app.py

import streamlit as st
import os
import time
import re
import json
import requests
from PIL import Image
from openai import OpenAI

# ------------------ App Configuration ------------------
st.set_page_config(page_title="Document AI Assistant", layout="wide")
st.title("📄 Document AI Assistant")
st.caption("Chat with an AI Assistant on your medical/pathology documents")

# ------------------ Load API Key and Assistant ID ------------------
OPENAI_API_KEY = os.environ.get("OPENAI_API_KEY")
ASSISTANT_ID = os.environ.get("ASSISTANT_ID")

if not OPENAI_API_KEY or not ASSISTANT_ID:
    st.error("Missing secrets. Please ensure both OPENAI_API_KEY and ASSISTANT_ID are set in your Hugging Face Space secrets.")
    st.stop()

client = OpenAI(api_key=OPENAI_API_KEY)

# ------------------ Load Structured JSON ------------------
STRUCTURED_JSON_PATH = "51940670-Manual-of-Surgical-Pathology-Third-Edition_1_structured_output.json"
try:
    with open(STRUCTURED_JSON_PATH, "r") as f:
        structured_data = json.load(f)
except Exception as e:
    st.error(f"❌ Failed to load structured summary file: {e}")
    st.stop()

# ------------------ Session State Initialization ------------------
if "messages" not in st.session_state:
    st.session_state.messages = []
if "thread_id" not in st.session_state:
    st.session_state.thread_id = None
if "image_url" not in st.session_state:
    st.session_state.image_url = None
if "image_updated" not in st.session_state:
    st.session_state.image_updated = False

# ------------------ Sidebar Controls ------------------
st.sidebar.header("🔧 Settings")
if st.sidebar.button("🔄 Clear Chat"):
    st.session_state.messages = []
    st.session_state.thread_id = None
    st.session_state.image_url = None
    st.session_state.image_updated = False
    st.rerun()

show_image = st.sidebar.checkbox("📖 Show Document Image", value=True)

# ------------------ Assistant Query Function ------------------
def query_assistant(prompt):
    st.session_state.messages.insert(0, {"role": "user", "content": prompt})  # insert at top

    try:
        if st.session_state.thread_id is None:
            thread = client.beta.threads.create()
            st.session_state.thread_id = thread.id

        thread_id = st.session_state.thread_id

        client.beta.threads.messages.create(
            thread_id=thread_id,
            role="user",
            content=prompt
        )

        run = client.beta.threads.runs.create(
            thread_id=thread_id,
            assistant_id=ASSISTANT_ID
        )

        with st.spinner("Assistant is thinking..."):
            while True:
                run_status = client.beta.threads.runs.retrieve(
                    thread_id=thread_id,
                    run_id=run.id
                )
                if run_status.status == "completed":
                    break
                time.sleep(1)

        messages = client.beta.threads.messages.list(thread_id=thread_id)
        for message in reversed(messages.data):
            if message.role == "assistant":
                assistant_message = message.content[0].text.value
                st.session_state.messages.insert(0, {"role": "assistant", "content": assistant_message})

                # Extract GitHub image URL if available
                image_match = re.search(
                    r'https://raw\.githubusercontent\.com/AndrewLORTech/surgical-pathology-manual/main/[\w\-/]*\.png',
                    assistant_message
                )
                if image_match:
                    st.session_state.image_url = image_match.group(0)
                    st.session_state.image_updated = True
                return assistant_message

    except Exception as e:
        st.error(f"❌ Error: {str(e)}")
        return None

# ------------------ Layout ------------------
left, center = st.columns([1, 2])

# ------------------ Center Column: Chat UI with Static Input on Top ------------------
with center:
    st.subheader("💬 Document AI Assistant")

    # Static Chat Input Bar
    with st.container():
        prompt = st.text_input("💡 Ask a question about the document:", key="chat_input")
        if prompt:
            query_assistant(prompt)
            st.experimental_rerun()

    # Show messages: latest at top
    for message in st.session_state.messages:
        role = message["role"]
        with st.chat_message(role):
            st.markdown(message["content"])

# ------------------ Left Column: Document Image ------------------
with left:
    st.subheader("📄 Document Image")
    if show_image and st.session_state.image_url:
        try:
            image = Image.open(requests.get(st.session_state.image_url, stream=True).raw)
            st.image(image, caption="📑 Extracted Page", use_container_width=True)
            st.session_state.image_updated = False
        except Exception as e:
            st.warning("⚠️ Could not load image.")