docling-core 2.15.0__py3-none-any.whl → 3.0.0__py3-none-any.whl

docling_core/types/doc/base.py

@@ -1,6 +1,5 @@
 """Models for the base data types."""
 
-import copy
 from enum import Enum
 from typing import Tuple
 
@@ -53,33 +52,53 @@ class BoundingBox(BaseModel):
         """height."""
         return abs(self.t - self.b)
 
-    def scaled(self, scale: float) -> "BoundingBox":
-        """scaled.
-
-        :param scale: float:
-
-        """
-        out_bbox = copy.deepcopy(self)
-        out_bbox.l *= scale
-        out_bbox.r *= scale
-        out_bbox.t *= scale
-        out_bbox.b *= scale
-
-        return out_bbox
-
-    def normalized(self, page_size: Size) -> "BoundingBox":
-        """normalized.
-
-        :param page_size: Size:
-
-        """
-        out_bbox = copy.deepcopy(self)
-        out_bbox.l /= page_size.width
-        out_bbox.r /= page_size.width
-        out_bbox.t /= page_size.height
-        out_bbox.b /= page_size.height
-
-        return out_bbox
+    def resize_by_scale(self, x_scale: float, y_scale: float):
+        """resize_by_scale."""
+        return BoundingBox(
+            l=self.l * x_scale,
+            r=self.r * x_scale,
+            t=self.t * y_scale,
+            b=self.b * y_scale,
+            coord_origin=self.coord_origin,
+        )
+
+    def scale_to_size(self, old_size: Size, new_size: Size):
+        """scale_to_size."""
+        return self.resize_by_scale(
+            x_scale=new_size.width / old_size.width,
+            y_scale=new_size.height / old_size.height,
+        )
+
+    # same as before, but using the implementation above
+    def scaled(self, scale: float):
+        """scaled."""
+        return self.resize_by_scale(x_scale=scale, y_scale=scale)
+
+    # same as before, but using the implementation above
+    def normalized(self, page_size: Size):
+        """normalized."""
+        return self.scale_to_size(
+            old_size=page_size, new_size=Size(height=1.0, width=1.0)
+        )
+
+    def expand_by_scale(self, x_scale: float, y_scale: float) -> "BoundingBox":
+        """expand_to_size."""
+        if self.coord_origin == CoordOrigin.TOPLEFT:
+            return BoundingBox(
+                l=self.l - self.width * x_scale,
+                r=self.r + self.width * x_scale,
+                t=self.t - self.height * y_scale,
+                b=self.b + self.height * y_scale,
+                coord_origin=self.coord_origin,
+            )
+        elif self.coord_origin == CoordOrigin.BOTTOMLEFT:
+            return BoundingBox(
+                l=self.l - self.width * x_scale,
+                r=self.r + self.width * x_scale,
+                t=self.t + self.height * y_scale,
+                b=self.b - self.height * y_scale,
+                coord_origin=self.coord_origin,
+            )
 
     def as_tuple(self) -> Tuple[float, float, float, float]:
         """as_tuple."""
@@ -116,26 +135,27 @@ class BoundingBox(BaseModel):
 
     def area(self) -> float:
         """area."""
-        area = (self.r - self.l) * (self.b - self.t)
-        if self.coord_origin == CoordOrigin.BOTTOMLEFT:
-            area = -area
-        return area
+        return abs(self.r - self.l) * abs(self.b - self.t)
 
     def intersection_area_with(self, other: "BoundingBox") -> float:
-        """intersection_area_with.
-
-        :param other: "BoundingBox":
+        """Calculate the intersection area with another bounding box."""
+        if self.coord_origin != other.coord_origin:
+            raise ValueError("BoundingBoxes have different CoordOrigin")
 
-        """
         # Calculate intersection coordinates
         left = max(self.l, other.l)
-        top = max(self.t, other.t)
         right = min(self.r, other.r)
-        bottom = min(self.b, other.b)
+
+        if self.coord_origin == CoordOrigin.TOPLEFT:
+            bottom = max(self.t, other.t)
+            top = min(self.b, other.b)
+        elif self.coord_origin == CoordOrigin.BOTTOMLEFT:
+            top = min(self.t, other.t)
+            bottom = max(self.b, other.b)
 
         # Calculate intersection dimensions
         width = right - left
-        height = bottom - top
+        height = top - bottom
 
         # If the bounding boxes do not overlap, width or height will be negative
         if width <= 0 or height <= 0:
@@ -143,6 +163,27 @@ class BoundingBox(BaseModel):
 
         return width * height
 
+    def intersection_over_union(
+        self, other: "BoundingBox", eps: float = 1.0e-6
+    ) -> float:
+        """intersection_over_union."""
+        intersection_area = self.intersection_area_with(other=other)
+
+        union_area = (
+            abs(self.l - self.r) * abs(self.t - self.b)
+            + abs(other.l - other.r) * abs(other.t - other.b)
+            - intersection_area
+        )
+
+        return intersection_area / (union_area + eps)
+
+    def intersection_over_self(
+        self, other: "BoundingBox", eps: float = 1.0e-6
+    ) -> float:
+        """intersection_over_self."""
+        intersection_area = self.intersection_area_with(other=other)
+        return intersection_area / self.area()
+
     def to_bottom_left_origin(self, page_height: float) -> "BoundingBox":
         """to_bottom_left_origin.
 
@@ -176,3 +217,151 @@ class BoundingBox(BaseModel):
                 b=page_height - self.b,  # self.t
                 coord_origin=CoordOrigin.TOPLEFT,
             )
+
+    def overlaps(self, other: "BoundingBox") -> bool:
+        """overlaps."""
+        return self.overlaps_horizontally(other=other) and self.overlaps_vertically(
+            other=other
+        )
+
+    def overlaps_horizontally(self, other: "BoundingBox") -> bool:
+        """Check if two bounding boxes overlap horizontally."""
+        return not (self.r <= other.l or other.r <= self.l)
+
+    def overlaps_vertically(self, other: "BoundingBox") -> bool:
+        """Check if two bounding boxes overlap vertically."""
+        if self.coord_origin != other.coord_origin:
+            raise ValueError("BoundingBoxes have different CoordOrigin")
+
+        # Normalize coordinates if needed
+        if self.coord_origin == CoordOrigin.BOTTOMLEFT:
+            return not (self.t <= other.b or other.t <= self.b)
+        elif self.coord_origin == CoordOrigin.TOPLEFT:
+            return not (self.b <= other.t or other.b <= self.t)
+
+    def overlaps_vertically_with_iou(self, other: "BoundingBox", iou: float) -> bool:
+        """overlaps_y_with_iou."""
+        if (
+            self.coord_origin == CoordOrigin.BOTTOMLEFT
+            and other.coord_origin == CoordOrigin.BOTTOMLEFT
+        ):
+
+            if self.overlaps_vertically(other=other):
+
+                u0 = min(self.b, other.b)
+                u1 = max(self.t, other.t)
+
+                i0 = max(self.b, other.b)
+                i1 = min(self.t, other.t)
+
+                iou_ = float(i1 - i0) / float(u1 - u0)
+                return (iou_) > iou
+
+            return False
+
+        elif (
+            self.coord_origin == CoordOrigin.TOPLEFT
+            and other.coord_origin == CoordOrigin.TOPLEFT
+        ):
+            if self.overlaps_vertically(other=other):
+                u0 = min(self.t, other.t)
+                u1 = max(self.b, other.b)
+
+                i0 = max(self.t, other.t)
+                i1 = min(self.b, other.b)
+
+                iou_ = float(i1 - i0) / float(u1 - u0)
+                return (iou_) > iou
+
+            return False
+        else:
+            raise ValueError("BoundingBoxes have different CoordOrigin")
+
+        return False
+
+    def is_left_of(self, other: "BoundingBox") -> bool:
+        """is_left_of."""
+        return self.l < other.l
+
+    def is_strictly_left_of(self, other: "BoundingBox", eps: float = 0.001) -> bool:
+        """is_strictly_left_of."""
+        return (self.r + eps) < other.l
+
+    def is_above(self, other: "BoundingBox") -> bool:
+        """is_above."""
+        if (
+            self.coord_origin == CoordOrigin.BOTTOMLEFT
+            and other.coord_origin == CoordOrigin.BOTTOMLEFT
+        ):
+            return self.t > other.t
+
+        elif (
+            self.coord_origin == CoordOrigin.TOPLEFT
+            and other.coord_origin == CoordOrigin.TOPLEFT
+        ):
+            return self.t < other.t
+
+        else:
+            raise ValueError("BoundingBoxes have different CoordOrigin")
+
+        return False
+
+    def is_strictly_above(self, other: "BoundingBox", eps: float = 1.0e-3) -> bool:
+        """is_strictly_above."""
+        if (
+            self.coord_origin == CoordOrigin.BOTTOMLEFT
+            and other.coord_origin == CoordOrigin.BOTTOMLEFT
+        ):
+            return (self.b + eps) > other.t
+
+        elif (
+            self.coord_origin == CoordOrigin.TOPLEFT
+            and other.coord_origin == CoordOrigin.TOPLEFT
+        ):
+            return (self.b + eps) < other.t
+
+        else:
+            raise ValueError("BoundingBoxes have different CoordOrigin")
+
+        return False
+
+    def is_horizontally_connected(
+        self, elem_i: "BoundingBox", elem_j: "BoundingBox"
+    ) -> bool:
+        """is_horizontally_connected."""
+        if (
+            self.coord_origin == CoordOrigin.BOTTOMLEFT
+            and elem_i.coord_origin == CoordOrigin.BOTTOMLEFT
+            and elem_j.coord_origin == CoordOrigin.BOTTOMLEFT
+        ):
+            min_ij = min(elem_i.b, elem_j.b)
+            max_ij = max(elem_i.t, elem_j.t)
+
+            if self.b < max_ij and min_ij < self.t:  # overlap_y
+                return False
+
+            if self.l < elem_i.r and elem_j.l < self.r:
+                return True
+
+            return False
+
+        elif (
+            self.coord_origin == CoordOrigin.TOPLEFT
+            and elem_i.coord_origin == CoordOrigin.TOPLEFT
+            and elem_j.coord_origin == CoordOrigin.TOPLEFT
+        ):
+            min_ij = min(elem_i.t, elem_j.t)
+            max_ij = max(elem_i.b, elem_j.b)
+
+            if self.t < max_ij and min_ij < self.b:  # overlap_y
+                return False
+
+            if self.l < elem_i.r and elem_j.l < self.r:
+                return True
+
+            return False
+
+        else:
+            raise ValueError("BoundingBoxes have different CoordOrigin")
+
+        return False

docling_core/types/doc/document.py

@@ -585,7 +585,8 @@ class DocItem(
         crop_bbox = (
             self.prov[0]
             .bbox.to_top_left_origin(page_height=page.size.height)
-            .scaled(scale=page_image.height / page.size.height)
+            .scale_to_size(old_size=page.size, new_size=page.image.size)
+            # .scaled(scale=page_image.height / page.size.height)
         )
         return page_image.crop(crop_bbox.as_tuple())
 
@@ -1515,6 +1516,9 @@ class DoclingDocument(BaseModel):
         elif label in [DocItemLabel.SECTION_HEADER]:
             return self.add_heading(text=text, orig=orig, prov=prov, parent=parent)
 
+        elif label in [DocItemLabel.CODE]:
+            return self.add_code(text=text, orig=orig, prov=prov, parent=parent)
+
         else:
 
             if not parent:
@@ -1991,6 +1995,7 @@ class DoclingDocument(BaseModel):
         to_element: int = sys.maxsize,
         labels: set[DocItemLabel] = DEFAULT_EXPORT_LABELS,
         strict_text: bool = False,
+        escaping_underscores: bool = True,
         image_placeholder: str = "<!-- image -->",
         image_mode: ImageRefMode = ImageRefMode.PLACEHOLDER,
         indent: int = 4,
@@ -2013,6 +2018,7 @@ class DoclingDocument(BaseModel):
             to_element=to_element,
             labels=labels,
             strict_text=strict_text,
+            escaping_underscores=escaping_underscores,
             image_placeholder=image_placeholder,
             image_mode=image_mode,
             indent=indent,
@@ -2030,6 +2036,7 @@ class DoclingDocument(BaseModel):
         to_element: int = sys.maxsize,
         labels: set[DocItemLabel] = DEFAULT_EXPORT_LABELS,
         strict_text: bool = False,
+        escaping_underscores: bool = True,
         image_placeholder: str = "<!-- image -->",
         image_mode: ImageRefMode = ImageRefMode.PLACEHOLDER,
         indent: int = 4,
@@ -2055,6 +2062,9 @@ class DoclingDocument(BaseModel):
         :param strict_text: bool: Whether to only include the text content
             of the document. (Default value = False).
         :type strict_text: bool = False
+        :param escaping_underscores: bool: Whether to escape underscores in the
+            text content of the document. (Default value = True).
+        :type escaping_underscores: bool = True
         :param image_placeholder: The placeholder to include to position
             images in the markdown. (Default value = "\<!-- image --\>").
         :type image_placeholder: str = "<!-- image -->"
@@ -2157,6 +2167,10 @@ class DoclingDocument(BaseModel):
                 text = f"{list_indent}{marker} {item.text}"
                 mdtexts.append(text)
 
+            elif isinstance(item, TextItem) and item.label in [DocItemLabel.FORMULA]:
+                in_list = False
+                mdtexts.append(f"$${item.text}$$")
+
             elif isinstance(item, TextItem) and item.label in labels:
                 in_list = False
                 if len(item.text) and text_width > 0:
@@ -2205,10 +2219,14 @@ class DoclingDocument(BaseModel):
             """Escape underscores but leave them intact in the URL.."""
             # Firstly, identify all the URL patterns.
             url_pattern = r"!\[.*?\]\((.*?)\)"
+            # Matches both inline ($...$) and block ($$...$$) LaTeX equations:
+            latex_pattern = r"\$\$?(?:\\.|[^$\\])*\$\$?"
+            combined_pattern = f"({url_pattern})|({latex_pattern})"
+
             parts = []
             last_end = 0
 
-            for match in re.finditer(url_pattern, text):
+            for match in re.finditer(combined_pattern, text):
                 # Text to add before the URL (needs to be escaped)
                 before_url = text[last_end : match.start()]
                 parts.append(re.sub(r"(?<!\\)_", r"\_", before_url))
@@ -2223,7 +2241,8 @@ class DoclingDocument(BaseModel):
 
             return "".join(parts)
 
-        mdtext = escape_underscores(mdtext)
+        if escaping_underscores:
+            mdtext = escape_underscores(mdtext)
 
         return mdtext
 
@@ -2241,6 +2260,7 @@ class DoclingDocument(BaseModel):
             to_element,
             labels,
             strict_text=True,
+            escaping_underscores=False,
             image_placeholder="",
         )
 

{docling_core-2.15.0.dist-info → docling_core-3.0.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: docling-core
-Version: 2.15.0
+Version: 3.0.0
 Summary: A python library to define and validate data types in Docling.
 Home-page: https://ds4sd.github.io/
 License: MIT

{docling_core-2.15.0.dist-info → docling_core-3.0.0.dist-info}/RECORD

@@ -23,8 +23,8 @@ docling_core/transforms/chunker/hybrid_chunker.py,sha256=kokjDdxjc_gygOokQwYFVnH
 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=O5GH3psrV5dSWhNzZRpOtGEImk-XpqLJ7Et5VGm0tcU,4651
-docling_core/types/doc/document.py,sha256=VlB_scwBHTKUJD7q89PE1zH4L3Q5T-esQ1Sg9csnxbo,93847
+docling_core/types/doc/base.py,sha256=lMRNq1DUK7K26L2VNZRqFaItCSZ6m9BdYTVaJA98PZQ,11495
+docling_core/types/doc/document.py,sha256=1AGj1xQJ1XLP2nkgHIn8-BDNDic5OPll35cg051B6Ts,94839
 docling_core/types/doc/labels.py,sha256=8Luymal9SKXTwyqq1ONKiUTxuMo_nRMYfBkRPFkdSSo,5306
 docling_core/types/doc/tokens.py,sha256=GMtm5TsNljBPaMYkgmD3WWZmC0FHqKF9imKEEySz4ps,6020
 docling_core/types/doc/utils.py,sha256=YDOh_ZD1Y7OmCEDdCLJ_MO5K3HA67nc_acfhOK6WztU,1439
@@ -56,8 +56,8 @@ docling_core/utils/generate_jsonschema.py,sha256=uNX1O5XnjyB5nA66XqZXTt3YbGuR2ty
 docling_core/utils/legacy.py,sha256=xfp7U0JqjI60K3loWiNTk8w08_KfCUzTb2MNULBOIz4,24396
 docling_core/utils/validate.py,sha256=aQ11UbFyl8iD_N7yTTZmm_VVeXz8KcCyn3GLXgkfYRM,2049
 docling_core/utils/validators.py,sha256=azcrndLzhNkTWnbFSu9shJ5D3j_znnLrIFA5R8hzmGU,2798
-docling_core-2.15.0.dist-info/LICENSE,sha256=2M9-6EoQ1sxFztTOkXGAtwUDJvnWaAHdB9BYWVwGkIw,1087
-docling_core-2.15.0.dist-info/METADATA,sha256=FyObOk5pW4t7tc5FjK8AshHE1FfXIFx3ok2i2M3lyiw,5744
-docling_core-2.15.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
-docling_core-2.15.0.dist-info/entry_points.txt,sha256=oClcdb2L2RKx4jdqUykY16Kum_f0_whwWhGzIodyidc,216
-docling_core-2.15.0.dist-info/RECORD,,
+docling_core-3.0.0.dist-info/LICENSE,sha256=2M9-6EoQ1sxFztTOkXGAtwUDJvnWaAHdB9BYWVwGkIw,1087
+docling_core-3.0.0.dist-info/METADATA,sha256=xwCWMsl3H5K_OCT_4Iz77zE4RVAW8JnKy19X02P-rAI,5743
+docling_core-3.0.0.dist-info/WHEEL,sha256=sP946D7jFCHeNz5Iq4fL4Lu-PrWrFsgfLXbbkciIZwg,88
+docling_core-3.0.0.dist-info/entry_points.txt,sha256=oClcdb2L2RKx4jdqUykY16Kum_f0_whwWhGzIodyidc,216
+docling_core-3.0.0.dist-info/RECORD,,