Update app.py

app.py

@@ -5,12 +5,13 @@ import re
 import requests
 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 DocAIA", layout="wide")
+st.set_page_config(page_title="Schlaeger Forrestdale DocAIA", layout="wide", initial_sidebar_state="collapsed")
 st.title("📄 Schlaeger Forrestdale Document Assistant")
-st.caption("Explore technical construction documents via AI with OCR references")
+st.caption("Explore City of Armadale construction documents using AI + OCR 🧠")
 
 # ------------------ Load API Key and Assistant ID ------------------
 OPENAI_API_KEY = os.environ.get("OPENAI_API_KEY")
@@ -31,30 +32,27 @@ 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
-if "search_history" not in st.session_state:
-    st.session_state.search_history = []
-if "run_in_progress" not in st.session_state:
-    st.session_state.run_in_progress = False
+if "pending_prompt" not in st.session_state:
+    st.session_state.pending_prompt = None
 
-# ------------------ Sidebar (Left Utility Panel) ------------------
+# ------------------ Sidebar ------------------
 st.sidebar.header("ℹ️ Information")
 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.session_state.search_history = []
-    st.session_state.run_in_progress = False
+    st.session_state.pending_prompt = None
     st.rerun()
 
 show_image = st.sidebar.toggle("📑 Show Page Image", value=True)
 
-st.sidebar.markdown("### 🔍 Navigate or Query Document")
-keyword = st.sidebar.text_input("Jump to clause keyword or term", placeholder="e.g. defects, payment, WHS")
+st.sidebar.subheader("📘 Document Tools")
+st.sidebar.markdown("Use the tools below to locate relevant clauses and actions:")
+
+keyword = st.sidebar.text_input("Search by Keyword", placeholder="e.g. defects, WHS, delay")
 if st.sidebar.button("🔎 Search Keyword") and keyword:
-    st.session_state.search_history.append(keyword)
-    if not st.session_state.run_in_progress:
-        st.session_state.messages.append({"role": "user", "content": f"Find clauses or references related to: {keyword}"})
+    st.session_state.pending_prompt = f"Find clauses or references related to: {keyword}"
 
 section_options = [
     "Select a section...",
@@ -62,15 +60,15 @@ section_options = [
     "2. Offer and Acceptance",
     "3. Key Personnel",
     "4. Contract Pricing",
-    "5. Contract Specifications",
+    "5. Specifications",
     "6. WHS Policies",
     "7. Penalties and Delays",
     "8. Dispute Resolution",
     "9. Principal Obligations"
 ]
-selected_section = st.sidebar.selectbox("📄 Browse Sections", section_options)
-if selected_section and selected_section != section_options[0]:
-    st.session_state.messages.append({"role": "user", "content": f"Summarize or list key points from section: {selected_section}"})
+section_select = st.sidebar.selectbox("📄 Jump to Section", section_options)
+if section_select != section_options[0]:
+    st.session_state.pending_prompt = f"Summarize or list key points from section: {section_select}"
 
 actions = [
     "Select an action...",
@@ -80,75 +78,93 @@ actions = [
     "Find delay-related penalties",
     "Extract dispute resolution steps"
 ]
-selected_action = st.sidebar.selectbox("⚙️ Common Actions", actions)
-if selected_action and selected_action != actions[0]:
-    st.session_state.messages.append({"role": "user", "content": selected_action})
-
-if st.session_state.search_history:
-    st.sidebar.markdown("---")
-    st.sidebar.markdown("#### 🔁 Recent Searches")
-    for term in reversed(st.session_state.search_history[-5:]):
-        st.sidebar.markdown(f"- {term}")
-
-if show_image and st.session_state.image_url:
-    st.sidebar.markdown("---")
-    st.sidebar.markdown("#### 🖼️ Page Image (Toggle)")
-    try:
-        response = requests.get(st.session_state.image_url)
-        response.raise_for_status()
-        img = Image.open(BytesIO(response.content))
-        st.sidebar.image(img, caption="OCR Page", use_container_width=True)
-    except Exception as e:
-        st.sidebar.error(f"❗ Failed to load image: {e}")
-
-# ------------------ Center Panel: Chat ------------------
-st.markdown("### 🧠 Ask a Document-Specific Question")
-prompt = st.chat_input("Example: What is the defects liability period?")
-
-for msg in st.session_state.messages:
-    with st.chat_message(msg["role"]):
-        st.markdown(msg["content"], unsafe_allow_html=True)
-
-if prompt and not st.session_state.run_in_progress:
-    st.session_state.messages.append({"role": "user", "content": prompt})
-    st.session_state.run_in_progress = True
-
-    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("🤖 Parsing and responding with referenced content..."):
-            while True:
-                run_status = client.beta.threads.runs.retrieve(thread_id=st.session_state.thread_id, run_id=run.id)
-                if run_status.status in ("completed", "failed", "cancelled"):
-                    break
-                time.sleep(1)
-
-        if run_status.status != "completed":
-            st.error(f"⚠️ Assistant failed: {run_status.status}")
-        else:
-            messages = client.beta.threads.messages.list(thread_id=st.session_state.thread_id)
-            for message in reversed(messages.data):
-                if message.role == "assistant":
-                    assistant_message = message.content[0].text.value
-                    st.session_state.messages.append({"role": "assistant", "content": assistant_message})
-                    break
-
-            match = re.search(r'Document Reference:\s+(.+?),\s+Page\s+(\d+)', assistant_message)
-            if match:
-                doc_name = match.group(1).strip()
-                page = int(match.group(2))
-                page_str = f"{page:04d}"
-                image_url = f"https://raw.githubusercontent.com/AndrewLORTech/c2ozschlaegerforrestdale/main/{doc_name}/{doc_name}_page_{page_str}.png"
-                st.session_state.image_url = image_url
-                st.session_state.image_updated = True
-
-        st.session_state.run_in_progress = False
-        st.rerun()
-    except Exception as e:
-        st.session_state.run_in_progress = False
-        st.error(f"❌ Error: {e}")
+action_select = st.sidebar.selectbox("⚙️ Common Contract Queries", actions)
+if action_select != actions[0]:
+    st.session_state.pending_prompt = action_select
+
+# ------------------ Layout: Chat + Image ------------------
+chat_col, image_col = st.columns([2, 1])
+
+# ------------------ Chat Interface ------------------
+with chat_col:
+    st.markdown("### 🧠 Ask a Document-Specific Question")
+    user_prompt = st.chat_input("Example: What is the defects liability period?")
+
+    # Use pending prompt from sidebar if no new chat prompt
+    if user_prompt:
+        st.session_state.messages.append({"role": "user", "content": user_prompt})
+    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("🤖 Parsing and responding with referenced content..."):
+                while True:
+                    run_status = client.beta.threads.runs.retrieve(
+                        thread_id=st.session_state.thread_id,
+                        run_id=run.id
+                    )
+                    if run_status.status in ("completed", "failed", "cancelled"):
+                        break
+                    time.sleep(1)
+
+            if run_status.status != "completed":
+                st.error(f"⚠️ Assistant failed: {run_status.status}")
+            else:
+                messages = client.beta.threads.messages.list(thread_id=st.session_state.thread_id)
+                for message in reversed(messages.data):
+                    if message.role == "assistant":
+                        assistant_reply = message.content[0].text.value
+                        st.session_state.messages.append({"role": "assistant", "content": assistant_reply})
+
+                        # Parse Document Reference and Page, then construct image URL with encoding
+                        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/"
+                                f"{folder}/{folder}_page_{page_str}.png"
+                            )
+                            st.session_state.image_url = image_url
+                            st.session_state.image_updated = True
+                        break
+
+            st.rerun()
+        except Exception as e:
+            st.error(f"❌ Error: {e}")
+
+    for msg in st.session_state.messages:
+        with st.chat_message(msg["role"]):
+            st.markdown(msg["content"], unsafe_allow_html=True)
+
+# ------------------ Image Display ------------------
+with image_col:
+    if show_image and st.session_state.image_url:
+        with st.spinner("Loading document preview..."):
+            try:
+                response = requests.get(st.session_state.image_url)
+                response.raise_for_status()
+                img = Image.open(BytesIO(response.content))
+                st.image(img, caption="📄 OCR Page Image", use_container_width=True)
+                st.session_state.image_updated = False
+            except Exception as e:
+                st.error(f"❗ Failed to load image: {e}")