docling-core 2.34.0__tar.gz → 2.34.2__tar.gz

{docling_core-2.34.0 → docling_core-2.34.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-core
-Version: 2.34.0
+Version: 2.34.2
 Summary: A python library to define and validate data types in Docling.
 Author-email: Cesar Berrospi Ramis <ceb@zurich.ibm.com>, Panos Vagenas <pva@zurich.ibm.com>, Michele Dolfi <dol@zurich.ibm.com>, Christoph Auer <cau@zurich.ibm.com>, Peter Staar <taa@zurich.ibm.com>
 Maintainer-email: Panos Vagenas <pva@zurich.ibm.com>, Michele Dolfi <dol@zurich.ibm.com>, Christoph Auer <cau@zurich.ibm.com>, Peter Staar <taa@zurich.ibm.com>, Cesar Berrospi Ramis <ceb@zurich.ibm.com>

{docling_core-2.34.0 → docling_core-2.34.2}/docling_core/types/doc/document.py

@@ -1874,12 +1874,19 @@ class DoclingDocument(BaseModel):
 
         return item.get_ref()
 
-    def _delete_items(self, refs: list[RefItem]) -> bool:
+    def _delete_items(self, refs: list[RefItem]):
         """Delete document item using the self-reference."""
         to_be_deleted_items: dict[tuple[int, ...], str] = {}  # stack to cref
 
+        if not refs:
+            return
+
         # Identify the to_be_deleted_items
-        for item, stack in self._iterate_items_with_stack(with_groups=True):
+        for item, stack in self._iterate_items_with_stack(
+            with_groups=True,
+            traverse_pictures=True,
+            included_content_layers={c for c in ContentLayer},
+        ):
             ref = item.get_ref()
 
             if ref in refs:
@@ -1890,8 +1897,10 @@ class DoclingDocument(BaseModel):
                 if tuple(substack) in to_be_deleted_items:
                     to_be_deleted_items[tuple(stack)] = ref.cref
 
-        if len(to_be_deleted_items) == 0:
-            raise ValueError("Nothing to be deleted ...")
+        if len(to_be_deleted_items) < len(refs):
+            raise ValueError(
+                f"Cannot find all provided RefItems in doc: {[r.cref for r in refs]}"
+            )
 
         # Clean the tree, reverse the order to not have to update
         for stack_, ref_ in reversed(sorted(to_be_deleted_items.items())):
@@ -1931,8 +1940,6 @@ class DoclingDocument(BaseModel):
             node=self.body, refs_to_be_deleted=refs, lookup=lookup
         )
 
-        return True
-
     # Update the references
     def _update_ref_with_lookup(
         self, item_label: str, item_index: int, lookup: dict[str, dict[int, int]]
@@ -2167,7 +2174,10 @@ class DoclingDocument(BaseModel):
 
         """
         if not isinstance(parent, (OrderedList, UnorderedList)):
-            raise ValueError("ListItem's parent must be a list group")
+            warnings.warn("ListItem's parent must be a list group.", DeprecationWarning)
+
+        if not parent:
+            parent = self.body
 
         if not orig:
             orig = text
@@ -4270,11 +4280,12 @@ class DoclingDocument(BaseModel):
                     item.parent.resolve(doc=self), (OrderedList, UnorderedList)
                 )
             ):
-                # non_group_list_items.append(item)
-                if prev is None or not isinstance(prev, ListItem):  # if new list
-                    misplaced_list_items.append([item])
-                else:
+                if isinstance(prev, ListItem) and (
+                    prev.parent is None or prev.parent.resolve(self) == self.body
+                ):  # case of continuing list
                     misplaced_list_items[-1].append(item)
+                else:  # case of new list
+                    misplaced_list_items.append([item])
             prev = item
 
         for curr_list_items in reversed(misplaced_list_items):

{docling_core-2.34.0 → docling_core-2.34.2}/docling_core.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-core
-Version: 2.34.0
+Version: 2.34.2
 Summary: A python library to define and validate data types in Docling.
 Author-email: Cesar Berrospi Ramis <ceb@zurich.ibm.com>, Panos Vagenas <pva@zurich.ibm.com>, Michele Dolfi <dol@zurich.ibm.com>, Christoph Auer <cau@zurich.ibm.com>, Peter Staar <taa@zurich.ibm.com>
 Maintainer-email: Panos Vagenas <pva@zurich.ibm.com>, Michele Dolfi <dol@zurich.ibm.com>, Christoph Auer <cau@zurich.ibm.com>, Peter Staar <taa@zurich.ibm.com>, Cesar Berrospi Ramis <ceb@zurich.ibm.com>

{docling_core-2.34.0 → docling_core-2.34.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "docling-core"
-version = "2.34.0"  # DO NOT EDIT, updated automatically
+version = "2.34.2"  # DO NOT EDIT, updated automatically
 description = "A python library to define and validate data types in Docling."
 license = "MIT"
 license-files = ["LICENSE"]

{docling_core-2.34.0 → docling_core-2.34.2}/test/test_docling_doc.py

@@ -757,9 +757,6 @@ def _construct_doc() -> DoclingDocument:
     leading_list = doc.add_group(parent=None, label=GroupLabel.LIST)
     doc.add_list_item(parent=leading_list, text="item of leading list")
 
-    with pytest.raises(ValueError, match="list group"):
-        doc.add_list_item(text="Misplaced list item")
-
     title = doc.add_title(
         text="Title of the Document"
     )  # can be done if such information is present, or ommitted.
@@ -1080,6 +1077,9 @@ def _construct_doc() -> DoclingDocument:
 
     doc.add_list_item(text="Item 4 in A", enumerated=True, parent=parent_A)
 
+    with pytest.warns(DeprecationWarning, match="list group"):
+        doc.add_list_item(text="List item without parent list group")
+
     doc.add_text(label=DocItemLabel.TEXT, text="The end.", parent=None)
 
     return doc
@@ -1519,7 +1519,9 @@ def test_document_manipulation():
         DoclingDocument.load_from_json(filename=_gt_filename(filename=filename))
 
         # test if the document is the same as the stored GT
-        _verify_loaded_output(filename=filename, pred=doc)
+        _verify_loaded_output(
+            filename=filename, pred=DoclingDocument.model_validate(document)
+        )
 
     image_dir = Path("./test/data/doc/constructed_images/")