Merge from main, update tests

Signed-off-by: Christoph Auer <[email protected]>

Author: cau-git

CHANGELOG.md

@@ -1,3 +1,34 @@
+## [v2.34.2](https://github.com/docling-project/docling-core/releases/tag/v2.34.2) - 2025-06-10
+
+### Fix
+
+* Fix doc traversal for item deletion ([#324](https://github.com/docling-project/docling-core/issues/324)) ([`076ad2b`](https://github.com/docling-project/docling-core/commit/076ad2ba9bd551daf1ee6a0141d360a7be8a18bd))
+
+## [v2.34.1](https://github.com/docling-project/docling-core/releases/tag/v2.34.1) - 2025-06-08
+
+### Fix
+
+* Warn when adding misplaced ListItem via API ([#321](https://github.com/docling-project/docling-core/issues/321)) ([`01b27b5`](https://github.com/docling-project/docling-core/commit/01b27b57b32c8a1a23e65ce3bf60cd1c027ca915))
+
+## [v2.34.0](https://github.com/docling-project/docling-core/releases/tag/v2.34.0) - 2025-06-06
+
+### Feature
+
+* **doctags:** Add enclosing bbox to inline ([#302](https://github.com/docling-project/docling-core/issues/302)) ([`dcc198f`](https://github.com/docling-project/docling-core/commit/dcc198f7c6231fe8c781abe4a83194be2ee8d23b))
+* Add subscript & superscript formatting ([#319](https://github.com/docling-project/docling-core/issues/319)) ([`ae96129`](https://github.com/docling-project/docling-core/commit/ae961299a5f729acecf1b2b346113ac23e8b97f0))
+* Add table annotations ([#304](https://github.com/docling-project/docling-core/issues/304)) ([`d8a5256`](https://github.com/docling-project/docling-core/commit/d8a5256b2cb654ceb35a70be1b656ac7463ad335))
+
+### Fix
+
+* Fix misplaced list items ([#317](https://github.com/docling-project/docling-core/issues/317)) ([`c383f64`](https://github.com/docling-project/docling-core/commit/c383f64c44b4e1eb760d19d9422948fea127331c))
+
+## [v2.33.1](https://github.com/docling-project/docling-core/releases/tag/v2.33.1) - 2025-06-04
+
+### Fix
+
+* New typer version with new click ([#315](https://github.com/docling-project/docling-core/issues/315)) ([`e17eabf`](https://github.com/docling-project/docling-core/commit/e17eabf0f92c3e3fde9f47fea083e44081bb9669))
+* Support section_header levels in doctags deserialization ([#313](https://github.com/docling-project/docling-core/issues/313)) ([`defd49e`](https://github.com/docling-project/docling-core/commit/defd49efae08c0cc2ce002f847724b6d65bd1407))
+
 ## [v2.33.0](https://github.com/docling-project/docling-core/releases/tag/v2.33.0) - 2025-06-02
 
 ### Feature

docling_core/transforms/serializer/base.py

@@ -202,6 +202,16 @@ def serialize_strikethrough(self, text: str, **kwargs: Any) -> str:
         """Hook for strikethrough formatting serialization."""
         ...
 
+    @abstractmethod
+    def serialize_subscript(self, text: str, **kwargs: Any) -> str:
+        """Hook for subscript formatting serialization."""
+        ...
+
+    @abstractmethod
+    def serialize_superscript(self, text: str, **kwargs: Any) -> str:
+        """Hook for superscript formatting serialization."""
+        ...
+
     @abstractmethod
     def serialize_hyperlink(
         self,
@@ -239,6 +249,15 @@ def serialize_captions(
         """Serialize the item's captions."""
         ...
 
+    @abstractmethod
+    def serialize_annotations(
+        self,
+        item: DocItem,
+        **kwargs: Any,
+    ) -> SerializationResult:
+        """Serialize the item's annotations."""
+        ...
+
     @abstractmethod
     def get_excluded_refs(self, **kwargs: Any) -> set[str]:
         """Get references to excluded items."""
@@ -257,3 +276,18 @@ class BaseSerializerProvider(ABC):
     def get_serializer(self, doc: DoclingDocument) -> BaseDocSerializer:
         """Get a the associated serializer."""
         ...
+
+
+class BaseAnnotationSerializer(ABC):
+    """Base class for annotation serializers."""
+
+    @abstractmethod
+    def serialize(
+        self,
+        *,
+        item: DocItem,
+        doc: DoclingDocument,
+        **kwargs: Any,
+    ) -> SerializationResult:
+        """Serializes the passed annotation."""
+        ...

docling_core/transforms/serializer/common.py

@@ -15,6 +15,7 @@
 from typing_extensions import Self, override
 
 from docling_core.transforms.serializer.base import (
+    BaseAnnotationSerializer,
     BaseDocSerializer,
     BaseFallbackSerializer,
     BaseFormSerializer,
@@ -30,6 +31,7 @@
 from docling_core.types.doc.document import (
     DOCUMENT_TOKENS_EXPORT_LABELS,
     ContentLayer,
+    DescriptionAnnotation,
     DocItem,
     DoclingDocument,
     FloatingItem,
@@ -41,9 +43,10 @@
     OrderedList,
     PictureClassificationData,
     PictureDataType,
-    PictureDescriptionData,
     PictureItem,
     PictureMoleculeData,
+    Script,
+    TableAnnotationType,
     TableItem,
     TextItem,
     UnorderedList,
@@ -122,7 +125,9 @@ def _iterate_items(
         yield item
 
 
-def _get_picture_annotation_text(annotation: PictureDataType) -> Optional[str]:
+def _get_annotation_text(
+    annotation: Union[PictureDataType, TableAnnotationType],
+) -> Optional[str]:
     result = None
     if isinstance(annotation, PictureClassificationData):
         predicted_class = (
@@ -132,7 +137,7 @@ def _get_picture_annotation_text(annotation: PictureDataType) -> Optional[str]:
         )
         if predicted_class is not None:
             result = predicted_class.replace("_", " ")
-    elif isinstance(annotation, PictureDescriptionData):
+    elif isinstance(annotation, DescriptionAnnotation):
         result = annotation.text
     elif isinstance(annotation, PictureMoleculeData):
         result = annotation.smi
@@ -211,6 +216,8 @@ class DocSerializer(BaseModel, BaseDocSerializer):
     list_serializer: BaseListSerializer
     inline_serializer: BaseInlineSerializer
 
+    annotation_serializer: BaseAnnotationSerializer
+
     params: CommonParams = CommonParams()
 
     _excluded_refs_cache: dict[str, set[str]] = {}
@@ -449,6 +456,10 @@ def post_process(
                 res = self.serialize_underline(text=res)
             if formatting.strikethrough:
                 res = self.serialize_strikethrough(text=res)
+            if formatting.script == Script.SUB:
+                res = self.serialize_subscript(text=res)
+            elif formatting.script == Script.SUPER:
+                res = self.serialize_superscript(text=res)
         if params.include_hyperlinks and hyperlink:
             res = self.serialize_hyperlink(text=res, hyperlink=hyperlink)
         return res
@@ -473,6 +484,16 @@ def serialize_strikethrough(self, text: str, **kwargs: Any) -> str:
         """Hook for strikethrough formatting serialization."""
         return text
 
+    @override
+    def serialize_subscript(self, text: str, **kwargs: Any) -> str:
+        """Hook for subscript formatting serialization."""
+        return text
+
+    @override
+    def serialize_superscript(self, text: str, **kwargs: Any) -> str:
+        """Hook for superscript formatting serialization."""
+        return text
+
     @override
     def serialize_hyperlink(
         self,
@@ -505,6 +526,19 @@ def serialize_captions(
             text_res = ""
         return create_ser_result(text=text_res, span_source=results)
 
+    @override
+    def serialize_annotations(
+        self,
+        item: DocItem,
+        **kwargs: Any,
+    ) -> SerializationResult:
+        """Serialize the item's annotations."""
+        return self.annotation_serializer.serialize(
+            item=item,
+            doc=self.doc,
+            **kwargs,
+        )
+
     def _get_applicable_pages(self) -> Optional[list[int]]:
         pages = {
             item.prov[0].page_no: ...

docling_core/transforms/serializer/doctags.py

@@ -7,6 +7,7 @@
 from typing_extensions import override
 
 from docling_core.transforms.serializer.base import (
+    BaseAnnotationSerializer,
     BaseDocSerializer,
     BaseFallbackSerializer,
     BaseFormSerializer,
@@ -17,12 +18,14 @@
     BaseTableSerializer,
     BaseTextSerializer,
     SerializationResult,
+    Span,
 )
 from docling_core.transforms.serializer.common import (
     CommonParams,
     DocSerializer,
     create_ser_result,
 )
+from docling_core.types.doc.base import BoundingBox
 from docling_core.types.doc.document import (
     CodeItem,
     DocItem,
@@ -38,6 +41,7 @@
     PictureItem,
     PictureMoleculeData,
     PictureTabularChartData,
+    ProvenanceItem,
     TableItem,
     TextItem,
     UnorderedList,
@@ -414,6 +418,39 @@ def serialize(
 class DocTagsInlineSerializer(BaseInlineSerializer):
     """DocTags-specific inline group serializer."""
 
+    def _get_inline_location_tags(
+        self, doc: DoclingDocument, item: InlineGroup, params: DocTagsParams
+    ) -> SerializationResult:
+
+        prov: Optional[ProvenanceItem] = None
+        boxes: list[BoundingBox] = []
+        doc_items: list[DocItem] = []
+        for it, _ in doc.iterate_items(root=item):
+            if isinstance(it, DocItem):
+                for prov in it.prov:
+                    boxes.append(prov.bbox)
+                    doc_items.append(it)
+        if prov is None:
+            return create_ser_result()
+
+        bbox = BoundingBox.enclosing_bbox(boxes=boxes)
+
+        # using last seen prov as reference for page dims
+        page_w, page_h = doc.pages[prov.page_no].size.as_tuple()
+
+        loc_str = DocumentToken.get_location(
+            bbox=bbox.to_top_left_origin(page_h).as_tuple(),
+            page_w=page_w,
+            page_h=page_h,
+            xsize=params.xsize,
+            ysize=params.ysize,
+        )
+
+        return SerializationResult(
+            text=loc_str,
+            spans=[Span(item=it) for it in doc_items],
+        )
+
     @override
     def serialize(
         self,
@@ -428,12 +465,23 @@ def serialize(
         """Serializes the passed item."""
         my_visited = visited if visited is not None else set()
         params = DocTagsParams(**kwargs)
-        parts = doc_serializer.get_parts(
-            item=item,
-            list_level=list_level,
-            is_inline_scope=True,
-            visited=my_visited,
-            **kwargs,
+        parts: List[SerializationResult] = []
+        if params.add_location:
+            inline_loc_tags_ser_res = self._get_inline_location_tags(
+                doc=doc,
+                item=item,
+                params=params,
+            )
+            parts.append(inline_loc_tags_ser_res)
+            params.add_location = False  # suppress children location serialization
+        parts.extend(
+            doc_serializer.get_parts(
+                item=item,
+                list_level=list_level,
+                is_inline_scope=True,
+                visited=my_visited,
+                **{**kwargs, **params.model_dump()},
+            )
         )
         wrap_tag = DocumentToken.INLINE.value
         delim = _get_delim(params=params)
@@ -460,6 +508,15 @@ def serialize(
         return create_ser_result()
 
 
+class DocTagsAnnotationSerializer(BaseAnnotationSerializer):
+    """DocTags-specific annotation serializer."""
+
+    @override
+    def serialize(self, *, item: DocItem, **kwargs: Any) -> SerializationResult:
+        """Serializes the item's annotations."""
+        return create_ser_result()
+
+
 class DocTagsDocSerializer(DocSerializer):
     """DocTags-specific document serializer."""
 
@@ -473,6 +530,8 @@ class DocTagsDocSerializer(DocSerializer):
     list_serializer: BaseListSerializer = DocTagsListSerializer()
     inline_serializer: BaseInlineSerializer = DocTagsInlineSerializer()
 
+    annotation_serializer: BaseAnnotationSerializer = DocTagsAnnotationSerializer()
+
     params: DocTagsParams = DocTagsParams()
 
     @override