docling-core 2.34.2__tar.gz → 2.36.0__tar.gz

{docling_core-2.34.2 → docling_core-2.36.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-core
-Version: 2.34.2
+Version: 2.36.0
 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-2.34.2 → docling_core-2.36.0}/docling_core/transforms/visualizer/layout_visualizer.py

@@ -40,6 +40,7 @@ class LayoutVisualizer(BaseVisualizer):
         """Layout visualization parameters."""
 
         show_label: bool = True
+        content_layers: set[ContentLayer] = {cl for cl in ContentLayer}
 
     base_visualizer: Optional[BaseVisualizer] = None
     params: Params = Params()
@@ -119,7 +120,10 @@ class LayoutVisualizer(BaseVisualizer):
                     )
 
     def _draw_doc_layout(
-        self, doc: DoclingDocument, images: Optional[dict[Optional[int], Image]] = None
+        self,
+        doc: DoclingDocument,
+        images: Optional[dict[Optional[int], Image]] = None,
+        included_content_layers: Optional[set[ContentLayer]] = None,
     ):
         """Draw the document clusters and optionaly the reading order."""
         clusters = []
@@ -128,6 +132,9 @@ class LayoutVisualizer(BaseVisualizer):
         if images is not None:
             my_images = images
 
+        if included_content_layers is None:
+            included_content_layers = {c for c in ContentLayer}
+
         # Initialise `my_images` beforehand: sometimes, you have the
         # page-images but no DocItems!
         for page_nr, page in doc.pages.items():
@@ -141,9 +148,7 @@ class LayoutVisualizer(BaseVisualizer):
         prev_image = None
         prev_page_nr = None
         for idx, (elem, _) in enumerate(
-            doc.iterate_items(
-                included_content_layers={ContentLayer.BODY, ContentLayer.FURNITURE}
-            )
+            doc.iterate_items(included_content_layers=included_content_layers)
         ):
             if not isinstance(elem, DocItem):
                 continue

{docling_core-2.34.2 → docling_core-2.36.0}/docling_core/transforms/visualizer/reading_order_visualizer.py

@@ -5,6 +5,7 @@ from typing import Optional
 
 from PIL import ImageDraw
 from PIL.Image import Image
+from pydantic import BaseModel
 from typing_extensions import override
 
 from docling_core.transforms.visualizer.base import BaseVisualizer
@@ -14,7 +15,16 @@ from docling_core.types.doc.document import ContentLayer, DocItem, DoclingDocume
 class ReadingOrderVisualizer(BaseVisualizer):
     """Reading order visualizer."""
 
+    class Params(BaseModel):
+        """Layout visualization parameters."""
+
+        show_label: bool = True
+        content_layers: set[ContentLayer] = {
+            cl for cl in ContentLayer if cl != ContentLayer.BACKGROUND
+        }
+
     base_visualizer: Optional[BaseVisualizer] = None
+    params: Params = Params()
 
     def _draw_arrow(
         self,
@@ -71,7 +81,7 @@ class ReadingOrderVisualizer(BaseVisualizer):
         my_images: dict[Optional[int], Image] = images or {}
         prev_page = None
         for elem, _ in doc.iterate_items(
-            included_content_layers={ContentLayer.BODY, ContentLayer.FURNITURE},
+            included_content_layers=self.params.content_layers,
         ):
             if not isinstance(elem, DocItem):
                 continue

docling_core-2.36.0/docling_core/transforms/visualizer/table_visualizer.py

@@ -0,0 +1,135 @@
+"""Define classes for layout visualization."""
+
+import logging
+from copy import deepcopy
+from typing import Optional
+
+from PIL import ImageDraw
+from PIL.Image import Image
+from pydantic import BaseModel
+from typing_extensions import override
+
+from docling_core.transforms.visualizer.base import BaseVisualizer
+from docling_core.types.doc.document import ContentLayer, DoclingDocument, TableItem
+
+_log = logging.getLogger(__name__)
+
+
+class TableVisualizer(BaseVisualizer):
+    """Table visualizer."""
+
+    class Params(BaseModel):
+        """Table visualization parameters."""
+
+        # show_Label: bool = False
+        show_cells: bool = True
+        # show_rows: bool = False
+        # show_cols: bool = False
+
+    base_visualizer: Optional[BaseVisualizer] = None
+    params: Params = Params()
+
+    def _draw_table_cells(
+        self,
+        table: TableItem,
+        page_image: Image,
+        page_height: float,
+        scale_x: float,
+        scale_y: float,
+    ):
+        """Draw individual table cells."""
+        draw = ImageDraw.Draw(page_image, "RGBA")
+
+        for cell in table.data.table_cells:
+            if cell.bbox is not None:
+
+                tl_bbox = cell.bbox.to_top_left_origin(page_height=page_height)
+
+                cell_color = (256, 0, 0, 32)  # Transparent black for cells
+
+                cx0, cy0, cx1, cy1 = tl_bbox.as_tuple()
+                cx0 *= scale_x
+                cx1 *= scale_x
+                cy0 *= scale_y
+                cy1 *= scale_y
+
+                draw.rectangle(
+                    [(cx0, cy0), (cx1, cy1)],
+                    outline=(256, 0, 0, 128),
+                    fill=cell_color,
+                )
+
+    def _draw_doc_tables(
+        self,
+        doc: DoclingDocument,
+        images: Optional[dict[Optional[int], Image]] = None,
+        included_content_layers: Optional[set[ContentLayer]] = None,
+    ):
+        """Draw the document tables."""
+        my_images: dict[Optional[int], Image] = {}
+
+        if images is not None:
+            my_images = images
+
+        if included_content_layers is None:
+            included_content_layers = {c for c in ContentLayer}
+
+        # Initialise `my_images` beforehand: sometimes, you have the
+        # page-images but no DocItems!
+        for page_nr, page in doc.pages.items():
+            page_image = doc.pages[page_nr].image
+            if page_image is None or (pil_img := page_image.pil_image) is None:
+                raise RuntimeError("Cannot visualize document without images")
+            elif page_nr not in my_images:
+                image = deepcopy(pil_img)
+                my_images[page_nr] = image
+
+        for idx, (elem, _) in enumerate(
+            doc.iterate_items(included_content_layers=included_content_layers)
+        ):
+            if not isinstance(elem, TableItem):
+                continue
+            if len(elem.prov) == 0:
+                continue  # Skip elements without provenances
+
+            if len(elem.prov) == 1:
+
+                page_nr = elem.prov[0].page_no
+
+                if page_nr in my_images:
+                    image = my_images[page_nr]
+
+                    if self.params.show_cells:
+                        self._draw_table_cells(
+                            table=elem,
+                            page_height=doc.pages[page_nr].size.height,
+                            page_image=image,
+                            scale_x=image.width / doc.pages[page_nr].size.width,
+                            scale_y=image.height / doc.pages[page_nr].size.height,
+                        )
+
+                else:
+                    raise RuntimeError(f"Cannot visualize page-image for {page_nr}")
+
+            else:
+                _log.error("Can not yet visualise tables with multiple provenances")
+
+        return my_images
+
+    @override
+    def get_visualization(
+        self,
+        *,
+        doc: DoclingDocument,
+        **kwargs,
+    ) -> dict[Optional[int], Image]:
+        """Get visualization of the document as images by page."""
+        base_images = (
+            self.base_visualizer.get_visualization(doc=doc, **kwargs)
+            if self.base_visualizer
+            else None
+        )
+        return self._draw_doc_tables(
+            doc=doc,
+            images=base_images,
+        )

{docling_core-2.34.2 → docling_core-2.36.0}/docling_core/types/doc/document.py

@@ -623,6 +623,7 @@ class ContentLayer(str, Enum):
 
     BODY = "body"
     FURNITURE = "furniture"
+    BACKGROUND = "background"
 
 
 DEFAULT_CONTENT_LAYERS = {ContentLayer.BODY}
@@ -860,6 +861,7 @@ class TextItem(DocItem):
         DocItemLabel.PARAGRAPH,
         DocItemLabel.REFERENCE,
         DocItemLabel.TEXT,
+        DocItemLabel.EMPTY_VALUE,
     ]
 
     orig: str  # untreated representation
@@ -2867,23 +2869,46 @@ class DoclingDocument(BaseModel):
 
     def print_element_tree(self):
         """Print_element_tree."""
-        for ix, (item, level) in enumerate(self.iterate_items(with_groups=True)):
+        for ix, (item, level) in enumerate(
+            self.iterate_items(
+                with_groups=True,
+                traverse_pictures=True,
+                included_content_layers={cl for cl in ContentLayer},
+            )
+        ):
             if isinstance(item, GroupItem):
                 print(
                     " " * level,
                     f"{ix}: {item.label.value} with name={item.name}",
                 )
+            elif isinstance(item, TextItem):
+                print(
+                    " " * level,
+                    f"{ix}: {item.label.value}: {item.text[:min(len(item.text), 100)]}",
+                )
+
             elif isinstance(item, DocItem):
                 print(" " * level, f"{ix}: {item.label.value}")
 
     def export_to_element_tree(self) -> str:
         """Export_to_element_tree."""
         texts = []
-        for ix, (item, level) in enumerate(self.iterate_items(with_groups=True)):
+        for ix, (item, level) in enumerate(
+            self.iterate_items(
+                with_groups=True,
+                traverse_pictures=True,
+                included_content_layers={cl for cl in ContentLayer},
+            )
+        ):
             if isinstance(item, GroupItem):
                 texts.append(
                     " " * level + f"{ix}: {item.label.value} with name={item.name}"
                 )
+            elif isinstance(item, TextItem):
+                texts.append(
+                    " " * level
+                    + f"{ix}: {item.label.value}: {item.text[:min(len(item.text), 100)]}"
+                )
             elif isinstance(item, DocItem):
                 texts.append(" " * level + f"{ix}: {item.label.value}")
 

{docling_core-2.34.2 → docling_core-2.36.0}/docling_core/types/doc/labels.py

@@ -27,6 +27,9 @@ class DocItemLabel(str, Enum):
     KEY_VALUE_REGION = "key_value_region"
     GRADING_SCALE = "grading_scale"  # for elements in forms, questionaires representing a grading scale
     # e.g. [strongly disagree | ... | ... | strongly agree]
+    # e.g. ★★☆☆☆
+    HANDWRITTEN_TEXT = "handwritten_text"
+    EMPTY_VALUE = "empty_value"  # used for empty value fields in fillable forms
 
     # Additional labels for markup-based formats (e.g. HTML, Word)
     PARAGRAPH = "paragraph"
@@ -60,6 +63,9 @@ class DocItemLabel(str, Enum):
             DocItemLabel.KEY_VALUE_REGION: (183, 65, 14),
             DocItemLabel.PARAGRAPH: (255, 255, 153),
             DocItemLabel.REFERENCE: (176, 224, 230),
+            DocItemLabel.GRADING_SCALE: (255, 204, 204),
+            DocItemLabel.HANDWRITTEN_TEXT: (204, 255, 204),
+            DocItemLabel.EMPTY_VALUE: (220, 220, 220),
         }
         return color_map.get(label, (0, 0, 0))
 
@@ -166,7 +172,6 @@ class GraphCellLabel(str, Enum):
     KEY = "key"  # used to designate a key (label) of a key-value element
     VALUE = "value"  # Data value with or without explicit Key, but filled in,
     # e.g. telephone number, address, quantity, name, date
-    EMPTY_VALUE = "empty_value"  # used for empty value fields in fillable forms
     CHECKBOX = "checkbox"
 
     def __str__(self):
@@ -179,7 +184,6 @@ class GraphCellLabel(str, Enum):
         color_map = {
             GraphCellLabel.KEY: (255, 0, 0),
             GraphCellLabel.VALUE: (0, 255, 0),
-            GraphCellLabel.EMPTY_VALUE: (0, 0, 255),
         }
         return color_map.get(label, (0, 0, 0))
 

{docling_core-2.34.2 → docling_core-2.36.0}/docling_core.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-core
-Version: 2.34.2
+Version: 2.36.0
 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-2.34.2 → docling_core-2.36.0}/docling_core.egg-info/SOURCES.txt

@@ -45,6 +45,7 @@ docling_core/transforms/visualizer/__init__.py
 docling_core/transforms/visualizer/base.py
 docling_core/transforms/visualizer/layout_visualizer.py
 docling_core/transforms/visualizer/reading_order_visualizer.py
+docling_core/transforms/visualizer/table_visualizer.py
 docling_core/types/__init__.py
 docling_core/types/base.py
 docling_core/types/doc/__init__.py

{docling_core-2.34.2 → docling_core-2.36.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "docling-core"
-version = "2.34.2"  # DO NOT EDIT, updated automatically
+version = "2.36.0"  # DO NOT EDIT, updated automatically
 description = "A python library to define and validate data types in Docling."
 license = "MIT"
 license-files = ["LICENSE"]

{docling_core-2.34.2 → docling_core-2.36.0}/test/test_visualization.py

@@ -2,6 +2,7 @@ from pathlib import Path
 
 import PIL.Image
 
+from docling_core.transforms.visualizer.table_visualizer import TableVisualizer
 from docling_core.types.doc.document import DoclingDocument
 
 from .test_data_gen_flag import GEN_TEST_DATA
@@ -52,3 +53,16 @@ def test_doc_visualization_no_label():
                 exp_file=VIZ_TEST_DATA_PATH / f"{src.stem}_viz_wout_lbl_p{k}.png",
                 actual=viz_pages[k],
             )
+
+
+def test_table_visualization_no_label():
+    src = Path("./test/data/doc/2408.09869v3_enriched.json")
+    doc = DoclingDocument.load_from_json(src)
+
+    visualizer = TableVisualizer()
+    viz_pages = visualizer.get_visualization(doc=doc)
+
+    verify(
+        exp_file=VIZ_TEST_DATA_PATH / f"{src.stem}_table_viz_wout_lbl_p5.png",
+        actual=viz_pages[5],
+    )