app.py

import streamlit as st
import os
import time
import re
import requests
import json
from PIL import Image
from io import BytesIO
from urllib.parse import quote
from openai import OpenAI

# ------------------ App Configuration ------------------
st.set_page_config(page_title="Schlaeger Forrestdale TechDocAIA", layout="wide", initial_sidebar_state="collapsed")
st.title("📄 Schlaeger Forrestdale Document Assistant")
st.caption("Explore City of Armadale construction documents using AI + OCR 🧐")

# ------------------ Load API Key ------------------
OPENAI_API_KEY = os.environ.get("OPENAI_API_KEY")
if not OPENAI_API_KEY:
    st.error("❌ Missing OPENAI_API_KEY. Please set it in Hugging Face Space secrets.")
    st.stop()

client = OpenAI(api_key=OPENAI_API_KEY)

# ------------------ Tabs Setup ------------------
tabs = st.tabs(["📑 Contract", "📐 Technical"])

# ------------------ Contract Tab ------------------
with tabs[0]:
    from urllib.parse import quote

    ASSISTANT_ID_CONTRACT = "asst_KsQRedoJUnEeStzfox1o06lO"

    if "contract_messages" not in st.session_state:
        st.session_state.contract_messages = []
    if "contract_thread_id" not in st.session_state:
        st.session_state.contract_thread_id = None
    if "contract_image_url" not in st.session_state:
        st.session_state.contract_image_url = None
    if "contract_image_updated" not in st.session_state:
        st.session_state.contract_image_updated = False
    if "contract_pending_prompt" not in st.session_state:
        st.session_state.contract_pending_prompt = None

    col1, col2 = st.columns([2, 1])
    with col1:
        st.markdown("### 🧠 Ask a Document-Specific Question")
        user_prompt = st.chat_input("Example: What is the defects liability period?")

        if user_prompt:
            st.session_state.contract_messages.append({"role": "user", "content": user_prompt})
        elif st.session_state.contract_pending_prompt:
            st.session_state.contract_messages.append({"role": "user", "content": st.session_state.contract_pending_prompt})
            st.session_state.contract_pending_prompt = None

        if st.session_state.contract_messages and st.session_state.contract_messages[-1]["role"] == "user":
            try:
                if st.session_state.contract_thread_id is None:
                    thread = client.beta.threads.create()
                    st.session_state.contract_thread_id = thread.id

                client.beta.threads.messages.create(
                    thread_id=st.session_state.contract_thread_id,
                    role="user",
                    content=st.session_state.contract_messages[-1]["content"]
                )

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

                with st.spinner("🤖 Parsing and responding with referenced content..."):
                    while True:
                        run_status = client.beta.threads.runs.retrieve(
                            thread_id=st.session_state.contract_thread_id,
                            run_id=run.id
                        )
                        if run_status.status in ("completed", "failed", "cancelled"):
                            break
                        time.sleep(1)

                if run_status.status == "completed":
                    messages = client.beta.threads.messages.list(thread_id=st.session_state.contract_thread_id)
                    for message in reversed(messages.data):
                        if message.role == "assistant":
                            assistant_reply = message.content[0].text.value
                            st.session_state.contract_messages.append({"role": "assistant", "content": assistant_reply})

                            match = re.search(r'Document Reference:\s*(.*?),\s*Page\s*(\d+)', assistant_reply)
                            if match:
                                doc_name = match.group(1).strip()
                                page = int(match.group(2))
                                page_str = f"{page:04d}"
                                folder = quote(doc_name)
                                image_url = f"https://raw.githubusercontent.com/AndrewLORTech/c2ozschlaegerforrestdale/main/{folder}/{folder}_page_{page_str}.png"
                                st.session_state.contract_image_url = image_url
                                st.session_state.contract_image_updated = True
                            break
                else:
                    st.error(f"⚠️ Assistant failed: {run_status.status}")
                st.rerun()

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

        for msg in st.session_state.contract_messages:
            with st.chat_message(msg["role"]):
                st.markdown(msg["content"], unsafe_allow_html=True)

    with col2:
        if st.session_state.contract_image_url:
            try:
                response = requests.get(st.session_state.contract_image_url)
                response.raise_for_status()
                img = Image.open(BytesIO(response.content))
                st.image(img, caption="📄 OCR Page Image", use_container_width=True)
            except Exception as e:
                st.error(f"❗ Failed to load image: {e}")

# ------------------ Technical Tab ------------------
with tabs[1]:
    ASSISTANT_ID_TECHNICAL = "asst_DjvuWBc7tCvMbAhY7n1em4BZ"

    if "tech_messages" not in st.session_state:
        st.session_state.tech_messages = []
    if "tech_thread_id" not in st.session_state:
        st.session_state.tech_thread_id = None
    if "tech_results" not in st.session_state:
        st.session_state.tech_results = []
    if "tech_lightbox_url" not in st.session_state:
        st.session_state.tech_lightbox_url = None

    user_prompt = st.chat_input("Ask about plans, drawings or components")
    if user_prompt:
        st.session_state.tech_messages.append({"role": "user", "content": user_prompt})

    if st.session_state.tech_messages and st.session_state.tech_messages[-1]["role"] == "user":
        try:
            if st.session_state.tech_thread_id is None:
                thread = client.beta.threads.create()
                st.session_state.tech_thread_id = thread.id

            client.beta.threads.messages.create(
                thread_id=st.session_state.tech_thread_id,
                role="user",
                content=st.session_state.tech_messages[-1]["content"]
            )

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

            with st.spinner("🤖 Parsing and responding..."):
                while True:
                    run_status = client.beta.threads.runs.retrieve(
                        thread_id=st.session_state.tech_thread_id,
                        run_id=run.id
                    )
                    if run_status.status in ("completed", "failed", "cancelled"):
                        break
                    time.sleep(1)

            if run_status.status == "completed":
                messages = client.beta.threads.messages.list(thread_id=st.session_state.tech_thread_id)
                for message in reversed(messages.data):
                    if message.role == "assistant":
                        assistant_reply = message.content[0].text.value
                        st.session_state.tech_messages.append({"role": "assistant", "content": assistant_reply})
                        try:
                            json_data = json.loads(assistant_reply.strip("`json "))
                            st.session_state.tech_results = json_data
                        except:
                            st.session_state.tech_results = []
                        break
            else:
                st.error(f"⚠️ Assistant failed: {run_status.status}")
            st.rerun()

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

    if st.session_state.tech_results:
        disciplines = sorted(set(d.get("discipline", "") for d in st.session_state.tech_results))
        selected_discipline = st.selectbox("🌍 Filter by discipline", ["All"] + disciplines)
        page_size = 8
        page_num = st.number_input("Page", min_value=1, step=1, value=1)

        filtered_results = [r for r in st.session_state.tech_results if selected_discipline == "All" or r.get("discipline") == selected_discipline]
        paged = filtered_results[(page_num - 1) * page_size : page_num * page_size]

        st.markdown("---")
        st.subheader("📂 Drawing Results")
        cols = st.columns(4)

        for i, item in enumerate(paged):
            with cols[i % 4]:
                st.markdown(f"**{item['drawing_number']}**")
                st.markdown(f"_Discipline: {item['discipline']}_")
                st.caption(item.get("summary", ""))
                for url in item.get("images", [])[:1]:
                    if st.button("🖼️ View Image", key=f"view_{i}"):
                        st.session_state.tech_lightbox_url = url

        if st.session_state.tech_lightbox_url:
            col_a, col_b = st.columns([1, 2])
            with col_b:
                st.image(st.session_state.tech_lightbox_url, caption="🔍 Enlarged Preview", use_container_width=True)
                if st.button("❌ Close Viewer"):
                    st.session_state.tech_lightbox_url = None
                    st.rerun()
    else:
        for msg in st.session_state.tech_messages:
            with st.chat_message(msg["role"]):
                st.markdown(msg["content"], unsafe_allow_html=True)