docling-core 2.19.0__tar.gz → 2.20.0__tar.gz

{docling_core-2.19.0 → docling_core-2.20.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: docling-core
-Version: 2.19.0
+Version: 2.20.0
 Summary: A python library to define and validate data types in Docling.
 Home-page: https://ds4sd.github.io/
 License: MIT

{docling_core-2.19.0 → docling_core-2.20.0}/docling_core/types/doc/base.py

@@ -1,7 +1,7 @@
 """Models for the base data types."""
 
 from enum import Enum
-from typing import Tuple
+from typing import List, Tuple
 
 from pydantic import BaseModel
 
@@ -365,3 +365,30 @@ class BoundingBox(BaseModel):
             raise ValueError("BoundingBoxes have different CoordOrigin")
 
         return False
+
+    @classmethod
+    def enclosing_bbox(cls, boxes: List["BoundingBox"]) -> "BoundingBox":
+        """Create a bounding box that covers all of the given boxes."""
+        if not boxes:
+            raise ValueError("No bounding boxes provided for union.")
+
+        origin = boxes[0].coord_origin
+        if any(box.coord_origin != origin for box in boxes):
+            raise ValueError(
+                "All bounding boxes must have the same \
+                CoordOrigin to compute their union."
+            )
+
+        left = min(box.l for box in boxes)
+        right = max(box.r for box in boxes)
+
+        if origin == CoordOrigin.TOPLEFT:
+            top = min(box.t for box in boxes)
+            bottom = max(box.b for box in boxes)
+        elif origin == CoordOrigin.BOTTOMLEFT:
+            top = max(box.t for box in boxes)
+            bottom = min(box.b for box in boxes)
+        else:
+            raise ValueError("BoundingBoxes have different CoordOrigin")
+
+        return cls(l=left, t=top, r=right, b=bottom, coord_origin=origin)

{docling_core-2.19.0 → docling_core-2.20.0}/docling_core/types/doc/document.py

@@ -43,7 +43,13 @@ from docling_core.search.package import VERSION_PATTERN
 from docling_core.types.base import _JSON_POINTER_REGEX
 from docling_core.types.doc import BoundingBox, Size
 from docling_core.types.doc.base import ImageRefMode
-from docling_core.types.doc.labels import CodeLanguageLabel, DocItemLabel, GroupLabel
+from docling_core.types.doc.labels import (
+    CodeLanguageLabel,
+    DocItemLabel,
+    GraphCellLabel,
+    GraphLinkLabel,
+    GroupLabel,
+)
 from docling_core.types.doc.tokens import DocumentToken, TableToken
 from docling_core.types.doc.utils import (
     get_html_tag_with_text_direction,
@@ -1101,7 +1107,9 @@ class TableItem(FloatingItem):
         return md_table
 
     def export_to_html(
-        self, doc: Optional["DoclingDocument"] = None, add_caption: bool = True
+        self,
+        doc: Optional["DoclingDocument"] = None,
+        add_caption: bool = True,
     ) -> str:
         """Export the table as html."""
         if doc is None:
@@ -1330,11 +1338,73 @@ class TableItem(FloatingItem):
         return body
 
 
-class KeyValueItem(DocItem):
+class GraphCell(BaseModel):
+    """GraphCell."""
+
+    label: GraphCellLabel
+
+    cell_id: int
+
+    text: str  # sanitized text
+    orig: str  # text as seen on document
+
+    prov: Optional[ProvenanceItem] = None
+
+    # in case you have a text, table or picture item
+    item_ref: Optional[RefItem] = None
+
+
+class GraphLink(BaseModel):
+    """GraphLink."""
+
+    label: GraphLinkLabel
+
+    source_cell_id: int
+    target_cell_id: int
+
+
+class GraphData(BaseModel):
+    """GraphData."""
+
+    cells: List[GraphCell] = Field(default_factory=list)
+    links: List[GraphLink] = Field(default_factory=list)
+
+    @field_validator("links")
+    @classmethod
+    def validate_links(cls, links, info):
+        """Ensure that each link is valid."""
+        cells = info.data.get("cells", [])
+
+        valid_cell_ids = {cell.cell_id for cell in cells}
+
+        for link in links:
+            if link.source_cell_id not in valid_cell_ids:
+                raise ValueError(
+                    f"Invalid source_cell_id {link.source_cell_id} in GraphLink"
+                )
+            if link.target_cell_id not in valid_cell_ids:
+                raise ValueError(
+                    f"Invalid target_cell_id {link.target_cell_id} in GraphLink"
+                )
+
+        return links
+
+
+class KeyValueItem(FloatingItem):
     """KeyValueItem."""
 
     label: typing.Literal[DocItemLabel.KEY_VALUE_REGION] = DocItemLabel.KEY_VALUE_REGION
 
+    graph: GraphData
+
+
+class FormItem(FloatingItem):
+    """FormItem."""
+
+    label: typing.Literal[DocItemLabel.FORM] = DocItemLabel.FORM
+
+    graph: GraphData
+
 
 ContentItem = Annotated[
     Union[
@@ -1446,7 +1516,9 @@ class DoclingDocument(BaseModel):
     )
 
     furniture: Annotated[GroupItem, Field(deprecated=True)] = GroupItem(
-        name="_root_", self_ref="#/furniture", content_layer=ContentLayer.FURNITURE
+        name="_root_",
+        self_ref="#/furniture",
+        content_layer=ContentLayer.FURNITURE,
     )  # List[RefItem] = []
     body: GroupItem = GroupItem(name="_root_", self_ref="#/body")  # List[RefItem] = []
 
@@ -1455,6 +1527,7 @@ class DoclingDocument(BaseModel):
     pictures: List[PictureItem] = []
     tables: List[TableItem] = []
     key_value_items: List[KeyValueItem] = []
+    form_items: List[FormItem] = []
 
     pages: Dict[int, PageItem] = {}  # empty as default
 
@@ -1851,6 +1924,68 @@ class DoclingDocument(BaseModel):
 
         return section_header_item
 
+    def add_key_values(
+        self,
+        graph: GraphData,
+        prov: Optional[ProvenanceItem] = None,
+        parent: Optional[NodeItem] = None,
+    ):
+        """add_key_values.
+
+        :param graph: GraphData:
+        :param prov: Optional[ProvenanceItem]:  (Default value = None)
+        :param parent: Optional[NodeItem]:  (Default value = None)
+        """
+        if not parent:
+            parent = self.body
+
+        key_value_index = len(self.key_value_items)
+        cref = f"#/key_value_items/{key_value_index}"
+
+        kv_item = KeyValueItem(
+            graph=graph,
+            self_ref=cref,
+            parent=parent.get_ref(),
+        )
+        if prov:
+            kv_item.prov.append(prov)
+
+        self.key_value_items.append(kv_item)
+        parent.children.append(RefItem(cref=cref))
+
+        return kv_item
+
+    def add_form(
+        self,
+        graph: GraphData,
+        prov: Optional[ProvenanceItem] = None,
+        parent: Optional[NodeItem] = None,
+    ):
+        """add_form.
+
+        :param graph: GraphData:
+        :param prov: Optional[ProvenanceItem]:  (Default value = None)
+        :param parent: Optional[NodeItem]:  (Default value = None)
+        """
+        if not parent:
+            parent = self.body
+
+        form_index = len(self.form_items)
+        cref = f"#/form_items/{form_index}"
+
+        form_item = FormItem(
+            graph=graph,
+            self_ref=cref,
+            parent=parent.get_ref(),
+        )
+        if prov:
+            form_item.prov.append(prov)
+
+        self.form_items.append(form_item)
+        parent.children.append(RefItem(cref=cref))
+
+        return form_item
+
     def num_pages(self):
         """num_pages."""
         return len(self.pages.values())
@@ -2009,7 +2144,8 @@ class DoclingDocument(BaseModel):
                             img.save(loc_path)
                             if reference_path is not None:
                                 obj_path = relative_path(
-                                    reference_path.resolve(), loc_path.resolve()
+                                    reference_path.resolve(),
+                                    loc_path.resolve(),
                                 )
                             else:
                                 obj_path = loc_path
@@ -2027,7 +2163,10 @@ class DoclingDocument(BaseModel):
         """Print_element_tree."""
         for ix, (item, level) in enumerate(self.iterate_items(with_groups=True)):
             if isinstance(item, GroupItem):
-                print(" " * level, f"{ix}: {item.label.value} with name={item.name}")
+                print(
+                    " " * level,
+                    f"{ix}: {item.label.value} with name={item.name}",
+                )
             elif isinstance(item, DocItem):
                 print(" " * level, f"{ix}: {item.label.value}")
 
@@ -2126,6 +2265,7 @@ class DoclingDocument(BaseModel):
         indent: int = 4,
         text_width: int = -1,
         page_no: Optional[int] = None,
+        included_content_layers: set[ContentLayer] = DEFAULT_CONTENT_LAYERS,
     ):
         """Save to markdown."""
         artifacts_dir, reference_path = self._get_output_paths(filename, artifacts_dir)
@@ -2149,6 +2289,7 @@ class DoclingDocument(BaseModel):
             indent=indent,
             text_width=text_width,
             page_no=page_no,
+            included_content_layers=included_content_layers,
         )
 
         with open(filename, "w", encoding="utf-8") as fw:
@@ -2167,6 +2308,7 @@ class DoclingDocument(BaseModel):
         indent: int = 4,
         text_width: int = -1,
         page_no: Optional[int] = None,
+        included_content_layers: set[ContentLayer] = DEFAULT_CONTENT_LAYERS,
     ) -> str:
         r"""Serialize to Markdown.
 
@@ -2248,7 +2390,12 @@ class DoclingDocument(BaseModel):
             mdtexts.append(text)
 
         for ix, (item, level) in enumerate(
-            self.iterate_items(self.body, with_groups=True, page_no=page_no)
+            self.iterate_items(
+                self.body,
+                with_groups=True,
+                page_no=page_no,
+                included_content_layers=included_content_layers,
+            )
         ):
             # If we've moved to a lower level, we're exiting one or more groups
             if level < previous_level:
@@ -2417,6 +2564,7 @@ class DoclingDocument(BaseModel):
         page_no: Optional[int] = None,
         html_lang: str = "en",
         html_head: str = _HTML_DEFAULT_HEAD,
+        included_content_layers: set[ContentLayer] = DEFAULT_CONTENT_LAYERS,
     ):
         """Save to HTML."""
         artifacts_dir, reference_path = self._get_output_paths(filename, artifacts_dir)
@@ -2437,6 +2585,7 @@ class DoclingDocument(BaseModel):
             page_no=page_no,
             html_lang=html_lang,
             html_head=html_head,
+            included_content_layers=included_content_layers,
         )
 
         with open(filename, "w", encoding="utf-8") as fw:
@@ -2484,6 +2633,7 @@ class DoclingDocument(BaseModel):
         page_no: Optional[int] = None,
         html_lang: str = "en",
         html_head: str = _HTML_DEFAULT_HEAD,
+        included_content_layers: set[ContentLayer] = DEFAULT_CONTENT_LAYERS,
     ) -> str:
         r"""Serialize to HTML."""
 
@@ -2508,7 +2658,11 @@ class DoclingDocument(BaseModel):
 
             return (in_ordered_list, html_texts)
 
-        head_lines = ["<!DOCTYPE html>", f'<html lang="{html_lang}">', html_head]
+        head_lines = [
+            "<!DOCTYPE html>",
+            f'<html lang="{html_lang}">',
+            html_head,
+        ]
         html_texts: list[str] = []
 
         prev_level = 0  # Track the previous item's level
@@ -2525,7 +2679,12 @@ class DoclingDocument(BaseModel):
             return text
 
         for ix, (item, curr_level) in enumerate(
-            self.iterate_items(self.body, with_groups=True, page_no=page_no)
+            self.iterate_items(
+                self.body,
+                with_groups=True,
+                page_no=page_no,
+                included_content_layers=included_content_layers,
+            )
         ):
             # If we've moved to a lower level, we're exiting one or more groups
             if curr_level < prev_level and len(in_ordered_list) > 0:
@@ -2583,7 +2742,8 @@ class DoclingDocument(BaseModel):
                 section_level: int = min(item.level + 1, 6)
 
                 text = get_html_tag_with_text_direction(
-                    html_tag=f"h{section_level}", text=_prepare_tag_content(item.text)
+                    html_tag=f"h{section_level}",
+                    text=_prepare_tag_content(item.text),
                 )
                 html_texts.append(text)
 
@@ -2840,13 +3000,19 @@ class DoclingDocument(BaseModel):
             self.iterate_items(
                 self.body,
                 with_groups=True,
-                included_content_layers={ContentLayer.BODY, ContentLayer.FURNITURE},
+                included_content_layers={
+                    ContentLayer.BODY,
+                    ContentLayer.FURNITURE,
+                },
             )
         ):
             # Close lists if we've moved to a lower nesting level
             if current_level < previous_level and ordered_list_stack:
                 ordered_list_stack = _close_lists(
-                    current_level, previous_level, ordered_list_stack, output_parts
+                    current_level,
+                    previous_level,
+                    ordered_list_stack,
+                    output_parts,
                 )
             previous_level = current_level
 
@@ -2954,7 +3120,10 @@ class DoclingDocument(BaseModel):
         return "".join(output_parts)
 
     def _export_to_indented_text(
-        self, indent="  ", max_text_len: int = -1, explicit_tables: bool = False
+        self,
+        indent="  ",
+        max_text_len: int = -1,
+        explicit_tables: bool = False,
     ):
         """Export the document to indented text to expose hierarchy."""
         result = []

{docling_core-2.19.0 → docling_core-2.20.0}/docling_core/types/doc/labels.py

@@ -140,6 +140,29 @@ class TableCellLabel(str, Enum):
         return str(self.value)
 
 
+class GraphCellLabel(str, Enum):
+    """GraphCellLabel."""
+
+    UNSPECIFIED = "unspecified"
+
+    KEY = "key"
+    VALUE = "value"
+
+    CHECKBOX = "checkbox"
+
+
+class GraphLinkLabel(str, Enum):
+    """GraphLinkLabel."""
+
+    UNSPECIFIED = "unspecified"
+
+    TO_VALUE = "to_value"
+    TO_KEY = "to_key"
+
+    TO_PARENT = "to_parent"
+    TO_CHILD = "to_child"
+
+
 class CodeLanguageLabel(str, Enum):
     """CodeLanguageLabel."""
 

{docling_core-2.19.0 → docling_core-2.20.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "docling-core"
-version = "2.19.0"
+version = "2.20.0"
 description = "A python library to define and validate data types in Docling."
 license = "MIT"
 authors = [