Fix tabby page slice parameters

ispras · Nov 21, 2023 · 9235cd5 · 9235cd5
1 parent d535b23
commit 9235cd5
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/dedoc/readers/pdf_reader/pdf_txtlayer_reader/pdf_tabby_reader.py b/dedoc/readers/pdf_reader/pdf_txtlayer_reader/pdf_tabby_reader.py
@@ -118,12 +118,12 @@ def __extract(self, path: str, parameters: dict, warnings: list)\
 
         # in java tabby reader page numeration starts with 1, end_page is included
         first_tabby_page = first_page + 1 if first_page is not None else 1
-        last_tabby_page = None if last_page is not None and last_page > page_count else last_page
+        last_tabby_page = page_count if (last_page is None) or (last_page is not None and last_page > page_count) else last_page
         self.logger.info(f"Reading PDF pages from {first_tabby_page} to {last_tabby_page}")
         document = self.__process_pdf(path=path, start_page=first_tabby_page, end_page=last_tabby_page)
 
         pages = document.get("pages", [])
-        for page in pages[first_page:last_page]:
+        for page in pages:
             page_lines = self.__get_lines_with_location(page, file_hash)
             if page_lines:
                 all_lines.extend(page_lines)