docling 2.28.1__tar.gz → 2.28.2__tar.gz

{docling-2.28.1 → docling-2.28.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: docling
-Version: 2.28.1
+Version: 2.28.2
 Summary: SDK and CLI for parsing PDF, DOCX, HTML, and more, to a unified document representation for powering downstream workflows such as gen AI applications.
 Home-page: https://github.com/docling-project/docling
 License: MIT

{docling-2.28.1 → docling-2.28.2}/docling/backend/html_backend.py

@@ -206,9 +206,9 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
         hlevel = int(element.name.replace("h", ""))
         text = element.text.strip()
 
-        if hlevel == 1:
-            self.content_layer = ContentLayer.BODY
+        self.content_layer = ContentLayer.BODY
 
+        if hlevel == 1:
             for key in self.parents.keys():
                 self.parents[key] = None
 
@@ -243,7 +243,7 @@ class HTMLDocumentBackend(DeclarativeDocumentBackend):
             self.parents[hlevel] = doc.add_heading(
                 parent=self.parents[hlevel - 1],
                 text=text,
-                level=hlevel,
+                level=hlevel - 1,
                 content_layer=self.content_layer,
             )
 

{docling-2.28.1 → docling-2.28.2}/docling/backend/md_backend.py

@@ -212,9 +212,16 @@ class MarkdownDocumentBackend(DeclarativeDocumentBackend):
             traverse(element)
             snippet_text = "".join(strings)
             if len(snippet_text) > 0:
-                parent_item = doc.add_text(
-                    label=doc_label, parent=parent_item, text=snippet_text
-                )
+                if doc_label == DocItemLabel.SECTION_HEADER:
+                    parent_item = doc.add_heading(
+                        text=snippet_text,
+                        level=element.level - 1,
+                        parent=parent_item,
+                    )
+                else:
+                    parent_item = doc.add_text(
+                        label=doc_label, parent=parent_item, text=snippet_text
+                    )
 
         elif isinstance(element, marko.block.List):
             has_non_empty_list_items = False
@@ -232,12 +239,15 @@ class MarkdownDocumentBackend(DeclarativeDocumentBackend):
                     label=label, name=f"list", parent=parent_item
                 )
 
-        elif isinstance(element, marko.block.ListItem) and len(element.children) > 0:
+        elif (
+            isinstance(element, marko.block.ListItem)
+            and len(element.children) > 0
+            and isinstance((first_child := element.children[0]), marko.block.Paragraph)
+        ):
             self._close_table(doc)
             self._process_inline_text(parent_item, doc)
             _log.debug(" - List item")
 
-            first_child = element.children[0]
             snippet_text = str(first_child.children[0].children)  # type: ignore
             is_numbered = False
             if (

{docling-2.28.1 → docling-2.28.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "docling"
-version = "2.28.1"  # DO NOT EDIT, updated automatically
+version = "2.28.2"  # DO NOT EDIT, updated automatically
 description = "SDK and CLI for parsing PDF, DOCX, HTML, and more, to a unified document representation for powering downstream workflows such as gen AI applications."
 authors = [
   "Christoph Auer <cau@zurich.ibm.com>",