docling-core 2.43.0__tar.gz → 2.43.1__tar.gz

{docling_core-2.43.0 → docling_core-2.43.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-core
-Version: 2.43.0
+Version: 2.43.1
 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-2.43.0 → docling_core-2.43.1}/docling_core/transforms/serializer/html.py

@@ -713,6 +713,23 @@ class HTMLListSerializer(BaseModel, BaseListSerializer):
             **kwargs,
         )
 
+        # Append nested list to parent list item:
+        i = 0
+        while i < len(parts):
+            prt = parts[i]
+            if prt.text.startswith(("<ul>", "<ol>")):
+                for j in range(i - 1, -1, -1):
+                    if parts[j].text.startswith(("<li>", "<li ")) and parts[
+                        j
+                    ].text.endswith("</li>"):
+                        before, _, _ = parts[j].text.rpartition("</li>")
+                        parts[j].text = f"{before}\n{prt.text}\n</li>"
+                        break
+                if j > -1:
+                    parts.pop(i)
+            else:
+                i += 1
+
         # Add all child parts
         text_res = "\n".join(
             [

{docling_core-2.43.0 → docling_core-2.43.1}/docling_core/transforms/visualizer/layout_visualizer.py

@@ -148,7 +148,9 @@ class LayoutVisualizer(BaseVisualizer):
         prev_image = None
         prev_page_nr = None
         for idx, (elem, _) in enumerate(
-            doc.iterate_items(included_content_layers=included_content_layers)
+            doc.iterate_items(
+                included_content_layers=included_content_layers, traverse_pictures=True
+            )
         ):
             if not isinstance(elem, DocItem):
                 continue

{docling_core-2.43.0 → docling_core-2.43.1}/docling_core.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-core
-Version: 2.43.0
+Version: 2.43.1
 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-2.43.0 → docling_core-2.43.1}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "docling-core"
-version = "2.43.0"  # DO NOT EDIT, updated automatically
+version = "2.43.1"  # DO NOT EDIT, updated automatically
 description = "A python library to define and validate data types in Docling."
 license = "MIT"
 license-files = ["LICENSE"]

{docling_core-2.43.0 → docling_core-2.43.1}/test/test_base.py

@@ -36,8 +36,8 @@ def test_identifier():
     )
 
     # schema_json(): no need to set by_alias since it is True by the default
-    tf = open("test/data/json_schemas/base_identifier.json", encoding="utf-8")
-    gold_json = json.load(tf)
+    with open("test/data/json_schemas/base_identifier.json", encoding="utf-8") as tf:
+        gold_json = json.load(tf)
 
     assert Identifier.model_json_schema() == gold_json
 

{docling_core-2.43.0 → docling_core-2.43.1}/test/test_docling_doc.py

@@ -1819,9 +1819,10 @@ def test_document_manipulation():
 
     # Test the handling of list items in insert_* methods, both with and without parent groups
 
-    li_sibling = doc.insert_list_item(
-        sibling=node, text="Inserted List Item, Incorrect Parent", after=False
-    )
+    with pytest.warns(DeprecationWarning, match="ListItem parent must be a ListGroup"):
+        li_sibling = doc.insert_list_item(
+            sibling=node, text="Inserted List Item, Incorrect Parent", after=False
+        )
     doc.insert_list_item(
         sibling=li_sibling, text="Inserted List Item, Correct Parent", after=True
     )
@@ -1831,12 +1832,13 @@ def test_document_manipulation():
         text="Inserted Text with LIST_ITEM Label, Correct Parent",
         after=False,
     )
-    doc.insert_text(
-        sibling=node,
-        label=DocItemLabel.LIST_ITEM,
-        text="Inserted Text with LIST_ITEM Label, Incorrect Parent",
-        after=True,
-    )
+    with pytest.warns(DeprecationWarning, match="ListItem parent must be a ListGroup"):
+        doc.insert_text(
+            sibling=node,
+            label=DocItemLabel.LIST_ITEM,
+            text="Inserted Text with LIST_ITEM Label, Incorrect Parent",
+            after=True,
+        )
 
     filename = Path(
         "test/data/doc/constructed_doc.inserted_list_items_with_insert_*.json"

{docling_core-2.43.0 → docling_core-2.43.1}/test/test_doctags_load.py

@@ -60,7 +60,8 @@ def test_doctags_load_from_files():
 
 def test_doctags_load_from_memory():
 
-    doctags = Path("test/data/doc/page_with_pic.dt").open("r").read()
+    with Path("test/data/doc/page_with_pic.dt").open() as file:
+        doctags = file.read()
     image = PILImage.open(Path("test/data/doc/page_with_pic.png"))
 
     doctags_doc = DocTagsDocument.from_doctags_and_image_pairs([doctags], [image])
@@ -75,7 +76,8 @@ def test_doctags_load_from_memory():
 
 
 def test_doctags_load_without_image():
-    doctags = Path("test/data/doc/page_with_pic.dt").open("r").read()
+    with Path("test/data/doc/page_with_pic.dt").open() as file:
+        doctags = file.read()
     doctags_doc = DocTagsDocument.from_doctags_and_image_pairs([doctags], None)
     doc = DoclingDocument.load_from_doctags(doctags_doc)
     exp = "test/data/doc/page_without_pic.dt.json"
@@ -86,7 +88,8 @@ def test_doctags_load_without_image():
 
 
 def test_doctags_load_for_kv_region():
-    doctags = Path("test/data/doc/doc_with_kv.dt").open("r").read()
+    with Path("test/data/doc/doc_with_kv.dt").open() as file:
+        doctags = file.read()
     image = PILImage.open(Path("test/data/doc/doc_with_kv.png"))
     doctags_doc = DocTagsDocument.from_doctags_and_image_pairs([doctags], [image])
     doc = DoclingDocument.load_from_doctags(doctags_doc)
@@ -98,7 +101,8 @@ def test_doctags_load_for_kv_region():
 
 
 def test_multipage_doctags_load():
-    doctags = Path("test/data/doc/2206.01062.yaml.dt").open("r").read()
+    with Path("test/data/doc/2206.01062.yaml.dt").open() as file:
+        doctags = file.read()
     doctags_doc = DocTagsDocument.from_multipage_doctags_and_images(doctags, None)
     doc = DoclingDocument.load_from_doctags(doctags_doc)
     exp = "test/data/doc/2206.01062.yaml.dt.json"