docling-core 2.26.4__py3-none-any.whl → 2.28.0__py3-none-any.whl

docling_core/types/doc/document.py

@@ -300,7 +300,7 @@ class TableCell(BaseModel):
     @classmethod
     def from_dict_format(cls, data: Any) -> Any:
         """from_dict_format."""
-        if isinstance(data, Dict):
+        if isinstance(data, dict):
             # Check if this is a native BoundingBox or a bbox from docling-ibm-models
             if (
                 # "bbox" not in data
@@ -2888,7 +2888,7 @@ class DoclingDocument(BaseModel):
         mode: str = "json",
         by_alias: bool = True,
         exclude_none: bool = True,
-    ) -> Dict:
+    ) -> Dict[str, Any]:
         """Export to dict."""
         out = self.model_dump(mode=mode, by_alias=by_alias, exclude_none=exclude_none)
 
@@ -3143,6 +3143,7 @@ class DoclingDocument(BaseModel):
         from_element: int = 0,
         to_element: int = sys.maxsize,
         labels: Optional[set[DocItemLabel]] = None,
+        enable_chart_tables: bool = True,
         image_mode: ImageRefMode = ImageRefMode.PLACEHOLDER,
         formula_to_mathml: bool = True,
         page_no: Optional[int] = None,
@@ -3176,6 +3177,7 @@ class DoclingDocument(BaseModel):
             start_idx=from_element,
             stop_idx=to_element,
             image_mode=image_mode,
+            enable_chart_tables=enable_chart_tables,
             formula_to_mathml=formula_to_mathml,
             html_head=html_head,
             html_lang=html_lang,
@@ -4042,6 +4044,28 @@ class DoclingDocument(BaseModel):
         self.pages[page_no] = pitem
         return pitem
 
+    def get_visualization(
+        self, show_label: bool = True
+    ) -> dict[Optional[int], PILImage.Image]:
+        """Get visualization of the document as images by page."""
+        from docling_core.transforms.visualizer.layout_visualizer import (
+            LayoutVisualizer,
+        )
+        from docling_core.transforms.visualizer.reading_order_visualizer import (
+            ReadingOrderVisualizer,
+        )
+
+        visualizer = ReadingOrderVisualizer(
+            base_visualizer=LayoutVisualizer(
+                params=LayoutVisualizer.Params(
+                    show_label=show_label,
+                ),
+            ),
+        )
+        images = visualizer.get_visualization(doc=self)
+
+        return images
+
     @field_validator("version")
     @classmethod
     def check_version_is_compatible(cls, v: str) -> str:

docling_core/types/doc/labels.py

@@ -45,6 +45,7 @@ class DocItemLabel(str, Enum):
             DocItemLabel.PAGE_FOOTER: (204, 255, 204),
             DocItemLabel.PAGE_HEADER: (204, 255, 204),
             DocItemLabel.PICTURE: (255, 204, 164),
+            DocItemLabel.CHART: (255, 204, 164),
             DocItemLabel.SECTION_HEADER: (255, 153, 153),
             DocItemLabel.TABLE: (255, 204, 204),
             DocItemLabel.TEXT: (255, 255, 153),
@@ -58,7 +59,7 @@ class DocItemLabel(str, Enum):
             DocItemLabel.PARAGRAPH: (255, 255, 153),
             DocItemLabel.REFERENCE: (176, 224, 230),
         }
-        return color_map[label]
+        return color_map.get(label, (0, 0, 0))
 
 
 class GroupLabel(str, Enum):

docling_core/types/doc/page.py

@@ -10,6 +10,7 @@ from enum import Enum
 from pathlib import Path
 from typing import (
     Annotated,
+    Any,
     Dict,
     Iterator,
     List,
@@ -538,7 +539,7 @@ class SegmentedPdfPage(SegmentedPage):
                 cells.append(pc)
         return cells
 
-    def export_to_dict(self) -> Dict:
+    def export_to_dict(self) -> Dict[str, Any]:
         """Export the page data to a dictionary.
 
         Returns:
@@ -1150,7 +1151,7 @@ class PdfTableOfContents(BaseModel):
 
     children: List["PdfTableOfContents"] = []
 
-    def export_to_dict(self, mode: str = "json") -> Dict:
+    def export_to_dict(self, mode: str = "json") -> Dict[str, Any]:
         """Export the table of contents to a dictionary.
 
         Args:
@@ -1212,7 +1213,7 @@ class ParsedPdfDocument(BaseModel):
     def export_to_dict(
         self,
         mode: str = "json",
-    ) -> Dict:
+    ) -> Dict[str, Any]:
         """Export the document to a dictionary.
 
         Args:

docling_core/types/legacy_doc/document.py

@@ -6,7 +6,7 @@
 """Models for the Docling Document data type."""
 
 from datetime import datetime
-from typing import Dict, Generic, Optional, Union
+from typing import Any, Dict, Generic, Optional, Union
 
 from pydantic import (
     AnyHttpUrl,
@@ -434,7 +434,7 @@ class ExportedCCSDocument(
 
         return pagedims
 
-    def export_to_dict(self) -> Dict:
+    def export_to_dict(self) -> Dict[str, Any]:
         """export_to_dict."""
         return self.model_dump(mode="json", by_alias=True, exclude_none=True)
 

{docling_core-2.26.4.dist-info → docling_core-2.28.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: docling-core
-Version: 2.26.4
+Version: 2.28.0
 Summary: A python library to define and validate data types in Docling.
 Home-page: https://github.com/docling-project
 License: MIT
@@ -26,6 +26,7 @@ Classifier: Topic :: Scientific/Engineering :: Information Analysis
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Classifier: Typing :: Typed
 Provides-Extra: chunking
+Provides-Extra: chunking-openai
 Requires-Dist: jsonref (>=1.1.0,<2.0.0)
 Requires-Dist: jsonschema (>=4.16.0,<5.0.0)
 Requires-Dist: latex2mathml (>=3.77.0,<4.0.0)
@@ -33,8 +34,9 @@ Requires-Dist: pandas (>=2.1.4,<3.0.0)
 Requires-Dist: pillow (>=10.0.0,<12.0.0)
 Requires-Dist: pydantic (>=2.6.0,<3.0.0,!=2.10.0,!=2.10.1,!=2.10.2)
 Requires-Dist: pyyaml (>=5.1,<7.0.0)
-Requires-Dist: semchunk (>=2.2.0,<3.0.0) ; extra == "chunking"
+Requires-Dist: semchunk (>=2.2.0,<3.0.0) ; extra == "chunking" or extra == "chunking-openai"
 Requires-Dist: tabulate (>=0.9.0,<0.10.0)
+Requires-Dist: tiktoken (>=0.9.0,<0.10.0) ; extra == "chunking-openai"
 Requires-Dist: transformers (>=4.34.0,<5.0.0) ; extra == "chunking"
 Requires-Dist: typer (>=0.12.5,<0.16.0)
 Requires-Dist: typing-extensions (>=4.12.2,<5.0.0)

{docling_core-2.26.4.dist-info → docling_core-2.28.0.dist-info}/RECORD

@@ -3,12 +3,12 @@ docling_core/cli/__init__.py,sha256=C63yWifzpA0IV7YWDatpAdrhoV8zjqxAKv0xMf09VdM,
 docling_core/cli/view.py,sha256=gwxSBYhGqwznMR8pdXaEuAh2bjFD5X_g11xFYSgFgtM,1764
 docling_core/experimental/__init__.py,sha256=XnAVSUHbA6OFhNSpoYqSD3u83-xVaUaki1DIKFw69Ew,99
 docling_core/experimental/serializer/__init__.py,sha256=CECQlMoCDUxkg4RAUdC3itA3I3qFhKhe2HcYghN6_xw,105
-docling_core/experimental/serializer/base.py,sha256=1sD1v5rWC4MT_Y6BWpMDjUAwuEqC0TR9YjQJZlhPt50,5901
-docling_core/experimental/serializer/common.py,sha256=iQUJPRZUhpGMi_s4makkZcINy5sdtxu2ehS9N8lnoMM,17332
-docling_core/experimental/serializer/doctags.py,sha256=e97FJHh77x--g2t1O2YprBzF8lkihn_xOr59EjnR7ag,17794
-docling_core/experimental/serializer/html.py,sha256=C-xf_PH28R6DoISFDoZwBK67rsfFmTemgZ_VW2i0haY,31896
+docling_core/experimental/serializer/base.py,sha256=9bgpWA0oMmZNRc3yIuZVnu5bJ1glClBsswtVF1vYwMI,6046
+docling_core/experimental/serializer/common.py,sha256=uviwBuYowzqvCbY-vy8v2VaEadJISk9aDETrkrfDo38,17437
+docling_core/experimental/serializer/doctags.py,sha256=RbHdqmFJ-t3oUvCsv0QjbIZqgUajPrt41jMaJGp4sdA,17874
+docling_core/experimental/serializer/html.py,sha256=By7NoDXQ4GDW-iFf8zWCYuU4f_TOHA8i86eGk60d4WM,33070
 docling_core/experimental/serializer/html_styles.py,sha256=-jBwS4EU7yfKoz0GSoxhwx90OmIKieO6TwPw57IuxcA,4692
-docling_core/experimental/serializer/markdown.py,sha256=5bvONhaA1EdAD0c3WlWfr2x2KmRaSZd8muG-91XVHgc,17733
+docling_core/experimental/serializer/markdown.py,sha256=WineuzwGDbFhbqEdz-sNWYewrUwBM0zfj88T8URaq6w,17877
 docling_core/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 docling_core/resources/schemas/doc/ANN.json,sha256=04U5j-PU9m5w7IagJ_rHcAx7qUtLkUuaWZO9GuYHnTA,4202
 docling_core/resources/schemas/doc/DOC.json,sha256=9tVKpCqDGGq3074Nn5qlUCdTN-5k1Q0ri_scJblwnLE,6686
@@ -27,14 +27,22 @@ docling_core/transforms/__init__.py,sha256=P81y_oqkiTN4Ld5crh1gQ6BbHqqR6C6nBt9AC
 docling_core/transforms/chunker/__init__.py,sha256=YdizSKXLmmK9eyYBsarHWr8Mx_AoA0PT7c0absibZMk,306
 docling_core/transforms/chunker/base.py,sha256=kJaRrGQynglG9wpy0IaAYTf4MKheWH5BAPzx4LE9yIg,2824
 docling_core/transforms/chunker/hierarchical_chunker.py,sha256=iYzA65INFo89klc94jixuzQP8ivywe-3aVYznt2Csv8,8287
-docling_core/transforms/chunker/hybrid_chunker.py,sha256=JPKKgfAdHqkYp4qyZWZyjJ3fYFq9lgD-mTaVVnm5T0Y,10936
+docling_core/transforms/chunker/hybrid_chunker.py,sha256=i2rxSE_6JZPClljcA_HVf0Pq5KgLyILhzG7CwRFcTIE,11888
+docling_core/transforms/chunker/tokenizer/__init__.py,sha256=-bhXOTpoI7SYk7vn47z8Ek-RZFjJk4TfZawxsFuNHnE,34
+docling_core/transforms/chunker/tokenizer/base.py,sha256=2gOBQPYJYC0iWXOgMG3DiNP7xEBtii7DYcib0iECq5o,575
+docling_core/transforms/chunker/tokenizer/huggingface.py,sha256=JQ-D3b5vTPQbvu4HaMfYqFzSBLbV_HnmoBGv7d6Kqn4,2220
+docling_core/transforms/chunker/tokenizer/openai.py,sha256=zt2kwcC-r8MafeEG0CESab8E4RIC9aaFXxxnxOGyTMA,918
+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=G_xPs5S_87RPPAIMKM6ryMU2aV_zGLYUTOlTQprIRD4,7336
+docling_core/transforms/visualizer/reading_order_visualizer.py,sha256=XXVuiI-Y0AH5uJCXINmfzcSSkTwR55-4fL6TOgzir6Y,5203
 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=sM3IyFXzVh2WT8IGh5nejXYh8sf39yBh8TBSlHeJ9CI,12611
-docling_core/types/doc/document.py,sha256=gK9-qiMi74p0jPeAHW6YGKl-O0ZSYE-p36MQAco3lx4,139341
-docling_core/types/doc/labels.py,sha256=Kmrrdmd6ejXomeXDlxjpmBEpPxMCYOc_3I2GSaoqqi4,5748
-docling_core/types/doc/page.py,sha256=QI1D5p63AxboT6PnHa7UlbPmH2i2_E3qIk_Gk2fdrxs,40270
+docling_core/types/doc/document.py,sha256=uYQTUEeZ40T5698Xff7NhC3iTbk1F76omZNvHIUmrfc,140174
+docling_core/types/doc/labels.py,sha256=3QgteZZ4jKi0fideTuTnuriviJBwew-5RKE4pse7Ppk,5812
+docling_core/types/doc/page.py,sha256=44tK6XM6Py0pK7zTyJ4kaZ5MLj8PvXIiw31hoQYa-Xs,40309
 docling_core/types/doc/tokens.py,sha256=z22l9J81_sg9CYMvOuLmPuLsNT7h_s7wao2UT89DvI8,9278
 docling_core/types/doc/utils.py,sha256=SaiQD-WMMooFm1bMqwatU-IGhtG048iKJb-ppnJit_k,2250
 docling_core/types/gen/__init__.py,sha256=C6TuCfvpSnSL5XDOFMcYHUY2-i08vvfOGRcdu6Af0pI,124
@@ -45,7 +53,7 @@ docling_core/types/legacy_doc/base.py,sha256=aBKBunw6M6nvEq4lqP1cfFWK3GpGa6PXwNQ
 docling_core/types/legacy_doc/doc_ann.py,sha256=CIQHW8yzu70bsMR9gtu7dqe4oz603Tq2eDDt9sh-tYo,1203
 docling_core/types/legacy_doc/doc_ocr.py,sha256=FfFqHAyMSbFt5cKeE7QLcxS0qUweBilBJoN9CH2TsQs,1394
 docling_core/types/legacy_doc/doc_raw.py,sha256=LrvQ9DhNjBRy98p_F9PUyHZeTGAxMKWqJzY4WJ7v-xs,3895
-docling_core/types/legacy_doc/document.py,sha256=AW8AIBM19k-HtTmXPsFKagqd6gi9THJdB4RsPb1C5F0,24534
+docling_core/types/legacy_doc/document.py,sha256=lEuxUS03YrY4dKvfzB1I208x6LtD0zukV9QU0hfjuwM,24549
 docling_core/types/legacy_doc/tokens.py,sha256=uU_MYW_p7ypf7eYICFBvxdnVaPZ7CQnvZmbJ6oPrtEA,6134
 docling_core/types/nlp/__init__.py,sha256=hGcztAeVK7xkRBqRRvc4zbY4PGeJ0r0QrEsetnSx9nI,119
 docling_core/types/nlp/qa.py,sha256=TyZjubqkEoREv0YzmuLKlq4WW_TnJNj7BoBY1_r2a1E,2731
@@ -65,8 +73,8 @@ docling_core/utils/generate_jsonschema.py,sha256=uNX1O5XnjyB5nA66XqZXTt3YbGuR2ty
 docling_core/utils/legacy.py,sha256=SqNQAxl97aHfoJEsC9vZcMJg5FNkmqKPFi-wdSrnfI0,24442
 docling_core/utils/validate.py,sha256=aQ11UbFyl8iD_N7yTTZmm_VVeXz8KcCyn3GLXgkfYRM,2049
 docling_core/utils/validators.py,sha256=azcrndLzhNkTWnbFSu9shJ5D3j_znnLrIFA5R8hzmGU,2798
-docling_core-2.26.4.dist-info/LICENSE,sha256=2M9-6EoQ1sxFztTOkXGAtwUDJvnWaAHdB9BYWVwGkIw,1087
-docling_core-2.26.4.dist-info/METADATA,sha256=S9qfeO5R35zkBwzIYnd5q3-O3LBl3-WOsQRlL50klqw,5843
-docling_core-2.26.4.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-docling_core-2.26.4.dist-info/entry_points.txt,sha256=oClcdb2L2RKx4jdqUykY16Kum_f0_whwWhGzIodyidc,216
-docling_core-2.26.4.dist-info/RECORD,,
+docling_core-2.28.0.dist-info/LICENSE,sha256=2M9-6EoQ1sxFztTOkXGAtwUDJvnWaAHdB9BYWVwGkIw,1087
+docling_core-2.28.0.dist-info/METADATA,sha256=uWyLwSsIWmUuQvfTYctf24fkDeYck3PAE9UsjSf85z8,5976
+docling_core-2.28.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+docling_core-2.28.0.dist-info/entry_points.txt,sha256=oClcdb2L2RKx4jdqUykY16Kum_f0_whwWhGzIodyidc,216
+docling_core-2.28.0.dist-info/RECORD,,