app.py

import streamlit as st
import os
import time
import re
import uuid
import threading
import sounddevice as sd
import numpy as np

from openai import OpenAI
from realtime_transcriber import WebSocketClient, connections, WEBSOCKET_URI, WEBSOCKET_HEADERS

# ------------------ 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 and voice")

# ------------------ 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)

# ------------------ 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 ------------------
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)

# ------------------ Section 1: Layout (Image + Chat) ------------------
col1, col2 = st.columns([1, 2])

# ----- Left Panel: Document Image -----
with col1:
    if show_image and st.session_state.image_url:
        st.image(st.session_state.image_url, caption="📑 Extracted Page", use_container_width=True)
        st.session_state.image_updated = False

# ----- Right Panel: Chat Assistant -----
with col2:
    prompt = st.chat_input("Type your question about the document...")

    # Pair user + assistant messages
    paired_messages = []
    buffer = []
    for msg in st.session_state.messages:
        buffer.append(msg)
        if msg["role"] == "assistant" and len(buffer) == 2:
            paired_messages.append(buffer.copy())
            buffer.clear()
    if buffer:
        paired_messages.append(buffer.copy())

    # Render messages from newest to oldest
    with st.container():
        for pair in reversed(paired_messages):
            for msg in pair:
                with st.chat_message(msg["role"]):
                    st.write(msg["content"])

    # Handle new prompt
    if prompt:
        st.session_state.messages.append({"role": "user", "content": prompt})

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

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

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

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

            messages = client.beta.threads.messages.list(thread_id=st.session_state.thread_id)
            assistant_message = None
            for message in reversed(messages.data):
                if message.role == "assistant":
                    assistant_message = message.content[0].text.value
                    break

            st.session_state.messages.append({"role": "assistant", "content": assistant_message})

            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

            st.rerun()

        except Exception as e:
            st.error(f"❌ Error: {str(e)}")

# ------------------ Section 3: Voice Transcription ------------------
st.markdown("---")
st.markdown("## 🎙️ Real-Time Voice Transcription")

# Init client
if "voice_client_id" not in st.session_state:
    client_id = str(uuid.uuid4())
    st.session_state.voice_client_id = client_id
    connections[client_id] = WebSocketClient(WEBSOCKET_URI, WEBSOCKET_HEADERS, client_id)
    threading.Thread(target=connections[client_id].run, daemon=True).start()

client_id = st.session_state.voice_client_id
transcript_placeholder = st.empty()

recording = st.checkbox("🎤 Start Recording")

if recording:
    st.warning("Recording is active. Speak clearly...")

    def audio_stream_callback(indata, frames, time_info, status):
        if status:
            print(f"⚠️ Audio status: {status}")
        if client_id in connections:
            connections[client_id].enqueue_audio_chunk(16000, indata.copy())
            transcript_placeholder.markdown(f"**Live Transcript:**\n\n{connections[client_id].transcript}")

    stream = sd.InputStream(callback=audio_stream_callback, channels=1, samplerate=16000)
    stream.start()
    st.session_state["stream"] = stream

else:
    if "stream" in st.session_state:
        st.session_state["stream"].stop()
        del st.session_state["stream"]
        st.success("Recording stopped.")

# Final transcript
if client_id in connections:
    st.markdown("**Final Transcript Output:**")
    st.markdown(connections[client_id].transcript)