fix: make sure page_items are sorted by page_no

Signed-off-by: Yusik Kim <kmyusk@gmail.com>
This commit is contained in:
Yusik Kim 2025-03-20 16:59:35 +01:00
parent 38a23eb50b
commit 3a09ca50bb

View File

@ -38,6 +38,7 @@ def remove_doctags_content(doctags: str, images: list[PILImage.Image]) -> str:
page_items[page_no].append(ser.serialize(item=item)) page_items[page_no].append(ser.serialize(item=item))
else: else:
page_items[page_no] = [ser.serialize(item=item)] page_items[page_no] = [ser.serialize(item=item)]
pages = [ser.serialize_page(parts=parts) for parts in page_items.values()] sorted_items = [page_items[key] for key in sorted(page_items.keys())]
pages = [ser.serialize_page(parts=parts) for parts in sorted_items]
return ser.serialize_doc(pages=pages).text return ser.serialize_doc(pages=pages).text