docling-core 2.35.0__py3-none-any.whl → 2.36.0__py3-none-any.whl

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()

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/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/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-2.35.0.dist-info → docling_core-2.36.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-core
-Version: 2.35.0
+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-2.35.0.dist-info → docling_core-2.36.0.dist-info}/RECORD

@@ -34,15 +34,15 @@ docling_core/transforms/serializer/html_styles.py,sha256=-jBwS4EU7yfKoz0GSoxhwx9
 docling_core/transforms/serializer/markdown.py,sha256=wfMNrjA4wMehWLCejAhEN1eQPRixUO1SyL6ojkKkzZY,20614
 docling_core/transforms/visualizer/__init__.py,sha256=gUfF25yiJ_KO46ZIUNqZQOZGy2PLx6gnnr6AZYxKHXI,35
 docling_core/transforms/visualizer/base.py,sha256=aEF7b3rHq6DVdX8zDYEPoq55BHDYe4Hh_97lBdcW4lY,555
-docling_core/transforms/visualizer/layout_visualizer.py,sha256=N3SA9sMkg2bEZ_2r52FpwRXcI3EJ2M5P9LYK4Az4jqQ,7968
-docling_core/transforms/visualizer/reading_order_visualizer.py,sha256=-ej5uLriNtr9C7YBHXMg8sZfB9Uc8cSRr1bJ8FVjpY8,5320
+docling_core/transforms/visualizer/layout_visualizer.py,sha256=hpq7OnyBgGxt3iW3_aNy9KH_0kmKdgoiJIFPcA2SSHU,8040
+docling_core/transforms/visualizer/reading_order_visualizer.py,sha256=yBra_W33bb16BxrTqP-ABu5NfRplTEJgu3dKdew3zKA,5601
 docling_core/transforms/visualizer/table_visualizer.py,sha256=XlLMSROyRW2UtAjKTltcESSs_rdQNKjO3QvO7ET7uc0,4275
 docling_core/types/__init__.py,sha256=MVRSgsk5focwGyAplh_TRR3dEecIXpd98g_u3zZ5HXo,260
 docling_core/types/base.py,sha256=PusJskRVL19y-hq0BgXr5e8--QEqSqLnFNJ8UbOqW88,8318
 docling_core/types/doc/__init__.py,sha256=bysJn2iwjAHwThSWDPXEdVUUij7p_ax12_nx2_0CMdg,653
 docling_core/types/doc/base.py,sha256=ndXquBrOKTFQApIJ5s2-zstj3xlVKRbJDSId0KOQnUg,14817
-docling_core/types/doc/document.py,sha256=UMQ7eSfXM0l2QPBpc4fg67OlYaWEJsgWqke_fk9jCUw,148549
-docling_core/types/doc/labels.py,sha256=vp4h3e7AmBvezRmgrfuPehjAHTZOufphErLB4ENhdME,7171
+docling_core/types/doc/document.py,sha256=elFR5J7O9FUWXiweNK2W7S-cPvAakdzkMls0Uh4ViU8,149361
+docling_core/types/doc/labels.py,sha256=JiciRK7_DOkebsrfQ6PVCvS__TsKgWn1ANk84BeB14k,7359
 docling_core/types/doc/page.py,sha256=1JMPwglaTITBvg959L_pcWPb-fXoDYGh-e_tGZMzVMQ,41060
 docling_core/types/doc/tokens.py,sha256=z22l9J81_sg9CYMvOuLmPuLsNT7h_s7wao2UT89DvI8,9278
 docling_core/types/doc/utils.py,sha256=SaiQD-WMMooFm1bMqwatU-IGhtG048iKJb-ppnJit_k,2250
@@ -74,9 +74,9 @@ docling_core/utils/generate_jsonschema.py,sha256=uNX1O5XnjyB5nA66XqZXTt3YbGuR2ty
 docling_core/utils/legacy.py,sha256=DrI3QGoL755ZCIoKHF74-pTWm8R0zfFo2C2vB5dT2aY,24463
 docling_core/utils/validate.py,sha256=aQ11UbFyl8iD_N7yTTZmm_VVeXz8KcCyn3GLXgkfYRM,2049
 docling_core/utils/validators.py,sha256=azcrndLzhNkTWnbFSu9shJ5D3j_znnLrIFA5R8hzmGU,2798
-docling_core-2.35.0.dist-info/licenses/LICENSE,sha256=2M9-6EoQ1sxFztTOkXGAtwUDJvnWaAHdB9BYWVwGkIw,1087
-docling_core-2.35.0.dist-info/METADATA,sha256=Gube58hbnoDQGoeGmaK-yrMulAuKHiw7lUeGKxzSDsc,6453
-docling_core-2.35.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-docling_core-2.35.0.dist-info/entry_points.txt,sha256=ER4zROQWkFMHIrY-oqY5E4HeCcCIg8dLkNztYGxdb7c,59
-docling_core-2.35.0.dist-info/top_level.txt,sha256=O-tcXpGiurlud-1ZxMq1b-OmrfAVA4sajcgWU32RtfA,13
-docling_core-2.35.0.dist-info/RECORD,,
+docling_core-2.36.0.dist-info/licenses/LICENSE,sha256=2M9-6EoQ1sxFztTOkXGAtwUDJvnWaAHdB9BYWVwGkIw,1087
+docling_core-2.36.0.dist-info/METADATA,sha256=8CnZkQHylNT1mgEEs_lIB18f2NL96R3kFAl-rBYVR0U,6453
+docling_core-2.36.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+docling_core-2.36.0.dist-info/entry_points.txt,sha256=ER4zROQWkFMHIrY-oqY5E4HeCcCIg8dLkNztYGxdb7c,59
+docling_core-2.36.0.dist-info/top_level.txt,sha256=O-tcXpGiurlud-1ZxMq1b-OmrfAVA4sajcgWU32RtfA,13
+docling_core-2.36.0.dist-info/RECORD,,