app.py

import streamlit as st
import pandas as pd
import numpy as np
from llama_cpp import Llama
from huggingface_hub import hf_hub_download
import io
import os

# Page configuration
st.set_page_config(
    page_title="📊 LLM Data Analyzer",
    page_icon="📊",
    layout="wide",
    initial_sidebar_state="expanded"
)

# Custom CSS for better UI
st.markdown("""
    <style>
    .main {
        padding: 0rem 1rem;
    }
    .stTabs [data-baseweb="tab-list"] {
        gap: 2px;
    }
    </style>
    """, unsafe_allow_html=True)

# Title and description
st.title("📊 LLM Data Analyzer")
st.markdown("""
    Analyze your CSV/Excel files and chat with an AI assistant powered by Llama 2.
    This app runs on **free Hugging Face CPU** - response time ~5-10 seconds per query.
    """)

# Cache model loading to avoid reloading
@st.cache_resource
def load_llm_model():
    """Load Llama 2 model from Hugging Face Hub"""
    st.info("📥 Downloading model (first time only, ~4GB)... This may take 2-3 minutes.")
    
    try:
        model_path = hf_hub_download(
            repo_id="TheBloke/Llama-2-7B-Chat-GGUF",
            filename="llama-2-7b-chat.Q4_K_M.gguf"
        )
        
        llm = Llama(
            model_path=model_path,
            n_ctx=2048,
            n_threads=4,
            n_gpu_layers=0,  # CPU only (free tier)
            verbose=False
        )
        return llm
    except Exception as e:
        st.error(f"Error loading model: {e}")
        return None

# Load model
llm = load_llm_model()

if llm is None:
    st.error("Failed to load model. Please refresh the page.")
    st.stop()

st.success("✅ Model loaded successfully!")

# Create tabs
tab1, tab2, tab3 = st.tabs(["📤 Upload & Analyze", "💬 Chat", "📊 About"])

# ============================================================================
# TAB 1: Upload & Analyze
# ============================================================================
with tab1:
    st.header("📤 Upload and Analyze Data")
    
    uploaded_file = st.file_uploader(
        "Upload a CSV or Excel file",
        type=["csv", "xlsx", "xls"],
        help="Supported formats: CSV, Excel"
    )
    
    if uploaded_file is not None:
        st.success(f"✅ File uploaded: {uploaded_file.name}")
        
        # Read the file
        try:
            if uploaded_file.name.endswith('.csv'):
                df = pd.read_csv(uploaded_file)
            else:
                df = pd.read_excel(uploaded_file)
            
            # Display data preview
            st.subheader("📋 Data Preview")
            st.dataframe(df.head(10), use_container_width=True)
            
            # Display statistics
            st.subheader("📊 Data Statistics")
            col1, col2, col3 = st.columns(3)
            
            with col1:
                st.metric("Rows", df.shape[0])
            with col2:
                st.metric("Columns", df.shape[1])
            with col3:
                st.metric("Memory", f"{df.memory_usage(deep=True).sum() / 1024:.2f} KB")
            
            # Detailed statistics
            st.write(df.describe().T)
            
            # Ask AI about the data
            st.subheader("❓ Ask AI About Your Data")
            question = st.text_input(
                "What would you like to know about this data?",
                placeholder="e.g., What is the average value in column X?"
            )
            
            if question:
                with st.spinner("🤔 AI is analyzing your data..."):
                    # Create prompt
                    data_summary = df.describe().to_string()
                    prompt = f"""You are a data analyst expert. You have the following data summary:

{data_summary}

Column names: {', '.join(df.columns.tolist())}

User's question: {question}

Please provide a clear, concise analysis based on the data summary. Focus on actionable insights."""
                    
                    # Generate response
                    response = llm(
                        prompt,
                        max_tokens=300,
                        stop=["\n\nUser:", "Question:"],
                        echo=False,
                        temperature=0.7
                    )
                    
                    answer = response['choices'][0]['text'].strip()
                    st.success("✅ Analysis Complete")
                    st.write(answer)
        
        except Exception as e:
            st.error(f"Error reading file: {e}")

# ============================================================================
# TAB 2: Chat
# ============================================================================
with tab2:
    st.header("💬 Chat with AI Assistant")
    st.write("Have a conversation with Llama 2. Ask anything!")
    
    # Initialize session state for chat history
    if "messages" not in st.session_state:
        st.session_state.messages = []
    
    # Display chat history
    for message in st.session_state.messages:
        with st.chat_message(message["role"]):
            st.markdown(message["content"])
    
    # Chat input
    user_input = st.chat_input("Type your message here...")
    
    if user_input:
        # Add user message to history
        st.session_state.messages.append({"role": "user", "content": user_input})
        
        # Display user message
        with st.chat_message("user"):
            st.markdown(user_input)
        
        # Generate AI response
        with st.chat_message("assistant"):
            with st.spinner("⏳ Generating response..."):
                prompt = f"""You are a helpful AI assistant. The user asks: {user_input}

Provide a clear, helpful, and concise response."""
                
                response = llm(
                    prompt,
                    max_tokens=300,
                    stop=["\n\nUser:", "User:"],
                    echo=False,
                    temperature=0.7
                )
                
                assistant_message = response['choices'][0]['text'].strip()
                st.markdown(assistant_message)
                
                # Add assistant message to history
                st.session_state.messages.append({
                    "role": "assistant",
                    "content": assistant_message
                })

# ============================================================================
# TAB 3: About
# ============================================================================
with tab3:
    st.header("ℹ️ About This App")
    
    st.markdown("""
    ### 🎯 What is this?
    
    **LLM Data Analyzer** is an AI-powered tool for analyzing data and having conversations with an intelligent assistant.
    
    ### 🔧 Technology Stack
    
    - **Model:** Llama 2 7B (quantized to 4-bit)
    - **Framework:** Llama.cpp (CPU inference)
    - **Frontend:** Streamlit
    - **Hosting:** Hugging Face Spaces (Free Tier)
    
    ### ⚡ Performance
    
    - **Speed:** ~5-10 tokens per second (free CPU)
    - **Context:** 2048 tokens max
    - **Model Size:** 4GB (quantized)
    - **Hardware:** Free tier CPU
    
    ### 💡 Use Cases
    
    1. **Data Analysis**: Upload CSV/Excel and ask questions
    2. **Chat**: General conversation with AI
    3. **Learning**: Understand your data better
    
    ### 🚀 Faster Version Available
    
    For **GPU acceleration** (70+ tokens/sec):
    - Run locally on Apple Silicon Mac using MLX
    - Upgrade to Hugging Face PRO tier
    - Deploy on GPU-enabled cloud servers
    
    ### 📝 Tips
    
    - Keep questions focused and specific for best results
    - First request takes longer (model loading)
    - Data is processed locally, not stored on server
    
    ### 🔗 Links
    
    - [GitHub Repository](#) - Source code
    - [Hugging Face Hub](#) - Model info
    - [Llama.cpp](#) - Inference engine
    
    ---
    
    **Version:** 1.0 | **Last Updated:** Dec 2025
    """)