docling-core 2.47.0__tar.gz → 2.48.0__tar.gz

{docling_core-2.47.0 → docling_core-2.48.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-core
-Version: 2.47.0
+Version: 2.48.0
 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-2.47.0 → docling_core-2.48.0}/docling_core/transforms/chunker/hierarchical_chunker.py

@@ -145,7 +145,7 @@ class TripletTableSerializer(BaseTableSerializer):
             parts.append(cap_res)
 
         if item.self_ref not in doc_serializer.get_excluded_refs(**kwargs):
-            table_df = item.export_to_dataframe()
+            table_df = item.export_to_dataframe(doc)
             if table_df.shape[0] >= 1 and table_df.shape[1] >= 2:
 
                 # copy header as first row and shift all rows by one

{docling_core-2.47.0 → docling_core-2.48.0}/docling_core/transforms/serializer/common.py

@@ -394,6 +394,7 @@ class DocSerializer(BaseModel, BaseDocSerializer):
                 item=item,
                 doc_serializer=self,
                 doc=self.doc,
+                visited=my_visited,
                 **my_kwargs,
             )
         return part

{docling_core-2.47.0 → docling_core-2.48.0}/docling_core/transforms/serializer/doctags.py

@@ -32,6 +32,7 @@ from docling_core.types.doc.document import (
     DoclingDocument,
     FloatingItem,
     FormItem,
+    GroupItem,
     InlineGroup,
     KeyValueItem,
     ListGroup,
@@ -42,6 +43,7 @@ from docling_core.types.doc.document import (
     PictureMoleculeData,
     PictureTabularChartData,
     ProvenanceItem,
+    SectionHeaderItem,
     TableItem,
     TextItem,
 )
@@ -94,11 +96,11 @@ class DocTagsTextSerializer(BaseModel, BaseTextSerializer):
         item: TextItem,
         doc_serializer: BaseDocSerializer,
         doc: DoclingDocument,
+        visited: Optional[set[str]] = None,
         **kwargs: Any,
     ) -> SerializationResult:
         """Serializes the passed item."""
-        from docling_core.types.doc.document import SectionHeaderItem
-
+        my_visited = visited if visited is not None else set()
         params = DocTagsParams(**kwargs)
         wrap_tag: Optional[str] = DocumentToken.create_token_name_from_doc_item_label(
             label=item.label,
@@ -116,12 +118,21 @@ class DocTagsTextSerializer(BaseModel, BaseTextSerializer):
                 parts.append(location)
 
         if params.add_content:
-            text_part = item.text
-            text_part = doc_serializer.post_process(
-                text=text_part,
-                formatting=item.formatting,
-                hyperlink=item.hyperlink,
-            )
+            if (
+                item.text == ""
+                and len(item.children) == 1
+                and isinstance(
+                    (child_group := item.children[0].resolve(doc)), InlineGroup
+                )
+            ):
+                ser_res = doc_serializer.serialize(item=child_group, visited=my_visited)
+                text_part = ser_res.text
+            else:
+                text_part = doc_serializer.post_process(
+                    text=item.text,
+                    formatting=item.formatting,
+                    hyperlink=item.hyperlink,
+                )
 
             if isinstance(item, CodeItem):
                 language_token = DocumentToken.get_code_language_token(
@@ -506,7 +517,12 @@ class DocTagsFallbackSerializer(BaseFallbackSerializer):
         **kwargs: Any,
     ) -> SerializationResult:
         """Serializes the passed item."""
-        return create_ser_result()
+        if isinstance(item, GroupItem):
+            parts = doc_serializer.get_parts(item=item, **kwargs)
+            text_res = "\n".join([p.text for p in parts if p.text])
+            return create_ser_result(text=text_res, span_source=parts)
+        else:
+            return create_ser_result()
 
 
 class DocTagsAnnotationSerializer(BaseAnnotationSerializer):

{docling_core-2.47.0 → docling_core-2.48.0}/docling_core/transforms/serializer/html.py

@@ -55,6 +55,7 @@ from docling_core.types.doc.document import (
     FormItem,
     FormulaItem,
     GraphData,
+    GroupItem,
     ImageRef,
     InlineGroup,
     KeyValueItem,
@@ -139,21 +140,34 @@ class HTMLTextSerializer(BaseModel, BaseTextSerializer):
         res_parts: list[SerializationResult] = []
         post_processed = False
 
-        # Prepare the HTML based on item type
-        if isinstance(item, TitleItem):
-            text_inner = self._prepare_content(item.text)
-            text = get_html_tag_with_text_direction(html_tag="h1", text=text_inner)
+        has_inline_repr = (
+            item.text == ""
+            and len(item.children) == 1
+            and isinstance((child_group := item.children[0].resolve(doc)), InlineGroup)
+        )
+        if has_inline_repr:
+            text = doc_serializer.serialize(item=child_group, visited=my_visited).text
+            post_processed = True
+        else:
+            text = item.text
+            if not isinstance(item, (CodeItem, FormulaItem)):
+                text = html.escape(text, quote=False)
+                text = text.replace("\n", "<br>")
 
-        elif isinstance(item, SectionHeaderItem):
-            section_level = min(item.level + 1, 6)
-            text_inner = self._prepare_content(item.text)
+        # Prepare the HTML based on item type
+        if isinstance(item, (TitleItem, SectionHeaderItem)):
+            section_level = (
+                min(item.level + 1, 6) if isinstance(item, SectionHeaderItem) else 1
+            )
             text = get_html_tag_with_text_direction(
-                html_tag=f"h{section_level}", text=text_inner
+                html_tag=f"h{section_level}", text=text
             )
 
         elif isinstance(item, FormulaItem):
             text = self._process_formula(
                 item=item,
+                text=text,
+                orig=item.orig,
                 doc=doc,
                 image_mode=params.image_mode,
                 formula_to_mathml=params.formula_to_mathml,
@@ -161,19 +175,26 @@ class HTMLTextSerializer(BaseModel, BaseTextSerializer):
             )
 
         elif isinstance(item, CodeItem):
-            text = self._process_code(item=item, is_inline_scope=is_inline_scope)
+            text = (
+                f"<code>{text}</code>"
+                if is_inline_scope
+                else f"<pre><code>{text}</code></pre>"
+            )
 
         elif isinstance(item, ListItem):
             # List items are handled by list serializer
             text_parts: list[str] = []
-            if item_text := self._prepare_content(item.text):
-                item_text = doc_serializer.post_process(
-                    text=item_text,
-                    formatting=item.formatting,
-                    hyperlink=item.hyperlink,
-                )
-                post_processed = True
-                text_parts.append(item_text)
+            if text:
+                if has_inline_repr:
+                    text = f"\n{text}\n"
+                else:
+                    text = doc_serializer.post_process(
+                        text=text,
+                        formatting=item.formatting,
+                        hyperlink=item.hyperlink,
+                    )
+                    post_processed = True
+                text_parts.append(text)
             nested_parts = [
                 r.text
                 for r in doc_serializer.get_parts(
@@ -184,29 +205,26 @@ class HTMLTextSerializer(BaseModel, BaseTextSerializer):
                 )
             ]
             text_parts.extend(nested_parts)
-            text_inner = "\n".join(text_parts)
+            text = "\n".join(text_parts)
             if nested_parts:
-                text_inner = f"\n{text_inner}\n"
+                text = f"\n{text}\n"
             text = (
                 get_html_tag_with_text_direction(
                     html_tag="li",
-                    text=text_inner,
+                    text=text,
                     attrs=(
                         {"style": f"list-style-type: '{item.marker} ';"}
                         if params.show_original_list_item_marker and item.marker
                         else {}
                     ),
                 )
-                if text_inner
+                if text
                 else ""
             )
 
-        elif is_inline_scope:
-            text = self._prepare_content(item.text)
-        else:
+        elif not is_inline_scope:
             # Regular text item
-            text_inner = self._prepare_content(item.text)
-            text = get_html_tag_with_text_direction(html_tag="p", text=text_inner)
+            text = get_html_tag_with_text_direction(html_tag="p", text=text)
 
         # Apply formatting and hyperlinks
         if not post_processed:
@@ -227,66 +245,44 @@ class HTMLTextSerializer(BaseModel, BaseTextSerializer):
 
         return create_ser_result(text=text, span_source=res_parts)
 
-    def _prepare_content(
-        self, text: str, do_escape_html=True, do_replace_newline=True
-    ) -> str:
-        """Prepare text content for HTML inclusion."""
-        if do_escape_html:
-            text = html.escape(text, quote=False)
-        if do_replace_newline:
-            text = text.replace("\n", "<br>")
-        return text
-
-    def _process_code(
-        self,
-        item: CodeItem,
-        is_inline_scope: bool,
-    ) -> str:
-        code_text = self._prepare_content(
-            item.text, do_escape_html=False, do_replace_newline=False
-        )
-        if is_inline_scope:
-            text = f"<code>{code_text}</code>"
-        else:
-            text = f"<pre><code>{code_text}</code></pre>"
-
-        return text
-
     def _process_formula(
         self,
-        item: FormulaItem,
+        *,
+        item: DocItem,
+        text: str,
+        orig: str,
         doc: DoclingDocument,
         image_mode: ImageRefMode,
         formula_to_mathml: bool,
         is_inline_scope: bool,
     ) -> str:
         """Process a formula item to HTML/MathML."""
-        math_formula = self._prepare_content(
-            item.text, do_escape_html=False, do_replace_newline=False
-        )
-
         # If formula is empty, try to use an image fallback
-        if item.text == "" and item.orig != "":
-            img_fallback = self._get_formula_image_fallback(item, doc)
-            if (
-                image_mode == ImageRefMode.EMBEDDED
-                and len(item.prov) > 0
-                and img_fallback
-            ):
-                return img_fallback
+        if (
+            text == ""
+            and orig != ""
+            and len(item.prov) > 0
+            and image_mode == ImageRefMode.EMBEDDED
+            and (
+                img_fallback := self._get_formula_image_fallback(
+                    item=item, orig=orig, doc=doc
+                )
+            )
+        ):
+            return img_fallback
 
         # Try to generate MathML
-        if formula_to_mathml and math_formula:
+        elif formula_to_mathml and text:
             try:
                 # Set display mode based on context
                 display_mode = "inline" if is_inline_scope else "block"
                 mathml_element = latex2mathml.converter.convert_to_element(
-                    math_formula, display=display_mode
+                    text, display=display_mode
                 )
                 annotation = SubElement(
                     mathml_element, "annotation", dict(encoding="TeX")
                 )
-                annotation.text = math_formula
+                annotation.text = text
                 mathml = unescape(tostring(mathml_element, encoding="unicode"))
 
                 # Don't wrap in div for inline formulas
@@ -296,40 +292,40 @@ class HTMLTextSerializer(BaseModel, BaseTextSerializer):
                     return f"<div>{mathml}</div>"
 
             except Exception:
-                img_fallback = self._get_formula_image_fallback(item, doc)
+                img_fallback = self._get_formula_image_fallback(
+                    item=item, orig=orig, doc=doc
+                )
                 if (
                     image_mode == ImageRefMode.EMBEDDED
                     and len(item.prov) > 0
                     and img_fallback
                 ):
                     return img_fallback
-                elif math_formula:
-                    return f"<pre>{math_formula}</pre>"
+                elif text:
+                    return f"<pre>{text}</pre>"
                 else:
                     return "<pre>Formula not decoded</pre>"
 
         _logger.warning("Could not parse formula with MathML")
 
         # Fallback options if we got here
-        if math_formula and is_inline_scope:
-            return f"<code>{math_formula}</code>"
-        elif math_formula and (not is_inline_scope):
-            f"<pre>{math_formula}</pre>"
+        if text and is_inline_scope:
+            return f"<code>{text}</code>"
+        elif text and (not is_inline_scope):
+            f"<pre>{text}</pre>"
         elif is_inline_scope:
             return '<span class="formula-not-decoded">Formula not decoded</span>'
 
         return '<div class="formula-not-decoded">Formula not decoded</div>'
 
     def _get_formula_image_fallback(
-        self, item: TextItem, doc: DoclingDocument
+        self, *, item: DocItem, orig: str, doc: DoclingDocument
     ) -> Optional[str]:
         """Try to get an image fallback for a formula."""
         item_image = item.get_image(doc=doc)
         if item_image is not None:
             img_ref = ImageRef.from_pil(item_image, dpi=72)
-            return (
-                "<figure>" f'<img src="{img_ref.uri}" alt="{item.orig}" />' "</figure>"
-            )
+            return "<figure>" f'<img src="{img_ref.uri}" alt="{orig}" />' "</figure>"
         return None
 
 
@@ -792,21 +788,30 @@ class HTMLFallbackSerializer(BaseFallbackSerializer):
     """HTML-specific fallback serializer."""
 
     @override
-    def serialize(self, *, item: NodeItem, **kwargs: Any) -> SerializationResult:
+    def serialize(
+        self,
+        *,
+        item: NodeItem,
+        doc_serializer: "BaseDocSerializer",
+        doc: DoclingDocument,
+        **kwargs: Any,
+    ) -> SerializationResult:
         """Fallback serializer for items not handled by other serializers."""
-        if isinstance(item, DocItem):
+        if isinstance(item, GroupItem):
+            parts = doc_serializer.get_parts(item=item, **kwargs)
+            text_res = "\n".join([p.text for p in parts if p.text])
+            return create_ser_result(text=text_res, span_source=parts)
+        else:
             return create_ser_result(
                 text=f"<!-- Unhandled item type: {item.__class__.__name__} -->",
-                span_source=item,
+                span_source=item if isinstance(item, DocItem) else [],
             )
-        else:
-            # For group items, we don't generate any markup
-            return create_ser_result()
 
 
 class HTMLAnnotationSerializer(BaseModel, BaseAnnotationSerializer):
     """HTML-specific annotation serializer."""
 
+    @override
     def serialize(
         self,
         *,

{docling_core-2.47.0 → docling_core-2.48.0}/docling_core/transforms/serializer/markdown.py

@@ -45,6 +45,7 @@ from docling_core.types.doc.document import (
     Formatting,
     FormItem,
     FormulaItem,
+    GroupItem,
     ImageRef,
     InlineGroup,
     KeyValueItem,
@@ -124,26 +125,24 @@ class MarkdownTextSerializer(BaseModel, BaseTextSerializer):
         my_visited = visited if visited is not None else set()
         params = MarkdownParams(**kwargs)
         res_parts: list[SerializationResult] = []
-        text = item.text
         escape_html = True
         escape_underscores = True
-        processing_pending = True
-        if isinstance(item, (ListItem, TitleItem, SectionHeaderItem)):
-            # case where processing/formatting should be applied first (in inner scope)
+
+        has_inline_repr = (
+            item.text == ""
+            and len(item.children) == 1
+            and isinstance((child_group := item.children[0].resolve(doc)), InlineGroup)
+        )
+        if has_inline_repr:
+            text = doc_serializer.serialize(item=child_group, visited=my_visited).text
             processing_pending = False
-            if (
-                text == ""
-                and len(item.children) == 1
-                and isinstance(
-                    (child_group := item.children[0].resolve(doc)), InlineGroup
-                )
-            ):
-                # case of inline within heading / list item
-                ser_res = doc_serializer.serialize(item=child_group)
-                text = ser_res.text
-                for span in ser_res.spans:
-                    my_visited.add(span.item.self_ref)
-            else:
+        else:
+            text = item.text
+            processing_pending = True
+
+        if isinstance(item, (ListItem, TitleItem, SectionHeaderItem)):
+            if not has_inline_repr:
+                # case where processing/formatting should be applied first (in inner scope)
                 text = doc_serializer.post_process(
                     text=text,
                     escape_html=escape_html,
@@ -151,6 +150,7 @@ class MarkdownTextSerializer(BaseModel, BaseTextSerializer):
                     formatting=item.formatting,
                     hyperlink=item.hyperlink,
                 )
+                processing_pending = False
 
             if isinstance(item, ListItem):
                 pieces: list[str] = []
@@ -600,13 +600,15 @@ class MarkdownFallbackSerializer(BaseFallbackSerializer):
         **kwargs: Any,
     ) -> SerializationResult:
         """Serializes the passed item."""
-        if isinstance(item, DocItem):
+        if isinstance(item, GroupItem):
+            parts = doc_serializer.get_parts(item=item, **kwargs)
+            text_res = "\n\n".join([p.text for p in parts if p.text])
+            return create_ser_result(text=text_res, span_source=parts)
+        else:
             return create_ser_result(
                 text="<!-- missing-text -->",
-                span_source=item,
+                span_source=item if isinstance(item, DocItem) else [],
             )
-        else:
-            return create_ser_result()
 
 
 class MarkdownDocSerializer(DocSerializer):

{docling_core-2.47.0 → docling_core-2.48.0}/docling_core/types/doc/document.py

@@ -60,7 +60,7 @@ _logger = logging.getLogger(__name__)
 
 Uint64 = typing.Annotated[int, Field(ge=0, le=(2**64 - 1))]
 LevelNumber = typing.Annotated[int, Field(ge=1, le=100)]
-CURRENT_VERSION: Final = "1.6.0"
+CURRENT_VERSION: Final = "1.7.0"
 
 DEFAULT_EXPORT_LABELS = {
     DocItemLabel.TITLE,
@@ -310,6 +310,7 @@ class TableCell(BaseModel):
     column_header: bool = False
     row_header: bool = False
     row_section: bool = False
+    fillable: bool = False
 
     @model_validator(mode="before")
     @classmethod

{docling_core-2.47.0 → docling_core-2.48.0}/docling_core.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-core
-Version: 2.47.0
+Version: 2.48.0
 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-2.47.0 → docling_core-2.48.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "docling-core"
-version = "2.47.0"  # DO NOT EDIT, updated automatically
+version = "2.48.0"  # DO NOT EDIT, updated automatically
 description = "A python library to define and validate data types in Docling."
 license = "MIT"
 license-files = ["LICENSE"]

{docling_core-2.47.0 → docling_core-2.48.0}/test/test_docling_doc.py

@@ -734,7 +734,7 @@ def _test_export_methods(
     for table in doc.tables:
         table.export_to_markdown()
         table.export_to_html(doc)
-        table.export_to_dataframe()
+        table.export_to_dataframe(doc)
         table.export_to_doctags(doc)
 
     # Test Images export ...
@@ -2102,7 +2102,7 @@ def _construct_rich_table_doc():
 
     table_item = doc.add_table(
         data=TableData(
-            num_rows=4,
+            num_rows=5,
             num_cols=2,
         ),
     )
@@ -2121,6 +2121,17 @@ def _construct_rich_table_doc():
     rich_item_3 = doc.add_table(
         data=TableData(num_rows=2, num_cols=3), parent=table_item
     )
+
+    rich_item_4 = doc.add_group(parent=table_item, label=GroupLabel.UNSPECIFIED)
+    doc.add_text(
+        parent=rich_item_4,
+        text="Some text in a generic group.",
+        label=DocItemLabel.TEXT,
+    )
+    doc.add_text(
+        parent=rich_item_4, text="More text in the group.", label=DocItemLabel.TEXT
+    )
+
     for i in range(rich_item_3.data.num_rows):
         for j in range(rich_item_3.data.num_cols):
             cell = TableCell(
@@ -2158,6 +2169,14 @@ def _construct_rich_table_doc():
                     end_col_offset_idx=j + 1,
                     ref=rich_item_3.get_ref(),
                 )
+            elif i == 4 and j == 0:
+                cell = RichTableCell(
+                    start_row_offset_idx=i,
+                    end_row_offset_idx=i + 1,
+                    start_col_offset_idx=j,
+                    end_col_offset_idx=j + 1,
+                    ref=rich_item_4.get_ref(),
+                )
             else:
                 cell = TableCell(
                     start_row_offset_idx=i,

{docling_core-2.47.0 → docling_core-2.48.0}/test/test_serialization.py

@@ -85,6 +85,11 @@ def verify(exp_file: Path, actual: str):
         assert expected == actual
 
 
+# ===============================
+# Markdown tests
+# ===============================
+
+
 def test_md_cross_page_list_page_break():
     src = Path("./test/data/doc/activities.json")
     doc = DoclingDocument.load_from_json(src)
@@ -99,7 +104,7 @@ def test_md_cross_page_list_page_break():
         ),
     )
     actual = ser.serialize().text
-    verify(exp_file=src.parent / f"{src.stem}.gt.md", actual=actual)
+    verify(exp_file=src.with_suffix(".gt.md"), actual=actual)
 
 
 def test_md_cross_page_list_page_break_none():
@@ -170,20 +175,6 @@ def test_md_cross_page_list_page_break_p2():
     verify(exp_file=src.parent / f"{src.stem}_p2.gt.md", actual=actual)
 
 
-def test_html_charts():
-    src = Path("./test/data/doc/barchart.json")
-    doc = DoclingDocument.load_from_json(src)
-
-    ser = HTMLDocSerializer(
-        doc=doc,
-        params=HTMLParams(
-            image_mode=ImageRefMode.PLACEHOLDER,
-        ),
-    )
-    actual = ser.serialize().text
-    verify(exp_file=src.parent / f"{src.stem}.gt.html", actual=actual)
-
-
 def test_md_charts():
     src = Path("./test/data/doc/barchart.json")
     doc = DoclingDocument.load_from_json(src)
@@ -195,7 +186,7 @@ def test_md_charts():
         ),
     )
     actual = ser.serialize().text
-    verify(exp_file=src.parent / f"{src.stem}.gt.md", actual=actual)
+    verify(exp_file=src.with_suffix(".gt.md"), actual=actual)
 
 
 def test_md_inline_and_formatting():
@@ -209,51 +200,7 @@ def test_md_inline_and_formatting():
         ),
     )
     actual = ser.serialize().text
-    verify(exp_file=src.parent / f"{src.stem}.md", actual=actual)
-
-
-def test_html_cross_page_list_page_break():
-    src = Path("./test/data/doc/activities.json")
-    doc = DoclingDocument.load_from_json(src)
-
-    ser = HTMLDocSerializer(
-        doc=doc,
-        params=HTMLParams(
-            image_mode=ImageRefMode.PLACEHOLDER,
-        ),
-    )
-    actual = ser.serialize().text
-    verify(exp_file=src.parent / f"{src.stem}.gt.html", actual=actual)
-
-
-def test_html_cross_page_list_page_break_p1():
-    src = Path("./test/data/doc/activities.json")
-    doc = DoclingDocument.load_from_json(src)
-
-    ser = HTMLDocSerializer(
-        doc=doc,
-        params=HTMLParams(
-            image_mode=ImageRefMode.PLACEHOLDER,
-            pages={1},
-        ),
-    )
-    actual = ser.serialize().text
-    verify(exp_file=src.parent / f"{src.stem}_p1.gt.html", actual=actual)
-
-
-def test_html_cross_page_list_page_break_p2():
-    src = Path("./test/data/doc/activities.json")
-    doc = DoclingDocument.load_from_json(src)
-
-    ser = HTMLDocSerializer(
-        doc=doc,
-        params=HTMLParams(
-            image_mode=ImageRefMode.PLACEHOLDER,
-            pages={2},
-        ),
-    )
-    actual = ser.serialize().text
-    verify(exp_file=src.parent / f"{src.stem}_p2.gt.html", actual=actual)
+    verify(exp_file=src.with_suffix(".gt.md"), actual=actual)
 
 
 def test_md_pb_placeholder_and_page_filter():
@@ -269,7 +216,7 @@ def test_md_pb_placeholder_and_page_filter():
         ),
     )
     actual = ser.serialize().text
-    verify(exp_file=src.parent / f"{src.stem}.gt.md", actual=actual)
+    verify(exp_file=src.with_suffix(".gt.md"), actual=actual)
 
 
 def test_md_list_item_markers():
@@ -358,7 +305,7 @@ def test_md_nested_lists():
 
     ser = MarkdownDocSerializer(doc=doc)
     actual = ser.serialize().text
-    verify(exp_file=src.parent / f"{src.stem}.gt.md", actual=actual)
+    verify(exp_file=src.with_suffix(".gt.md"), actual=actual)
 
 
 def test_md_rich_table():
@@ -370,6 +317,69 @@ def test_md_rich_table():
     verify(exp_file=exp_file, actual=actual)
 
 
+# ===============================
+# HTML tests
+# ===============================
+
+
+def test_html_charts():
+    src = Path("./test/data/doc/barchart.json")
+    doc = DoclingDocument.load_from_json(src)
+
+    ser = HTMLDocSerializer(
+        doc=doc,
+        params=HTMLParams(
+            image_mode=ImageRefMode.PLACEHOLDER,
+        ),
+    )
+    actual = ser.serialize().text
+    verify(exp_file=src.with_suffix(".gt.html"), actual=actual)
+
+
+def test_html_cross_page_list_page_break():
+    src = Path("./test/data/doc/activities.json")
+    doc = DoclingDocument.load_from_json(src)
+
+    ser = HTMLDocSerializer(
+        doc=doc,
+        params=HTMLParams(
+            image_mode=ImageRefMode.PLACEHOLDER,
+        ),
+    )
+    actual = ser.serialize().text
+    verify(exp_file=src.with_suffix(".gt.html"), actual=actual)
+
+
+def test_html_cross_page_list_page_break_p1():
+    src = Path("./test/data/doc/activities.json")
+    doc = DoclingDocument.load_from_json(src)
+
+    ser = HTMLDocSerializer(
+        doc=doc,
+        params=HTMLParams(
+            image_mode=ImageRefMode.PLACEHOLDER,
+            pages={1},
+        ),
+    )
+    actual = ser.serialize().text
+    verify(exp_file=src.parent / f"{src.stem}_p1.gt.html", actual=actual)
+
+
+def test_html_cross_page_list_page_break_p2():
+    src = Path("./test/data/doc/activities.json")
+    doc = DoclingDocument.load_from_json(src)
+
+    ser = HTMLDocSerializer(
+        doc=doc,
+        params=HTMLParams(
+            image_mode=ImageRefMode.PLACEHOLDER,
+            pages={2},
+        ),
+    )
+    actual = ser.serialize().text
+    verify(exp_file=src.parent / f"{src.stem}_p2.gt.html", actual=actual)
+
+
 def test_html_split_page():
     src = Path("./test/data/doc/2408.09869v3_enriched.json")
     doc = DoclingDocument.load_from_json(src)
@@ -506,7 +516,7 @@ def test_html_nested_lists():
 
     ser = HTMLDocSerializer(doc=doc)
     actual = ser.serialize().text
-    verify(exp_file=src.parent / f"{src.stem}.gt.html", actual=actual)
+    verify(exp_file=src.with_suffix(".gt.html"), actual=actual)
 
 
 def test_html_rich_table():
@@ -518,13 +528,27 @@ def test_html_rich_table():
     verify(exp_file=exp_file, actual=actual)
 
 
+def test_html_inline_and_formatting():
+    src = Path("./test/data/doc/inline_and_formatting.yaml")
+    doc = DoclingDocument.load_from_yaml(src)
+
+    ser = HTMLDocSerializer(doc=doc)
+    actual = ser.serialize().text
+    verify(exp_file=src.with_suffix(".gt.html"), actual=actual)
+
+
+# ===============================
+# DocTags tests
+# ===============================
+
+
 def test_doctags_inline_loc_tags():
     src = Path("./test/data/doc/2408.09869v3_enriched.json")
     doc = DoclingDocument.load_from_json(src)
 
     ser = DocTagsDocSerializer(doc=doc)
     actual = ser.serialize().text
-    verify(exp_file=src.parent / f"{src.stem}.out.dt", actual=actual)
+    verify(exp_file=src.with_suffix(".out.dt"), actual=actual)
 
 
 def test_doctags_rich_table():
@@ -535,3 +559,12 @@ def test_doctags_rich_table():
     ser = DocTagsDocSerializer(doc=doc)
     actual = ser.serialize().text
     verify(exp_file=exp_file, actual=actual)
+
+
+def test_doctags_inline_and_formatting():
+    src = Path("./test/data/doc/inline_and_formatting.yaml")
+    doc = DoclingDocument.load_from_yaml(src)
+
+    ser = DocTagsDocSerializer(doc=doc)
+    actual = ser.serialize().text
+    verify(exp_file=src.with_suffix(".gt.dt"), actual=actual)