Update app.py

app.py

@@ -15,6 +15,7 @@ db_name = os.environ.get("MONGO_DB", "grant_docs")
 mongo_uri = f"mongodb+srv://{user}:{password}@{cluster}/{db_name}?retryWrites=true&w=majority&tls=true&tlsAllowInvalidCertificates=true"
 client = MongoClient(mongo_uri, tls=True, tlsAllowInvalidCertificates=True, serverSelectionTimeoutMS=20000)
 db = client[db_name]
+st.set_page_config(page_title="Doc Chunker", layout="wide")
 def gate_ui():
     APP_PASSWORD=st.secrets.get("APP_PASSWORD", os.getenv("APP_PASSWORD")).strip()
     if "authed" not in st.session_state:
@@ -35,81 +36,82 @@ def gate_ui():
             st.error("Incorrect password.")
     return False
 # === Streamlit UI ===
-if not gate_ui():
-    return
-st.set_page_config(page_title="Doc Chunker", layout="wide")
-st.title("📄 Document Chunker & Uploader")
-
-with st.sidebar:
-    st.header("Settings")
-
-    # Fetch collection names for dropdown
-    try:
-        existing_categories = db["final_chunks"].distinct("collection_category") or []
-    except Exception:
-        existing_categories = []
-    existing_categories=sorted([c for c in existing_categories if c])+["Create New Category"]
-    selected_category = st.selectbox(
-        "Choose Category (collection_category)",
-        existing_categories,
-        index=existing_categories.index("Create New Category") if "Create New Category" in existing_categories else 0
-    )
-    if selected_category == "Create New Category":
-        selected_category = st.sidebar.text_input("Enter Category Name:")
-        if not selected_category:
-            st.warning("⚠️ Enter a category name to proceed.")
-            st.stop()
-
-    is_grant_app = st.toggle("Is this a Grant Application?", value=False)
-
-uploaded_file = st.file_uploader("Upload a DOCX, TXT, or PDF file", type=["docx", "txt", "pdf"])
-
-if uploaded_file:
-    temp_path = Path(tempfile.gettempdir()) / uploaded_file.name
-    with open(temp_path, "wb") as f:
-        f.write(uploaded_file.getbuffer())
-
-    st.success(f"Uploaded `{uploaded_file.name}`")
-
-    modified_time = datetime.now().isoformat()
-    collection = db['final_chunks']
-    already = collection.find_one({
-        "metadata.title": uploaded_file.name,
-        "collection_category": selected_category
-    })
-
-    if already:
-        st.warning(f"⚠️ `{uploaded_file.name}` already exists in category `{selected_category}`. Skipping…")
-    else:
-        st.write("⏳ Processing with DocumentChunker...")
-        chunker = DocumentChunker()
-        chunks = chunker.process_document(str(temp_path))
-
-        if chunks:
-            for chunk in chunks:
-                chunk['collection_category']=selected_category
-                chunk['metadata'].update({
-                    "title": uploaded_file.name,
-                    "uploaded_at": modified_time,
-                    "is_grant_app": is_grant_app,
-                })
-                collection.insert_one(chunk)
-
-            st.success(f"✅ {len(chunks)} chunks inserted into `final_chunks` (category: `{selected_category}`)")
-
-            # Show a few previews
-            for i, c in enumerate(chunks[:3]):
-                st.subheader(f"Chunk {i+1}: {c['metadata'].get('header') or 'No Header'}")
-                st.markdown(c['text'][:400] + "...")
-                st.caption(f"Topics: {', '.join(c['metadata']['topics'])} | Category: {c['metadata']['category']}")
-                st.progress(c['metadata']['confidence_score'])
-
-            if len(chunks) > 3:
-                st.info(f"... and {len(chunks)-3} more chunks processed.")
-
+def main():
+    if not gate_ui():
+        return
+    st.title("📄 Document Chunker & Uploader")
+    
+    with st.sidebar:
+        st.header("Settings")
+    
+        # Fetch collection names for dropdown
+        try:
+            existing_categories = db["final_chunks"].distinct("collection_category") or []
+        except Exception:
+            existing_categories = []
+        existing_categories=sorted([c for c in existing_categories if c])+["Create New Category"]
+        selected_category = st.selectbox(
+            "Choose Category (collection_category)",
+            existing_categories,
+            index=existing_categories.index("Create New Category") if "Create New Category" in existing_categories else 0
+        )
+        if selected_category == "Create New Category":
+            selected_category = st.sidebar.text_input("Enter Category Name:")
+            if not selected_category:
+                st.warning("⚠️ Enter a category name to proceed.")
+                st.stop()
+    
+        is_grant_app = st.toggle("Is this a Grant Application?", value=False)
+    
+    uploaded_file = st.file_uploader("Upload a DOCX, TXT, or PDF file", type=["docx", "txt", "pdf"])
+    
+    if uploaded_file:
+        temp_path = Path(tempfile.gettempdir()) / uploaded_file.name
+        with open(temp_path, "wb") as f:
+            f.write(uploaded_file.getbuffer())
+    
+        st.success(f"Uploaded `{uploaded_file.name}`")
+    
+        modified_time = datetime.now().isoformat()
+        collection = db['final_chunks']
+        already = collection.find_one({
+            "metadata.title": uploaded_file.name,
+            "collection_category": selected_category
+        })
+    
+        if already:
+            st.warning(f"⚠️ `{uploaded_file.name}` already exists in category `{selected_category}`. Skipping…")
         else:
-            st.warning("⚠️ No chunks were generated.")
-
+            st.write("⏳ Processing with DocumentChunker...")
+            chunker = DocumentChunker()
+            chunks = chunker.process_document(str(temp_path))
+    
+            if chunks:
+                for chunk in chunks:
+                    chunk['collection_category']=selected_category
+                    chunk['metadata'].update({
+                        "title": uploaded_file.name,
+                        "uploaded_at": modified_time,
+                        "is_grant_app": is_grant_app,
+                    })
+                    collection.insert_one(chunk)
+    
+                st.success(f"✅ {len(chunks)} chunks inserted into `final_chunks` (category: `{selected_category}`)")
+    
+                # Show a few previews
+                for i, c in enumerate(chunks[:3]):
+                    st.subheader(f"Chunk {i+1}: {c['metadata'].get('header') or 'No Header'}")
+                    st.markdown(c['text'][:400] + "...")
+                    st.caption(f"Topics: {', '.join(c['metadata']['topics'])} | Category: {c['metadata']['category']}")
+                    st.progress(c['metadata']['confidence_score'])
+    
+                if len(chunks) > 3:
+                    st.info(f"... and {len(chunks)-3} more chunks processed.")
+    
+            else:
+                st.warning("⚠️ No chunks were generated.")
+if __name__ == "__main__":
+    main()
     # try:
     #     os.remove(temp_path)
     # except Exception as e: