Fix tabby partially read

ispras · Nov 20, 2023 · 5e37d6c · 5e37d6c
1 parent fa396ef
commit 5e37d6c
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 4 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
@@ -117,11 +117,10 @@ def __extract(self, path: str, parameters: dict, warnings: list)\
                 return all_lines, all_tables, all_tables_on_images, all_attached_images, document_metadata
 
         # 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
-        # document = self.__process_pdf(path=path, start_page=first_tabby_page, end_page=last_tabby_page) TODO TLDR-518
+        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
+        document = self.__process_pdf(path=path, start_page=first_tabby_page, end_page=last_tabby_page)
 
-        document = self.__process_pdf(path=path)
         pages = document.get("pages", [])
         for page in pages[first_page:last_page]:
             page_lines = self.__get_lines_with_location(page, file_hash)

diff --git a/dedoc/readers/pdf_reader/pdf_txtlayer_reader/tabbypdf/jars/ispras_tbl_extr.jar b/dedoc/readers/pdf_reader/pdf_txtlayer_reader/tabbypdf/jars/ispras_tbl_extr.jar