docling-core 2.39.0__py3-none-any.whl → 2.41.0__py3-none-any.whl

docling_core/types/doc/base.py

@@ -3,7 +3,7 @@
 from enum import Enum
 from typing import List, Tuple
 
-from pydantic import BaseModel
+from pydantic import BaseModel, FieldSerializationInfo, field_serializer
 
 
 class ImageRefMode(str, Enum):
@@ -21,12 +21,28 @@ class CoordOrigin(str, Enum):
     BOTTOMLEFT = "BOTTOMLEFT"
 
 
+_CTX_COORD_PREC = "coord_prec"
+
+
+def _serialize_precision(
+    value: float, info: FieldSerializationInfo, ctx_key: str
+) -> float:
+    precision = info.context.get(ctx_key) if info.context else None
+    if isinstance(precision, int):
+        return round(value, precision)
+    return value
+
+
 class Size(BaseModel):
     """Size."""
 
     width: float = 0.0
     height: float = 0.0
 
+    @field_serializer("width", "height")
+    def _serialize(self, value: float, info: FieldSerializationInfo) -> float:
+        return _serialize_precision(value, info, _CTX_COORD_PREC)
+
     def as_tuple(self):
         """as_tuple."""
         return (self.width, self.height)
@@ -52,6 +68,10 @@ class BoundingBox(BaseModel):
         """height."""
         return abs(self.t - self.b)
 
+    @field_serializer("l", "t", "r", "b")
+    def _serialize(self, value: float, info: FieldSerializationInfo) -> float:
+        return _serialize_precision(value, info, _CTX_COORD_PREC)
+
     def resize_by_scale(self, x_scale: float, y_scale: float):
         """resize_by_scale."""
         return BoundingBox(

docling_core/types/doc/document.py

@@ -26,8 +26,10 @@ from pydantic import (
     BaseModel,
     ConfigDict,
     Field,
+    FieldSerializationInfo,
     StringConstraints,
     computed_field,
+    field_serializer,
     field_validator,
     model_validator,
     validate_call,
@@ -38,7 +40,12 @@ from typing_extensions import Annotated, Self, deprecated
 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 CoordOrigin, ImageRefMode
+from docling_core.types.doc.base import (
+    _CTX_COORD_PREC,
+    CoordOrigin,
+    ImageRefMode,
+    _serialize_precision,
+)
 from docling_core.types.doc.labels import (
     CodeLanguageLabel,
     DocItemLabel,
@@ -85,6 +92,8 @@ DOCUMENT_TOKENS_EXPORT_LABELS.update(
     ]
 )
 
+_CTX_CONFID_PREC = "confid_prec"
+
 
 class BaseAnnotation(BaseModel):
     """Base class for all annotation types."""
@@ -98,6 +107,10 @@ class PictureClassificationClass(BaseModel):
     class_name: str
     confidence: float
 
+    @field_serializer("confidence")
+    def _serialize(self, value: float, info: FieldSerializationInfo) -> float:
+        return _serialize_precision(value, info, _CTX_CONFID_PREC)
+
 
 class PictureClassificationData(BaseAnnotation):
     """PictureClassificationData."""
@@ -125,6 +138,10 @@ class PictureMoleculeData(BaseAnnotation):
     segmentation: List[Tuple[float, float]]
     provenance: str
 
+    @field_serializer("confidence")
+    def _serialize(self, value: float, info: FieldSerializationInfo) -> float:
+        return _serialize_precision(value, info, _CTX_CONFID_PREC)
+
 
 class MiscAnnotation(BaseAnnotation):
     """MiscAnnotation."""
@@ -731,9 +748,11 @@ class ProvenanceItem(BaseModel):
 class ContentLayer(str, Enum):
     """ContentLayer."""
 
-    BODY = "body"
-    FURNITURE = "furniture"
-    BACKGROUND = "background"
+    BODY = "body"  # main content of the document
+    FURNITURE = "furniture"  # eg page-headers/footers
+    BACKGROUND = "background"  # eg watermarks
+    INVISIBLE = "invisible"  # hidden or invisible text
+    NOTES = "notes"  # author/speaker notes, corrections, etc
 
 
 DEFAULT_CONTENT_LAYERS = {ContentLayer.BODY}
@@ -3046,6 +3065,8 @@ class DoclingDocument(BaseModel):
         artifacts_dir: Optional[Path] = None,
         image_mode: ImageRefMode = ImageRefMode.EMBEDDED,
         indent: int = 2,
+        coord_precision: Optional[int] = None,
+        confid_precision: Optional[int] = None,
     ):
         """Save as json."""
         if isinstance(filename, str):
@@ -3059,7 +3080,9 @@ class DoclingDocument(BaseModel):
             artifacts_dir, image_mode, reference_path=reference_path
         )
 
-        out = new_doc.export_to_dict()
+        out = new_doc.export_to_dict(
+            coord_precision=coord_precision, confid_precision=confid_precision
+        )
         with open(filename, "w", encoding="utf-8") as fw:
             json.dump(out, fw, indent=indent)
 
@@ -3085,6 +3108,8 @@ class DoclingDocument(BaseModel):
         artifacts_dir: Optional[Path] = None,
         image_mode: ImageRefMode = ImageRefMode.EMBEDDED,
         default_flow_style: bool = False,
+        coord_precision: Optional[int] = None,
+        confid_precision: Optional[int] = None,
     ):
         """Save as yaml."""
         if isinstance(filename, str):
@@ -3098,7 +3123,9 @@ class DoclingDocument(BaseModel):
             artifacts_dir, image_mode, reference_path=reference_path
         )
 
-        out = new_doc.export_to_dict()
+        out = new_doc.export_to_dict(
+            coord_precision=coord_precision, confid_precision=confid_precision
+        )
         with open(filename, "w", encoding="utf-8") as fw:
             yaml.dump(out, fw, default_flow_style=default_flow_style)
 
@@ -3123,9 +3150,18 @@ class DoclingDocument(BaseModel):
         mode: str = "json",
         by_alias: bool = True,
         exclude_none: bool = True,
+        coord_precision: Optional[int] = None,
+        confid_precision: Optional[int] = None,
     ) -> Dict[str, Any]:
         """Export to dict."""
-        out = self.model_dump(mode=mode, by_alias=by_alias, exclude_none=exclude_none)
+        context = {}
+        if coord_precision is not None:
+            context[_CTX_COORD_PREC] = coord_precision
+        if confid_precision is not None:
+            context[_CTX_CONFID_PREC] = confid_precision
+        out = self.model_dump(
+            mode=mode, by_alias=by_alias, exclude_none=exclude_none, context=context
+        )
 
         return out
 

docling_core/types/doc/page.py

@@ -25,9 +25,21 @@ import numpy as np
 from PIL import Image as PILImage
 from PIL import ImageColor, ImageDraw, ImageFont
 from PIL.ImageFont import FreeTypeFont
-from pydantic import AnyUrl, BaseModel, Field, model_validator
+from pydantic import (
+    AnyUrl,
+    BaseModel,
+    Field,
+    FieldSerializationInfo,
+    field_serializer,
+    model_validator,
+)
 
-from docling_core.types.doc.base import BoundingBox, CoordOrigin
+from docling_core.types.doc.base import (
+    _CTX_COORD_PREC,
+    BoundingBox,
+    CoordOrigin,
+    _serialize_precision,
+)
 from docling_core.types.doc.document import ImageRef
 
 _logger = logging.getLogger(__name__)
@@ -105,6 +117,10 @@ class BoundingRectangle(BaseModel):
 
     coord_origin: CoordOrigin = CoordOrigin.BOTTOMLEFT
 
+    @field_serializer("r_x0", "r_y0", "r_x1", "r_y1", "r_x2", "r_y2", "r_x3", "r_y3")
+    def _serialize(self, value: float, info: FieldSerializationInfo) -> float:
+        return _serialize_precision(value, info, _CTX_COORD_PREC)
+
     @property
     def width(self) -> float:
         """Calculate the width of the rectangle."""
@@ -122,6 +138,8 @@ class BoundingRectangle(BaseModel):
         p_1 = ((self.r_x1 + self.r_x2) / 2.0, (self.r_y1 + self.r_y2) / 2.0)
 
         delta_x, delta_y = p_1[0] - p_0[0], p_1[1] - p_0[1]
+        if self.coord_origin == CoordOrigin.TOPLEFT:
+            delta_y = -delta_y
 
         if abs(delta_y) < 1.0e-3:
             angle = 0.0
@@ -131,8 +149,7 @@ class BoundingRectangle(BaseModel):
             angle = math.atan(delta_y / delta_x)
         if delta_x < 0:
             angle += np.pi
-        if angle < 0:
-            angle += 2 * np.pi
+        angle = angle % (2 * np.pi)
         return angle
 
     @property

{docling_core-2.39.0.dist-info → docling_core-2.41.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling-core
-Version: 2.39.0
+Version: 2.41.0
 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-2.39.0.dist-info → docling_core-2.41.0.dist-info}/RECORD

@@ -40,10 +40,10 @@ docling_core/transforms/visualizer/table_visualizer.py,sha256=iJPjk-XQSSCH3oujcj
 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=8hOhm5W9mArf3zwgfoMxDs1pHizhLFSAZlLu1tPBBRk,1641
-docling_core/types/doc/base.py,sha256=ndXquBrOKTFQApIJ5s2-zstj3xlVKRbJDSId0KOQnUg,14817
-docling_core/types/doc/document.py,sha256=Wu9st8xomDWF8a-GtHn6TSGbkqiMxPLzCUCnYXMhgOc,157212
+docling_core/types/doc/base.py,sha256=u8sFLA29x8QphvLzgy2wAKu3HXyM2GODfBXqEwQMrTY,15527
+docling_core/types/doc/document.py,sha256=YAJIIdT2fBnlp8ASWvzJTjUbil_ZCwuBBSjmiwhZ1KI,158630
 docling_core/types/doc/labels.py,sha256=-W1-LW6z0J9F9ExJqR0Wd1WeqWTaY3Unm-j1UkQGlC4,7330
-docling_core/types/doc/page.py,sha256=GV9UnGCvvqs6KD_ac3hF6b_NH6M6IevsL5iSt8WWVCI,41221
+docling_core/types/doc/page.py,sha256=CH9DY3LLgnUdhRuJBWfnkDkPBdRzz9yi4el1LsxJSME,41651
 docling_core/types/doc/tokens.py,sha256=z22l9J81_sg9CYMvOuLmPuLsNT7h_s7wao2UT89DvI8,9278
 docling_core/types/doc/utils.py,sha256=JpAi7x9DHksFlIj_gRJPcSZOHa8AHvVPEO_K9aSnw4c,2608
 docling_core/types/gen/__init__.py,sha256=C6TuCfvpSnSL5XDOFMcYHUY2-i08vvfOGRcdu6Af0pI,124
@@ -74,9 +74,9 @@ docling_core/utils/generate_jsonschema.py,sha256=uNX1O5XnjyB5nA66XqZXTt3YbGuR2ty
 docling_core/utils/legacy.py,sha256=5lghO48OEcV9V51tRnH3YSKgLtdqhr-Q5C_OcJZ8TOs,24392
 docling_core/utils/validate.py,sha256=aQ11UbFyl8iD_N7yTTZmm_VVeXz8KcCyn3GLXgkfYRM,2049
 docling_core/utils/validators.py,sha256=azcrndLzhNkTWnbFSu9shJ5D3j_znnLrIFA5R8hzmGU,2798
-docling_core-2.39.0.dist-info/licenses/LICENSE,sha256=2M9-6EoQ1sxFztTOkXGAtwUDJvnWaAHdB9BYWVwGkIw,1087
-docling_core-2.39.0.dist-info/METADATA,sha256=Yn1ptbMIE-Tj9y-pSXuKpdotyq34lyHnFWRxWNT-9qY,6453
-docling_core-2.39.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-docling_core-2.39.0.dist-info/entry_points.txt,sha256=ER4zROQWkFMHIrY-oqY5E4HeCcCIg8dLkNztYGxdb7c,59
-docling_core-2.39.0.dist-info/top_level.txt,sha256=O-tcXpGiurlud-1ZxMq1b-OmrfAVA4sajcgWU32RtfA,13
-docling_core-2.39.0.dist-info/RECORD,,
+docling_core-2.41.0.dist-info/licenses/LICENSE,sha256=2M9-6EoQ1sxFztTOkXGAtwUDJvnWaAHdB9BYWVwGkIw,1087
+docling_core-2.41.0.dist-info/METADATA,sha256=CqsXanxB2dd22G__-Ws0XdLDzkf9uwMGmx98V2h9f3k,6453
+docling_core-2.41.0.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+docling_core-2.41.0.dist-info/entry_points.txt,sha256=ER4zROQWkFMHIrY-oqY5E4HeCcCIg8dLkNztYGxdb7c,59
+docling_core-2.41.0.dist-info/top_level.txt,sha256=O-tcXpGiurlud-1ZxMq1b-OmrfAVA4sajcgWU32RtfA,13
+docling_core-2.41.0.dist-info/RECORD,,