fix: add export to xml and html (#17)

* added the XML export Signed-off-by: Peter Staar <[email protected]> * reformatted all Signed-off-by: Peter Staar <[email protected]> * fixed tests Signed-off-by: Peter Staar <[email protected]> * added the DocumentTokens class Signed-off-by: Peter Staar <[email protected]> * updating the to-xml method Signed-off-by: Peter Staar <[email protected]> * updating the to-xml method Signed-off-by: Peter Staar <[email protected]> * fixed the to-md method Signed-off-by: Peter Staar <[email protected]> * added the strict-text in the to-md method Signed-off-by: Peter Staar <[email protected]> * added page-tokens Signed-off-by: Peter Staar <[email protected]> * updated the location/page tokens Signed-off-by: Peter Staar <[email protected]> * small fix to have correct special document-tokens Signed-off-by: Peter Staar <[email protected]> * reformatted the code Signed-off-by: Peter Staar <[email protected]> --------- Signed-off-by: Peter Staar <[email protected]>
DS4SD · Sep 18, 2024 · 9bc256e · 9bc256e
1 parent 2f55d92
commit 9bc256e
Showing 1 changed file with 4 additions and 4 deletions.
diff --git a/docling_core/types/doc/document.py b/docling_core/types/doc/document.py
@@ -410,21 +410,21 @@ def get_special_tokens(
         special_tokens = [token.value for token in cls]
 
         # Adding dynamically generated row and col tokens
-        for i in range(0, max_rows):
+        for i in range(0, max_rows + 1):
             special_tokens += [f"<row_{i}>", f"</row_{i}>"]
 
-        for i in range(0, max_cols):
+        for i in range(0, max_cols + 1):
             special_tokens += [f"<col_{i}>", f"</col_{i}>"]
 
         for i in range(6):
             special_tokens += [f"<section-header-{i}>", f"</section-header-{i}>"]
 
         # Adding dynamically generated page-tokens
-        for i in range(0, max_pages):
+        for i in range(0, max_pages + 1):
             special_tokens.append(f"<page_{i}>")
 
         # Adding dynamically generated location-tokens
-        for i in range(0, max(page_dimension[0], page_dimension[1])):
+        for i in range(0, max(page_dimension[0] + 1, page_dimension[1] + 1)):
             special_tokens.append(f"<loc_{i}>")
 
         return special_tokens