streamlit_app.py

import streamlit as st
import os
from datetime import datetime
import tempfile
from pathlib import Path

# Import our modules
from utils import initialize_rag_system, clear_session_state, format_sources
from config import MODEL_NAME, EMBEDDING_MODEL

# Page config
st.set_page_config(
    page_title="AI RAG Assistant",
    page_icon="🤖",
    layout="wide",
    initial_sidebar_state="expanded"
)

# Custom CSS
st.markdown("""
<style>
    .main-header {font-size: 2.5rem; font-weight: 700; color: #1f77b4;}
    .chat-message {padding: 1rem; border-radius: 1rem; margin: 1rem 0;}
    .user-message {background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white;}
    .assistant-message {background: linear-gradient(135deg, #f093fb 0%, #f5576c 100%);}
    .source-container {background: #f8f9fa; border-left: 4px solid #007bff; padding: 1rem;}
</style>
""", unsafe_allow_html=True)

# Header
st.markdown('<h1 class="main-header">🤖 AI RAG Assistant</h1>', unsafe_allow_html=True)
st.markdown("Upload your documents and chat with your data using advanced RAG powered by Llama-4-Scout")

# Footer with attribution
st.markdown(
    "**Built with** [![anycoder](https://img.shields.io/badge/Built%20with-anycoder-3b82f6?style=for-the-badge&logo=huggingface)](https://huggingface.co/spaces/akhaliq/anycoder)"
)

# Sidebar
with st.sidebar:
    st.header("⚙️ Settings")
    
    # Model selection
    model_name = st.selectbox(
        "Response Model", 
        [MODEL_NAME],
        help="Llama-4-Scout for powerful reasoning"
    )
    
    embedding_model = st.selectbox(
        "Embedding Model",
        [EMBEDDING_MODEL],
        help="bge-m3: State-of-the-art multilingual embeddings"
    )
    
    # Similarity threshold
    similarity_threshold = st.slider(
        "Similarity Threshold", 0.5, 0.95, 0.8,
        help="Minimum similarity score for relevant chunks"
    )
    
    # Max new tokens
    max_new_tokens = st.slider("Max Tokens", 200, 2000, 1000)
    
    st.divider()
    if st.button("🗑️ Clear Chat & Memory", type="secondary"):
        clear_session_state()
        st.rerun()

# Initialize session state
if "messages" not in st.session_state:
    st.session_state.messages = []
if "rag_system" not in st.session_state:
    st.session_state.rag_system = None
if "documents_processed" not in st.session_state:
    st.session_state.documents_processed = 0

# File upload section
uploaded_files = st.file_uploader(
    "📁 Upload Documents",
    type=['pdf', 'txt', 'md', 'docx', 'doc', 'pptx', 'ppt'],
    accept_multiple_files=True,
    help="Supports PDF, TXT, MD, DOCX, PPTX and more"
)

# Process uploaded files
if uploaded_files:
    with st.spinner("Processing documents... This may take a moment."):
        try:
            temp_dir = tempfile.mkdtemp()
            for file in uploaded_files:
                file_path = Path(temp_dir) / file.name
                with open(file_path, "wb") as f:
                    f.write(file.getbuffer())
            
            # Initialize or update RAG system
            st.session_state.rag_system = initialize_rag_system(
                temp_dir,
                model_name,
                embedding_model,
                similarity_threshold
            )
            st.session_state.documents_processed = len(uploaded_files)
            
            st.success(f"✅ Processed {len(uploaded_files)} documents successfully!")
            st.info(f"📊 {st.session_state.documents_processed} documents indexed and ready for querying")
            
        except Exception as e:
            st.error(f"❌ Error processing documents: {str(e)}")

# Status indicator
if st.session_state.rag_system is not None:
    col1, col2 = st.columns([3, 1])
    with col1:
        st.success(f"✅ Ready! {st.session_state.documents_processed} documents loaded")
    with col2:
        st.caption(f"Model: {model_name}")

# Chat interface
st.markdown("---")

# Display chat messages
for message in st.session_state.messages:
    with st.chat_message(message["role"]):
        st.markdown(message["content"])
        
        # Display sources for assistant messages
        if message["role"] == "assistant" and "sources" in message:
            with st.expander("📚 Sources", expanded=False):
                st.markdown(format_sources(message["sources"]))

# Chat input
if prompt := st.chat_input("Ask a question about your documents..."):
    # Add user message
    st.session_state.messages.append({"role": "user", "content": prompt})
    with st.chat_message("user"):
        st.markdown(prompt)

    # Generate response
    if st.session_state.rag_system is not None:
        with st.chat_message("assistant"):
            with st.spinner("Thinking..."):
                try:
                    # Query RAG system
                    response = st.session_state.rag_system.query(prompt)
                    
                    # Display response
                    st.markdown(response.response)
                    
                    # Store full response with sources
                    full_message = {
                        "role": "assistant",
                        "content": response.response,
                        "sources": response.source_nodes
                    }
                    st.session_state.messages.append(full_message)
                    
                except Exception as e:
                    st.error(f"Error generating response: {str(e)}")
    else:
        with st.chat_message("assistant"):
            st.warning("👆 Please upload and process documents first!")

# Instructions
with st.expander("ℹ️ How to use", expanded=False):
    st.markdown("""
    1. **Upload documents** (PDF, TXT, MD, DOCX, PPTX supported)
    2. **Wait for processing** (indexing happens automatically)
    3. **Ask questions** about your documents
    4. **Click sources** to see exact references
    
    **Features:**
    - Multi-document support
    - Advanced semantic search
    - Source citations
    - Adjustable similarity threshold
    - Streaming responses
    """)

# Performance metrics
if st.session_state.rag_system is not None:
    with st.expander("📈 System Info", expanded=False):
        col1, col2, col3 = st.columns(3)
        with col1:
            st.metric("Documents", st.session_state.documents_processed)
        with col2:
            st.metric("Model", MODEL_NAME.split('/')[-1])
        with col3:
            st.metric("Embedding", EMBEDDING_MODEL.split('/')[-1])