f"Generated in {end_time - start_time:.2f}s using KV caching")
                
                response_container.markdown(full_response)
                st.session_state.messages.append({"role": "assistant", "content": full_response})
                
        except Exception as e:
            st.error(f"⚡ Generation error: {str(e)}")
    else:
        st.error("🤖 Model not loaded!")