Update app.py

app.py

@@ -89,8 +89,10 @@ class WebpageContentProcessor:
         structured_chunks = []
         for i, node in enumerate(nodes):
             content = node.get_content(metadata_mode=MetadataMode.NONE).strip()
+            if not content:
+                continue
             title_match = re.match(r"^(#+)\s*(.*)", content)
-            title = title_match.group(2).strip() if title_match else (content.split('\n')[0][:70] + "...")
+            title = title_match.group(2).strip() if title_match and title_match.group(2).strip() else (content.split('\n')[0][:70] + "...")
             structured_chunks.append({"id": i, "title": title, "content": content})
         return structured_chunks
 
@@ -159,6 +161,7 @@ class ChunkManager:
 
     def delete_chunk(self, chunk_id: int):
         self._chunks = [c for c in self._chunks if c["id"] != chunk_id]
+        # Re-index remaining chunks to maintain sequential IDs
         for i, chunk in enumerate(self._chunks):
             chunk['id'] = i
 
@@ -172,19 +175,26 @@ class ChunkManager:
         self.target_grade_max = grade_max
         self.target_min_chunk_words = min_words
         self.target_max_chunk_words = max_words
-        self.set_chunks(self.get_chunks()) # Recalculate stats with new targets
+        # Recalculate stats for all chunks with new targets
+        self.set_chunks(self.get_chunks()) 
 
+# --- Streamlit UI ---
 st.set_page_config(layout="wide", page_title="Webpage Content Editor")
 
-# Initialize session state variables
-if 'chunk_manager' not in st.session_state:
-    st.session_state.chunk_manager = ChunkManager()
-if 'content_processor' not in st.session_state:
-    st.session_state.content_processor = WebpageContentProcessor()
-if 'selected_chunk_id' not in st.session_state:
-    st.session_state.selected_chunk_id = None
-if 'status_message' not in st.session_state:
-    st.session_state.status_message = ""
+# Initialize session state variables if they don't exist
+def init_session_state():
+    if 'chunk_manager' not in st.session_state:
+        st.session_state.chunk_manager = ChunkManager()
+    if 'content_processor' not in st.session_state:
+        st.session_state.content_processor = WebpageContentProcessor()
+    if 'selected_chunk_id' not in st.session_state:
+        st.session_state.selected_chunk_id = None
+    if 'status_message' not in st.session_state:
+        st.session_state.status_message = ""
+    if 'url_input' not in st.session_state:
+        st.session_state.url_input = ""
+
+init_session_state()
 
 processor = st.session_state.content_processor
 manager = st.session_state.chunk_manager
@@ -198,29 +208,31 @@ st.info(
     icon="ℹ️"
 )
 
-url_input = st.text_input("Enter a webpage URL to begin", key="url_input")
+url_input = st.text_input("Enter a webpage URL to begin", value=st.session_state.url_input, key="url_input_widget")
 
 if st.button("Process URL", use_container_width=True):
-    if url_input:
+    st.session_state.url_input = st.session_state.url_input_widget
+    if st.session_state.url_input:
         with st.spinner("Fetching and processing content..."):
-            markdown = processor.fetch_and_convert_to_markdown(url_input)
+            markdown = processor.fetch_and_convert_to_markdown(st.session_state.url_input)
             if "Error" in markdown:
                 st.session_state.status_message = markdown
                 manager.set_chunks([])
+                st.session_state.selected_chunk_id = None
             else:
                 chunks = processor.parse_markdown_into_chunks(markdown)
                 manager.set_chunks(chunks)
-                st.session_state.status_message = f"Successfully processed {len(chunks)} chunks." if chunks else "Could not extract content chunks."
-            
-            if manager.get_chunks():
-                st.session_state.selected_chunk_id = manager.get_chunks()[0]['id']
-            else:
-                st.session_state.selected_chunk_id = None
+                if chunks:
+                    st.session_state.status_message = f"Successfully processed {len(chunks)} chunks."
+                    st.session_state.selected_chunk_id = chunks[0]['id']
+                else:
+                    st.session_state.status_message = "Could not extract content chunks."
+                    st.session_state.selected_chunk_id = None
             st.rerun()
 
 if st.session_state.status_message:
     st.toast(st.session_state.status_message)
-    st.session_state.status_message = "" # Clear message after showing
+    st.session_state.status_message = ""
 
 tab1, tab2 = st.tabs(["Chunk Editor", "Settings & Overview"])
 
@@ -229,23 +241,28 @@ with tab1:
     if not chunks:
         st.write("Process a URL to start editing chunks.")
     else:
+        chunk_ids = [c['id'] for c in chunks]
         # Ensure selected_chunk_id is valid
-        if st.session_state.selected_chunk_id not in [c['id'] for c in chunks]:
-            st.session_state.selected_chunk_id = chunks[0]['id'] if chunks else None
+        if st.session_state.selected_chunk_id not in chunk_ids:
+            st.session_state.selected_chunk_id = chunk_ids[0] if chunk_ids else None
 
         if st.session_state.selected_chunk_id is not None:
             chunk_options = {c['id']: f"Chunk {c['id']}: {c['title']}" for c in chunks}
             
-            # This selectbox now directly manages `selected_chunk_id` in the session state.
-            # When a user makes a selection, Streamlit automatically updates the state and reruns the script.
-            st.selectbox(
+            # The selectbox's state is now managed directly by st.session_state.selected_chunk_id
+            # When the user selects a new option, Streamlit automatically updates this state variable and reruns the script.
+            selected_id = st.selectbox(
                 "Select a chunk to edit",
-                options=list(chunk_options.keys()),
+                options=chunk_ids,
                 format_func=lambda x: chunk_options.get(x, "Invalid Chunk"),
-                key="selected_chunk_id", # The key is now the session state variable itself
-                index=list(chunk_options.keys()).index(st.session_state.selected_chunk_id)
+                index=chunk_ids.index(st.session_state.selected_chunk_id)
             )
             
+            # Update the session state ONLY if the user selection has changed
+            if selected_id != st.session_state.selected_chunk_id:
+                st.session_state.selected_chunk_id = selected_id
+                st.rerun()
+
             selected_chunk = manager.get_chunk_by_id(st.session_state.selected_chunk_id)
 
             if selected_chunk:
@@ -255,7 +272,7 @@ with tab1:
                     "Chunk Content",
                     value=selected_chunk['content'],
                     height=300,
-                    key=f"editor_{selected_chunk['id']}" # Unique key forces re-render
+                    key=f"editor_{selected_chunk['id']}" # Unique key forces widget to re-render on selection change
                 )
 
                 col1, col2, _ = st.columns([1, 1, 4])
@@ -265,10 +282,8 @@ with tab1:
                     st.rerun()
                 
                 if col2.button("Delete Chunk", use_container_width=True, key=f"delete_{selected_chunk['id']}"):
-                    old_id = selected_chunk['id']
-                    manager.delete_chunk(old_id)
+                    manager.delete_chunk(selected_chunk['id'])
                     st.session_state.status_message = "Chunk deleted!"
-                    # Select the next available chunk or none if empty
                     remaining_chunks = manager.get_chunks()
                     st.session_state.selected_chunk_id = remaining_chunks[0]['id'] if remaining_chunks else None
                     st.rerun()