Update app.py

app.py

@@ -10,12 +10,19 @@ from urllib.parse import quote
 from openai import OpenAI
 
 # ------------------ Authentication ------------------
+VALID_USERS = {
+    "[email protected]": "Pass.123",
+    "[email protected]": "Pass.123",
+    "[email protected]": "Pass.123",
+    "[email protected]": "Pass.123",
+}
+
 def login():
     st.title("🔐 Login Required")
     email = st.text_input("Email")
     password = st.text_input("Password", type="password")
     if st.button("Login"):
-        if email == "[email protected]" and password == "Pass.123":
+        if VALID_USERS.get(email) == password:
             st.session_state.authenticated = True
             st.rerun()
         else:
@@ -33,100 +40,147 @@ st.set_page_config(page_title="Schlager Forrestdale DocAIAssist", layout="wide",
 st.title("📄 Schlager Forrestdale Document Assistant")
 st.caption("Explore City of Armadale construction documents using AI + OCR 🧐")
 
-# ------------------ Tabs ------------------
-tab1, tab2 = st.tabs(["📑 Contract", "📁 Technical"])
-
-# ------------------ API Key ------------------
+# ------------------ Load API Key ------------------
 OPENAI_API_KEY = os.environ.get("OPENAI_API_KEY")
 if not OPENAI_API_KEY:
-    st.error("❌ OPENAI_API_KEY missing in environment variables.")
+    st.error("❌ Missing OPENAI_API_KEY environment variable.")
     st.stop()
+
 client = OpenAI(api_key=OPENAI_API_KEY)
 
+# ------------------ Tabs ------------------
+tab1, tab2 = st.tabs(["📑 Contract", "📁 Technical"])
+
 # ------------------ Contract Tab ------------------
 with tab1:
     ASSISTANT_ID = "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
-
-    user_input = st.chat_input("Ask about contract clauses, terms, or obligations")
-    if user_input:
-        st.session_state.contract_messages.append({"role": "user", "content": user_input})
-
-    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
-            )
-
-            with st.spinner("🤖 Thinking..."):
-                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)
+    for key in ["messages", "thread_id", "image_url", "image_updated", "pending_prompt"]:
+        if key not in st.session_state:
+            st.session_state[key] = None if "url" in key else []
 
-            if run_status.status == "completed":
-                messages = client.beta.threads.messages.list(thread_id=st.session_state.contract_thread_id)
-                for msg in reversed(messages.data):
-                    if msg.role == "assistant":
-                        content = msg.content[0].text.value
-                        st.session_state.contract_messages.append({"role": "assistant", "content": content})
-
-                        match = re.search(r'Document Reference:\s*(.*?),\s*Page\s*(\d+)', content)
-                        if match:
-                            doc_name = match.group(1).strip()
-                            page = int(match.group(2))
-                            page_str = f"{page:04d}"
-                            folder = quote(doc_name)
-                            img_url = f"https://raw.githubusercontent.com/AndrewLORTech/c2ozschlaegerforrestdale/main/{folder}/{folder}_page_{page_str}.png"
-                            st.session_state.contract_image_url = img_url
-                        break
-        except Exception as e:
-            st.error(f"Error: {e}")
+    with st.sidebar:
+        st.header("ℹ️ Contract Tools")
+        if st.button("🧹 Clear Chat"):
+            for k in ["messages", "thread_id", "image_url", "image_updated", "pending_prompt"]:
+                st.session_state[k] = None if "url" in k else []
+            st.rerun()
 
-    for m in st.session_state.contract_messages:
-        with st.chat_message(m["role"]):
-            st.markdown(m["content"], unsafe_allow_html=True)
+        show_image = st.toggle("📑 Show Page Image", value=True)
+        keyword = st.text_input("Search by Keyword", placeholder="e.g. defects, WHS, delay")
+        if st.button("🔎 Search Keyword") and keyword:
+            st.session_state.pending_prompt = f"Find clauses or references related to: {keyword}"
+
+        section_options = [
+            "Select a section...",
+            "1. Formal Instrument of Contract",
+            "2. Offer and Acceptance",
+            "3. Key Personnel",
+            "4. Contract Pricing",
+            "5. Specifications",
+            "6. WHS Policies",
+            "7. Penalties and Delays",
+            "8. Dispute Resolution",
+            "9. Principal Obligations"
+        ]
+        section = st.selectbox("📄 Jump to Section", section_options)
+        if section != section_options[0]:
+            st.session_state.pending_prompt = f"Summarize or list key points from section: {section}"
+
+        actions = [
+            "Select an action...",
+            "List all contractual obligations",
+            "Summarize payment terms",
+            "List WHS responsibilities",
+            "Find delay-related penalties",
+            "Extract dispute resolution steps"
+        ]
+        action = st.selectbox("⚙️ Common Queries", actions)
+        if action != actions[0]:
+            st.session_state.pending_prompt = action
+
+    chat_col, image_col = st.columns([2, 1])
+    with chat_col:
+        st.markdown("### 🧠 Ask a Document-Specific Question")
+        user_input = st.chat_input("Example: What is the defects liability period?")
+        if user_input:
+            st.session_state.messages.append({"role": "user", "content": user_input})
+        elif st.session_state.pending_prompt:
+            st.session_state.messages.append({"role": "user", "content": st.session_state.pending_prompt})
+            st.session_state.pending_prompt = None
+
+        if st.session_state.messages and st.session_state.messages[-1]["role"] == "user":
+            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=st.session_state.messages[-1]["content"]
+                )
+
+                run = client.beta.threads.runs.create(
+                    thread_id=st.session_state.thread_id,
+                    assistant_id=ASSISTANT_ID
+                )
+
+                with st.spinner("🤖 Thinking..."):
+                    while True:
+                        status = client.beta.threads.runs.retrieve(thread_id=st.session_state.thread_id, run_id=run.id)
+                        if status.status in ("completed", "failed", "cancelled"):
+                            break
+                        time.sleep(1)
+
+                if status.status == "completed":
+                    messages = client.beta.threads.messages.list(thread_id=st.session_state.thread_id)
+                    for m in reversed(messages.data):
+                        if m.role == "assistant":
+                            reply = m.content[0].text.value
+                            st.session_state.messages.append({"role": "assistant", "content": reply})
+                            match = re.search(r'Document Reference:\s*(.*?),\s*Page\s*(\d+)', reply)
+                            if match:
+                                doc, page = match.group(1).strip(), int(match.group(2))
+                                folder = quote(doc)
+                                img_url = f"https://raw.githubusercontent.com/AndrewLORTech/c2ozschlaegerforrestdale/main/{folder}/{folder}_page_{page:04d}.png"
+                                st.session_state.image_url = img_url
+                                st.session_state.image_updated = True
+                            break
+                else:
+                    st.error("❌ Assistant failed.")
+                st.rerun()
+            except Exception as e:
+                st.error(f"❌ Error: {e}")
 
-    if st.session_state.contract_image_url:
-        st.image(st.session_state.contract_image_url, caption="📄 Referenced Contract Page", use_container_width=True)
+        for msg in st.session_state.messages:
+            with st.chat_message(msg["role"]):
+                st.markdown(msg["content"], unsafe_allow_html=True)
+
+    with image_col:
+        if show_image and st.session_state.image_url:
+            try:
+                r = requests.get(st.session_state.image_url)
+                r.raise_for_status()
+                img = Image.open(BytesIO(r.content))
+                st.image(img, caption="📄 OCR Page Image", use_container_width=True)
+            except Exception as e:
+                st.error(f"🖼️ Image failed: {e}")
 
 # ------------------ Technical Tab ------------------
 with tab2:
     ASSISTANT_ID = "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 = []
-
     st.session_state.tech_lightbox = None
 
-    tech_prompt = st.chat_input("Ask about plans, drawings or components")
-    if tech_prompt:
-        st.session_state.tech_messages.append({"role": "user", "content": tech_prompt})
+    tech_input = st.chat_input("Ask about plans, drawings or components")
+    if tech_input:
+        st.session_state.tech_messages.append({"role": "user", "content": tech_input})
 
     if st.session_state.tech_messages and st.session_state.tech_messages[-1]["role"] == "user":
         try:
@@ -162,8 +216,7 @@ with tab2:
                         content = msg.content[0].text.value
                         st.session_state.tech_messages.append({"role": "assistant", "content": content})
                         try:
-                            json_data = json.loads(content.strip("`json "))
-                            st.session_state.tech_results = json_data
+                            st.session_state.tech_results = json.loads(content.strip("`json "))
                         except:
                             st.session_state.tech_results = []
                         break
@@ -172,16 +225,15 @@ with tab2:
 
     with st.expander("🔧 Options (Filter + Pagination)", expanded=False):
         disciplines = sorted(set(d.get("discipline", "") for d in st.session_state.tech_results))
-        selected_discipline = st.selectbox("🌍 Filter by discipline", ["All"] + disciplines)
+        selected = st.selectbox("🌍 Filter by discipline", ["All"] + disciplines)
         page_size = 8
-        page_num = st.number_input("Page", min_value=1, step=1, value=1)
+        page = st.number_input("Page", min_value=1, step=1, value=1)
 
     if st.session_state.tech_results:
         st.subheader("📂 Results")
-        filtered = [r for r in st.session_state.tech_results if selected_discipline == "All" or r.get("discipline") == selected_discipline]
-        paged = filtered[(page_num - 1) * page_size : page_num * page_size]
+        results = [r for r in st.session_state.tech_results if selected == "All" or r.get("discipline") == selected]
+        paged = results[(page - 1) * page_size : page * page_size]
         cols = st.columns(4)
-
         for i, item in enumerate(paged):
             with cols[i % 4]:
                 st.markdown(f"**📁 {item['drawing_number']} ({item['discipline']})**")
@@ -199,4 +251,4 @@ with tab2:
     else:
         for msg in st.session_state.tech_messages:
             with st.chat_message(msg["role"]):
-                st.markdown(msg["content"], unsafe_allow_html=True)
+                st.markdown(msg["content"], unsafe_allow_html=True)