Update app.py via AI Editor

app.py

@@ -96,41 +96,47 @@ def intelligent_pdf_split(input_path, session_dir, max_mb=5, min_split_mb=4):
     n_pages = len(reader.pages)
     splits = []
     current_writer = PdfWriter()
-    split_points = []
-    last_header = None
     last_split_at = 0
-    for i in range(n_pages):
-        page = reader.pages[i]
-        current_writer.add_page(page)
+    i = 0
+    last_header = None
+    force_split = False
+    while i < n_pages:
+        current_writer = PdfWriter()
+        part_start = i
+        current_writer.add_page(reader.pages[i])
         size = estimate_writer_size(current_writer) / (1024 * 1024)
-        header = extract_text_headers(reader, i)
-        blank = is_blank_page(reader, i)
-        chapter = is_chapter_header(header)
-        split_here = False
-
-        logging.debug(f"Page {i}: size={size:.2f}MB, header='{header}', blank={blank}, chapter={chapter}")
-
-        if size >= max_mb:
-            split_here = True
-        elif size >= min_split_mb:
-            if blank or chapter or (header and header != last_header):
-                split_here = True
-
-        if split_here:
-            splits.append((last_split_at, i+1))
-            logging.info(f"Splitting at pages {last_split_at}-{i+1} (size ~{size:.2f}MB)")
-            last_split_at = i+1
-            current_writer = PdfWriter()
-        last_header = header
-
-    if last_split_at < n_pages:
-        splits.append((last_split_at, n_pages))
-
+        if size > max_mb:
+            # Single page exceeds max_mb, must split after this page
+            splits.append((i, i+1))
+            i += 1
+            continue
+        # Add consecutive pages until approaching min_split_mb
+        j = i + 1
+        while j < n_pages:
+            tmp_writer = PdfWriter()
+            for k in range(part_start, j+1):
+                tmp_writer.add_page(reader.pages[k])
+            size = estimate_writer_size(tmp_writer) / (1024 * 1024)
+            if size > max_mb:
+                # Last added page makes it too big; break, split at previous
+                break
+            # If above min_split_mb, check for natural break
+            header = extract_text_headers(reader, j)
+            blank = is_blank_page(reader, j)
+            chapter = is_chapter_header(header)
+            if size >= min_split_mb and (blank or chapter or (header and header != last_header)):
+                # Good split point found
+                j += 1
+                break
+            last_header = header
+            j += 1
+        splits.append((part_start, j))
+        i = j
     split_files = []
     for idx, (start, end) in enumerate(splits):
         writer = PdfWriter()
-        for i in range(start, end):
-            writer.add_page(reader.pages[i])
+        for p in range(start, end):
+            writer.add_page(reader.pages[p])
         out_path = os.path.join(session_dir, f'split_part_{idx+1}.pdf')
         with open(out_path, 'wb') as f:
             writer.write(f)
@@ -303,7 +309,6 @@ def handle_upload(contents, filename, clear_n, delete_upload_n_list, split_n, de
     if session_data is None:
         session_data = {}
 
-    # Handle Clear Session
     if trigger == 'clear-session':
         logging.info(f"handle_upload: Clear session button pressed for {session_id}")
         clean_session(session_id)
@@ -311,7 +316,6 @@ def handle_upload(contents, filename, clear_n, delete_upload_n_list, split_n, de
         logging.info(f"Session cleared for {session_id}")
         return "", True, get_split_results_placeholder(), resp_data
 
-    # Handle Delete Upload (detect ANY delete button press)
     delete_pressed = False
     if isinstance(trigger, dict) and trigger.get('type') == 'delete-upload-btn':
         delete_pressed = True
@@ -332,7 +336,6 @@ def handle_upload(contents, filename, clear_n, delete_upload_n_list, split_n, de
         logging.info(f"Session files deleted for {session_id}")
         return "", True, get_split_results_placeholder(), {}
 
-    # Handle Delete Split File
     delete_split_pressed = False
     delete_split_idx = None
     if isinstance(trigger, dict) and trigger.get('type') == 'delete-split-btn':
@@ -354,7 +357,6 @@ def handle_upload(contents, filename, clear_n, delete_upload_n_list, split_n, de
                 logging.info(f"Deleted split file: {file_path} for session {session_id}")
             split_files = [f for i, f in enumerate(split_files) if i != delete_split_idx]
             session_data['split_files'] = split_files
-            # Recreate ZIP if any splits remain, else remove ZIP
             zip_path = os.path.join(session_dir, "split_files.zip")
             if split_files:
                 make_zip_of_splits(split_files, session_dir)
@@ -365,7 +367,6 @@ def handle_upload(contents, filename, clear_n, delete_upload_n_list, split_n, de
                     os.remove(zip_path)
                     logging.info(f"Deleted ZIP file as no splits remain for session {session_id}")
                 session_data['zip_ready'] = False
-            # Build UI
             orig_filename = session_data.get('orig_filename', '')
             file_info = dbc.Row([
                 dbc.Col(html.Div(f"Uploaded: {orig_filename}"), width=9, style={'display': 'flex', 'alignItems': 'center'}),
@@ -389,7 +390,6 @@ def handle_upload(contents, filename, clear_n, delete_upload_n_list, split_n, de
         else:
             logging.warning(f"Split file delete index {delete_split_idx} invalid for session {session_id}")
 
-    # Handle Upload
     if trigger == 'upload-pdf':
         logging.info(f"handle_upload: Upload triggered for filename={filename}, session_id={session_id}")
         if not contents:
@@ -426,7 +426,6 @@ def handle_upload(contents, filename, clear_n, delete_upload_n_list, split_n, de
             logging.error(f"Error processing PDF: {e}")
             return html.Div(f"Error: {e}", style={'color': 'red'}), True, get_split_results_placeholder(), {}
 
-    # Handle Split
     if trigger == 'split-btn':
         orig_filename = session_data.get('orig_filename')
         logging.info(f"handle_upload: Split button clicked for {session_id}, orig_filename={orig_filename}")
@@ -470,7 +469,6 @@ def handle_upload(contents, filename, clear_n, delete_upload_n_list, split_n, de
             logging.error(f"Error splitting PDF for session {session_id}: {e}")
             return html.Div(f"Error: {e}", style={'color': 'red'}), False, get_split_results_placeholder(), session_data
 
-    # Restore after split
     if session_data.get('split_files'):
         split_files = session_data['split_files']
         orig_filename = session_data.get('orig_filename', '')
@@ -494,7 +492,6 @@ def handle_upload(contents, filename, clear_n, delete_upload_n_list, split_n, de
         logging.info(f"handle_upload: Restoring split results for session {session_id}, {len(split_files)} files.")
         return file_info, False, results, session_data
 
-    # Restore after upload (before split)
     if session_data.get('orig_filename') and not session_data.get('split_files'):
         file_info = dbc.Row([
             dbc.Col(html.Div(f"Uploaded: {session_data['orig_filename']}"), width=9, style={'display': 'flex', 'alignItems': 'center'}),