app.py

import streamlit as st
import os
import logging

# Configure logging
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
logger = logging.getLogger("APP_ENTRY")

logger.info("🚀 app.py module loaded. Streamlit starting up...")

st.set_page_config(page_title="Gemini Research Assistant", layout="wide")
st.title("💎 Agentic RAG: Gemini 2.0 Research Assistant")

# --- AUTO-INGESTION SEQUENCE ---
# This ensures the vector DB exists before the agent tries to load it.
# --- CONFIGURATION ---
DB_PATH = "./chroma_db"
DATA_PATH = "./data"

@st.cache_resource(show_spinner=False)
def initialize_knowledge_base():
    """Checks and builds the vector database if missing."""
    if not os.path.exists(DB_PATH) or not os.listdir(DB_PATH):
        logger.info("⚠️ VectorDB not found. Checking for PDF data...")
        if os.path.exists(DATA_PATH) and any(f.endswith('.pdf') for f in os.listdir(DATA_PATH)):
            logger.info("📄 Data found. Starting ingestion process...")
            # We use a placeholder to show progress since st.spinner isn't thread-safe in early startup sometimes
            status_placeholder = st.empty()
            status_placeholder.info("🧠 Initializing Knowledge Base... Check Logs for progress.")
            
            from src.processor import build_index
            try:
                build_index(DATA_PATH, DB_PATH)
                status_placeholder.success("✅ Knowledge Base Built! Refreshing...")
                logger.info("✅ Ingestion complete.")
                status_placeholder.empty()
            except Exception as e:
                logger.error(f"❌ Ingestion FAILED: {e}")
                status_placeholder.error(f"Failed to build index: {e}")
                raise e
        else:
            logger.warning("No data found in 'data' directory.")
            st.warning("⚠️ No data found! Please add PDFs to the 'data' folder to use Local Research.")
    else:
        logger.info("✅ VectorDB exists. Skipping ingestion.")

# Run the initialization
initialize_knowledge_base()

# Lazy import agent AFTER DB check to prevent "Table not found" errors
logger.info("🤖 Loading Agent Logic...")
from src.agent import app as agent_app
logger.info("✅ Agent loaded. Ready to serve.")

if "messages" not in st.session_state:
    st.session_state.messages = []

# Display history
for msg in st.session_state.messages:
    with st.chat_message(msg["role"]):
        st.markdown(msg["content"])

# Chat input
if prompt := st.chat_input("Ask about internal docs or latest tech..."):
    st.session_state.messages.append({"role": "user", "content": prompt})
    with st.chat_message("user"):
        st.markdown(prompt)

    with st.chat_message("assistant"):
        inputs = {"messages": [("user", prompt)]}
        config = {"configurable": {"thread_id": "1"}}
        
        # Execute LangGraph brain
        response = agent_app.invoke(inputs, config=config)
        answer = response["messages"][-1].content
        
        st.markdown(answer)
        st.session_state.messages.append({"role": "assistant", "content": answer})