feat: Make Page.parsed_page the only source of truth for text cells, add OCR cells to it (#1745)

* Keep page.parsed_page.textline_cells and page.cells in sync, including OCR

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* Make page.parsed_page the only source of truth for text cells

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* Small fix

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* Correctly compute PDF boxes from pymupdf

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* Use different OCR engine order

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* Add type hints and fix mypy

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* One more test fix

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* Remove with pypdfium2_lock from caller sites

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* Fix typing

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

---------

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

docling/backend/docling_parse_backend.py

@@ -7,12 +7,17 @@ from typing import List, Optional, Union
 
 import pypdfium2 as pdfium
 from docling_core.types.doc import BoundingBox, CoordOrigin, Size
-from docling_core.types.doc.page import BoundingRectangle, SegmentedPdfPage, TextCell
+from docling_core.types.doc.page import (
+    BoundingRectangle,
+    SegmentedPdfPage,
+    TextCell,
+)
 from docling_parse.pdf_parsers import pdf_parser_v1
 from PIL import Image, ImageDraw
 from pypdfium2 import PdfPage
 
 from docling.backend.pdf_backend import PdfDocumentBackend, PdfPageBackend
+from docling.backend.pypdfium2_backend import get_pdf_page_geometry
 from docling.datamodel.document import InputDocument
 
 _log = logging.getLogger(__name__)
@@ -36,6 +41,51 @@ class DoclingParsePageBackend(PdfPageBackend):
     def is_valid(self) -> bool:
         return self.valid
 
+    def _compute_text_cells(self) -> List[TextCell]:
+        """Compute text cells from docling-parse data."""
+        cells: List[TextCell] = []
+        cell_counter = 0
+
+        if not self.valid:
+            return cells
+
+        page_size = self.get_size()
+
+        parser_width = self._dpage["width"]
+        parser_height = self._dpage["height"]
+
+        for i in range(len(self._dpage["cells"])):
+            rect = self._dpage["cells"][i]["box"]["device"]
+            x0, y0, x1, y1 = rect
+
+            if x1 < x0:
+                x0, x1 = x1, x0
+            if y1 < y0:
+                y0, y1 = y1, y0
+
+            text_piece = self._dpage["cells"][i]["content"]["rnormalized"]
+            cells.append(
+                TextCell(
+                    index=cell_counter,
+                    text=text_piece,
+                    orig=text_piece,
+                    from_ocr=False,
+                    rect=BoundingRectangle.from_bounding_box(
+                        BoundingBox(
+                            l=x0 * page_size.width / parser_width,
+                            b=y0 * page_size.height / parser_height,
+                            r=x1 * page_size.width / parser_width,
+                            t=y1 * page_size.height / parser_height,
+                            coord_origin=CoordOrigin.BOTTOMLEFT,
+                        )
+                    ).to_top_left_origin(page_size.height),
+                )
+            )
+
+            cell_counter += 1
+
+        return cells
+
     def get_text_in_rect(self, bbox: BoundingBox) -> str:
         if not self.valid:
             return ""
@@ -70,75 +120,27 @@ class DoclingParsePageBackend(PdfPageBackend):
         return text_piece
 
     def get_segmented_page(self) -> Optional[SegmentedPdfPage]:
-        return None
+        if not self.valid:
+            return None
+
+        text_cells = self._compute_text_cells()
+
+        # Get the PDF page geometry from pypdfium2
+        dimension = get_pdf_page_geometry(self._ppage)
+
+        # Create SegmentedPdfPage
+        return SegmentedPdfPage(
+            dimension=dimension,
+            textline_cells=text_cells,
+            char_cells=[],
+            word_cells=[],
+            has_lines=len(text_cells) > 0,
+            has_words=False,
+            has_chars=False,
+        )
 
     def get_text_cells(self) -> Iterable[TextCell]:
-        cells: List[TextCell] = []
+        return self._compute_text_cells()
-        cell_counter = 0
-
-        if not self.valid:
-            return cells
-
-        page_size = self.get_size()
-
-        parser_width = self._dpage["width"]
-        parser_height = self._dpage["height"]
-
-        for i in range(len(self._dpage["cells"])):
-            rect = self._dpage["cells"][i]["box"]["device"]
-            x0, y0, x1, y1 = rect
-
-            if x1 < x0:
-                x0, x1 = x1, x0
-            if y1 < y0:
-                y0, y1 = y1, y0
-
-            text_piece = self._dpage["cells"][i]["content"]["rnormalized"]
-            cells.append(
-                TextCell(
-                    index=cell_counter,
-                    text=text_piece,
-                    orig=text_piece,
-                    from_ocr=False,
-                    rect=BoundingRectangle.from_bounding_box(
-                        BoundingBox(
-                            # l=x0, b=y0, r=x1, t=y1,
-                            l=x0 * page_size.width / parser_width,
-                            b=y0 * page_size.height / parser_height,
-                            r=x1 * page_size.width / parser_width,
-                            t=y1 * page_size.height / parser_height,
-                            coord_origin=CoordOrigin.BOTTOMLEFT,
-                        )
-                    ).to_top_left_origin(page_size.height),
-                )
-            )
-
-            cell_counter += 1
-
-        def draw_clusters_and_cells():
-            image = (
-                self.get_page_image()
-            )  # make new image to avoid drawing on the saved ones
-            draw = ImageDraw.Draw(image)
-            for c in cells:
-                x0, y0, x1, y1 = c.rect.to_bounding_box().as_tuple()
-                cell_color = (
-                    random.randint(30, 140),
-                    random.randint(30, 140),
-                    random.randint(30, 140),
-                )
-                draw.rectangle([(x0, y0), (x1, y1)], outline=cell_color)
-            image.show()
-
-        # before merge:
-        # draw_clusters_and_cells()
-
-        # cells = merge_horizontal_cells(cells)
-
-        # after merge:
-        # draw_clusters_and_cells()
-
-        return cells
 
     def get_bitmap_rects(self, scale: float = 1) -> Iterable[BoundingBox]:
         AREA_THRESHOLD = 0  # 32 * 32

docling/backend/docling_parse_v2_backend.py

@@ -7,12 +7,19 @@ from typing import TYPE_CHECKING, List, Optional, Union
 
 import pypdfium2 as pdfium
 from docling_core.types.doc import BoundingBox, CoordOrigin
-from docling_core.types.doc.page import BoundingRectangle, SegmentedPdfPage, TextCell
+from docling_core.types.doc.page import (
+    BoundingRectangle,
+    PdfPageBoundaryType,
+    PdfPageGeometry,
+    SegmentedPdfPage,
+    TextCell,
+)
 from docling_parse.pdf_parsers import pdf_parser_v2
 from PIL import Image, ImageDraw
 from pypdfium2 import PdfPage
 
 from docling.backend.pdf_backend import PdfDocumentBackend, PdfPageBackend
+from docling.backend.pypdfium2_backend import get_pdf_page_geometry
 from docling.datamodel.base_models import Size
 from docling.utils.locks import pypdfium2_lock
 
@@ -40,6 +47,55 @@ class DoclingParseV2PageBackend(PdfPageBackend):
     def is_valid(self) -> bool:
         return self.valid
 
+    def _compute_text_cells(self) -> List[TextCell]:
+        """Compute text cells from docling-parse v2 data."""
+        cells: List[TextCell] = []
+        cell_counter = 0
+
+        if not self.valid:
+            return cells
+
+        page_size = self.get_size()
+
+        parser_width = self._dpage["sanitized"]["dimension"]["width"]
+        parser_height = self._dpage["sanitized"]["dimension"]["height"]
+
+        cells_data = self._dpage["sanitized"]["cells"]["data"]
+        cells_header = self._dpage["sanitized"]["cells"]["header"]
+
+        for i, cell_data in enumerate(cells_data):
+            x0 = cell_data[cells_header.index("x0")]
+            y0 = cell_data[cells_header.index("y0")]
+            x1 = cell_data[cells_header.index("x1")]
+            y1 = cell_data[cells_header.index("y1")]
+
+            if x1 < x0:
+                x0, x1 = x1, x0
+            if y1 < y0:
+                y0, y1 = y1, y0
+
+            text_piece = cell_data[cells_header.index("text")]
+            cells.append(
+                TextCell(
+                    index=cell_counter,
+                    text=text_piece,
+                    orig=text_piece,
+                    from_ocr=False,
+                    rect=BoundingRectangle.from_bounding_box(
+                        BoundingBox(
+                            l=x0 * page_size.width / parser_width,
+                            b=y0 * page_size.height / parser_height,
+                            r=x1 * page_size.width / parser_width,
+                            t=y1 * page_size.height / parser_height,
+                            coord_origin=CoordOrigin.BOTTOMLEFT,
+                        )
+                    ).to_top_left_origin(page_size.height),
+                )
+            )
+            cell_counter += 1
+
+        return cells
+
     def get_text_in_rect(self, bbox: BoundingBox) -> str:
         if not self.valid:
             return ""
@@ -81,73 +137,27 @@ class DoclingParseV2PageBackend(PdfPageBackend):
         return text_piece
 
     def get_segmented_page(self) -> Optional[SegmentedPdfPage]:
-        return None
+        if not self.valid:
+            return None
+
+        text_cells = self._compute_text_cells()
+
+        # Get the PDF page geometry from pypdfium2
+        dimension = get_pdf_page_geometry(self._ppage)
+
+        # Create SegmentedPdfPage
+        return SegmentedPdfPage(
+            dimension=dimension,
+            textline_cells=text_cells,
+            char_cells=[],
+            word_cells=[],
+            has_textlines=len(text_cells) > 0,
+            has_words=False,
+            has_chars=False,
+        )
 
     def get_text_cells(self) -> Iterable[TextCell]:
-        cells: List[TextCell] = []
+        return self._compute_text_cells()
-        cell_counter = 0
-
-        if not self.valid:
-            return cells
-
-        page_size = self.get_size()
-
-        parser_width = self._dpage["sanitized"]["dimension"]["width"]
-        parser_height = self._dpage["sanitized"]["dimension"]["height"]
-
-        cells_data = self._dpage["sanitized"]["cells"]["data"]
-        cells_header = self._dpage["sanitized"]["cells"]["header"]
-
-        for i, cell_data in enumerate(cells_data):
-            x0 = cell_data[cells_header.index("x0")]
-            y0 = cell_data[cells_header.index("y0")]
-            x1 = cell_data[cells_header.index("x1")]
-            y1 = cell_data[cells_header.index("y1")]
-
-            if x1 < x0:
-                x0, x1 = x1, x0
-            if y1 < y0:
-                y0, y1 = y1, y0
-
-            text_piece = cell_data[cells_header.index("text")]
-            cells.append(
-                TextCell(
-                    index=cell_counter,
-                    text=text_piece,
-                    orig=text_piece,
-                    from_ocr=False,
-                    rect=BoundingRectangle.from_bounding_box(
-                        BoundingBox(
-                            # l=x0, b=y0, r=x1, t=y1,
-                            l=x0 * page_size.width / parser_width,
-                            b=y0 * page_size.height / parser_height,
-                            r=x1 * page_size.width / parser_width,
-                            t=y1 * page_size.height / parser_height,
-                            coord_origin=CoordOrigin.BOTTOMLEFT,
-                        )
-                    ).to_top_left_origin(page_size.height),
-                )
-            )
-            cell_counter += 1
-
-        def draw_clusters_and_cells():
-            image = (
-                self.get_page_image()
-            )  # make new image to avoid drawing on the saved ones
-            draw = ImageDraw.Draw(image)
-            for c in cells:
-                x0, y0, x1, y1 = c.bbox.as_tuple()
-                cell_color = (
-                    random.randint(30, 140),
-                    random.randint(30, 140),
-                    random.randint(30, 140),
-                )
-                draw.rectangle([(x0, y0), (x1, y1)], outline=cell_color)
-            image.show()
-
-        # draw_clusters_and_cells()
-
-        return cells
 
     def get_bitmap_rects(self, scale: float = 1) -> Iterable[BoundingBox]:
         AREA_THRESHOLD = 0  # 32 * 32

docling/backend/docling_parse_v4_backend.py

@@ -59,20 +59,6 @@ class DoclingParseV4PageBackend(PdfPageBackend):
         return self._dpage
 
     def get_text_cells(self) -> Iterable[TextCell]:
-        page_size = self.get_size()
-
-        [tc.to_top_left_origin(page_size.height) for tc in self._dpage.textline_cells]
-
-        # for cell in self._dpage.textline_cells:
-        #     rect = cell.rect
-        #
-        #     assert (
-        #         rect.to_bounding_box().l <= rect.to_bounding_box().r
-        #     ), f"left is > right on bounding box {rect.to_bounding_box()} of rect {rect}"
-        #     assert (
-        #         rect.to_bounding_box().t <= rect.to_bounding_box().b
-        #     ), f"top is > bottom on bounding box {rect.to_bounding_box()} of rect {rect}"
-
         return self._dpage.textline_cells
 
     def get_bitmap_rects(self, scale: float = 1) -> Iterable[BoundingBox]:
@@ -171,12 +157,28 @@ class DoclingParseV4DocumentBackend(PdfDocumentBackend):
         self, page_no: int, create_words: bool = True, create_textlines: bool = True
     ) -> DoclingParseV4PageBackend:
         with pypdfium2_lock:
+            seg_page = self.dp_doc.get_page(
+                page_no + 1,
+                create_words=create_words,
+                create_textlines=create_textlines,
+            )
+
+            # In Docling, all TextCell instances are expected with top-left origin.
+            [
+                tc.to_top_left_origin(seg_page.dimension.height)
+                for tc in seg_page.textline_cells
+            ]
+            [
+                tc.to_top_left_origin(seg_page.dimension.height)
+                for tc in seg_page.char_cells
+            ]
+            [
+                tc.to_top_left_origin(seg_page.dimension.height)
+                for tc in seg_page.word_cells
+            ]
+
             return DoclingParseV4PageBackend(
-                self.dp_doc.get_page(
+                seg_page,
-                    page_no + 1,
-                    create_words=create_words,
-                    create_textlines=create_textlines,
-                ),
                 self._pdoc[page_no],
             )
 

docling/backend/pypdfium2_backend.py

@@ -8,7 +8,13 @@ from typing import TYPE_CHECKING, List, Optional, Union
 import pypdfium2 as pdfium
 import pypdfium2.raw as pdfium_c
 from docling_core.types.doc import BoundingBox, CoordOrigin, Size
-from docling_core.types.doc.page import BoundingRectangle, SegmentedPdfPage, TextCell
+from docling_core.types.doc.page import (
+    BoundingRectangle,
+    PdfPageBoundaryType,
+    PdfPageGeometry,
+    SegmentedPdfPage,
+    TextCell,
+)
 from PIL import Image, ImageDraw
 from pypdfium2 import PdfTextPage
 from pypdfium2._helpers.misc import PdfiumError
@@ -16,6 +22,76 @@ from pypdfium2._helpers.misc import PdfiumError
 from docling.backend.pdf_backend import PdfDocumentBackend, PdfPageBackend
 from docling.utils.locks import pypdfium2_lock
 
+
+def get_pdf_page_geometry(
+    ppage: pdfium.PdfPage,
+    angle: float = 0.0,
+    boundary_type: PdfPageBoundaryType = PdfPageBoundaryType.CROP_BOX,
+) -> PdfPageGeometry:
+    """
+    Create PdfPageGeometry from a pypdfium2 PdfPage object.
+
+    Args:
+        ppage: pypdfium2 PdfPage object
+        angle: Page rotation angle in degrees (default: 0.0)
+        boundary_type: The boundary type for the page (default: CROP_BOX)
+
+    Returns:
+        PdfPageGeometry with all the different bounding boxes properly set
+    """
+    with pypdfium2_lock:
+        # Get the main bounding box (intersection of crop_box and media_box)
+        bbox_tuple = ppage.get_bbox()
+        bbox = BoundingBox.from_tuple(bbox_tuple, CoordOrigin.BOTTOMLEFT)
+
+        # Get all the different page boxes from pypdfium2
+        media_box_tuple = ppage.get_mediabox()
+        crop_box_tuple = ppage.get_cropbox()
+        art_box_tuple = ppage.get_artbox()
+        bleed_box_tuple = ppage.get_bleedbox()
+        trim_box_tuple = ppage.get_trimbox()
+
+        # Convert to BoundingBox objects using existing from_tuple method
+        # pypdfium2 returns (x0, y0, x1, y1) in PDF coordinate system (bottom-left origin)
+        # Use bbox as fallback when specific box types are not defined
+        media_bbox = (
+            BoundingBox.from_tuple(media_box_tuple, CoordOrigin.BOTTOMLEFT)
+            if media_box_tuple
+            else bbox
+        )
+        crop_bbox = (
+            BoundingBox.from_tuple(crop_box_tuple, CoordOrigin.BOTTOMLEFT)
+            if crop_box_tuple
+            else bbox
+        )
+        art_bbox = (
+            BoundingBox.from_tuple(art_box_tuple, CoordOrigin.BOTTOMLEFT)
+            if art_box_tuple
+            else bbox
+        )
+        bleed_bbox = (
+            BoundingBox.from_tuple(bleed_box_tuple, CoordOrigin.BOTTOMLEFT)
+            if bleed_box_tuple
+            else bbox
+        )
+        trim_bbox = (
+            BoundingBox.from_tuple(trim_box_tuple, CoordOrigin.BOTTOMLEFT)
+            if trim_box_tuple
+            else bbox
+        )
+
+        return PdfPageGeometry(
+            angle=angle,
+            rect=BoundingRectangle.from_bounding_box(bbox),
+            boundary_type=boundary_type,
+            art_bbox=art_bbox,
+            bleed_bbox=bleed_bbox,
+            crop_bbox=crop_bbox,
+            media_bbox=media_bbox,
+            trim_bbox=trim_bbox,
+        )
+
+
 if TYPE_CHECKING:
     from docling.datamodel.document import InputDocument
 
@@ -41,38 +117,8 @@ class PyPdfiumPageBackend(PdfPageBackend):
     def is_valid(self) -> bool:
         return self.valid
 
-    def get_bitmap_rects(self, scale: float = 1) -> Iterable[BoundingBox]:
+    def _compute_text_cells(self) -> List[TextCell]:
-        AREA_THRESHOLD = 0  # 32 * 32
+        """Compute text cells from pypdfium."""
-        page_size = self.get_size()
-        with pypdfium2_lock:
-            for obj in self._ppage.get_objects(filter=[pdfium_c.FPDF_PAGEOBJ_IMAGE]):
-                pos = obj.get_pos()
-                cropbox = BoundingBox.from_tuple(
-                    pos, origin=CoordOrigin.BOTTOMLEFT
-                ).to_top_left_origin(page_height=page_size.height)
-
-                if cropbox.area() > AREA_THRESHOLD:
-                    cropbox = cropbox.scaled(scale=scale)
-
-                    yield cropbox
-
-    def get_text_in_rect(self, bbox: BoundingBox) -> str:
-        with pypdfium2_lock:
-            if not self.text_page:
-                self.text_page = self._ppage.get_textpage()
-
-        if bbox.coord_origin != CoordOrigin.BOTTOMLEFT:
-            bbox = bbox.to_bottom_left_origin(self.get_size().height)
-
-        with pypdfium2_lock:
-            text_piece = self.text_page.get_text_bounded(*bbox.as_tuple())
-
-        return text_piece
-
-    def get_segmented_page(self) -> Optional[SegmentedPdfPage]:
-        return None
-
-    def get_text_cells(self) -> Iterable[TextCell]:
         with pypdfium2_lock:
             if not self.text_page:
                 self.text_page = self._ppage.get_textpage()
@@ -203,30 +249,58 @@ class PyPdfiumPageBackend(PdfPageBackend):
 
             return merged_cells
 
-        def draw_clusters_and_cells():
+        return merge_horizontal_cells(cells)
-            image = (
-                self.get_page_image()
-            )  # make new image to avoid drawing on the saved ones
-            draw = ImageDraw.Draw(image)
-            for c in cells:
-                x0, y0, x1, y1 = c.rect.to_bounding_box().as_tuple()
-                cell_color = (
-                    random.randint(30, 140),
-                    random.randint(30, 140),
-                    random.randint(30, 140),
-                )
-                draw.rectangle([(x0, y0), (x1, y1)], outline=cell_color)
-            image.show()
 
-        # before merge:
+    def get_bitmap_rects(self, scale: float = 1) -> Iterable[BoundingBox]:
-        # draw_clusters_and_cells()
+        AREA_THRESHOLD = 0  # 32 * 32
+        page_size = self.get_size()
+        with pypdfium2_lock:
+            for obj in self._ppage.get_objects(filter=[pdfium_c.FPDF_PAGEOBJ_IMAGE]):
+                pos = obj.get_pos()
+                cropbox = BoundingBox.from_tuple(
+                    pos, origin=CoordOrigin.BOTTOMLEFT
+                ).to_top_left_origin(page_height=page_size.height)
 
-        cells = merge_horizontal_cells(cells)
+                if cropbox.area() > AREA_THRESHOLD:
+                    cropbox = cropbox.scaled(scale=scale)
 
-        # after merge:
+                    yield cropbox
-        # draw_clusters_and_cells()
 
-        return cells
+    def get_text_in_rect(self, bbox: BoundingBox) -> str:
+        with pypdfium2_lock:
+            if not self.text_page:
+                self.text_page = self._ppage.get_textpage()
+
+        if bbox.coord_origin != CoordOrigin.BOTTOMLEFT:
+            bbox = bbox.to_bottom_left_origin(self.get_size().height)
+
+        with pypdfium2_lock:
+            text_piece = self.text_page.get_text_bounded(*bbox.as_tuple())
+
+        return text_piece
+
+    def get_segmented_page(self) -> Optional[SegmentedPdfPage]:
+        if not self.valid:
+            return None
+
+        text_cells = self._compute_text_cells()
+
+        # Get the PDF page geometry from pypdfium2
+        dimension = get_pdf_page_geometry(self._ppage)
+
+        # Create SegmentedPdfPage
+        return SegmentedPdfPage(
+            dimension=dimension,
+            textline_cells=text_cells,
+            char_cells=[],
+            word_cells=[],
+            has_textlines=len(text_cells) > 0,
+            has_words=False,
+            has_chars=False,
+        )
+
+    def get_text_cells(self) -> Iterable[TextCell]:
+        return self._compute_text_cells()
 
     def get_page_image(
         self, scale: float = 1, cropbox: Optional[BoundingBox] = None

docling/datamodel/base_models.py

@@ -232,7 +232,6 @@ class Page(BaseModel):
     page_no: int
     # page_hash: Optional[str] = None
     size: Optional[Size] = None
-    cells: List[TextCell] = []
     parsed_page: Optional[SegmentedPdfPage] = None
     predictions: PagePredictions = PagePredictions()
     assembled: Optional[AssembledUnit] = None
@@ -245,6 +244,14 @@ class Page(BaseModel):
         float, Image
     ] = {}  # Cache of images in different scales. By default it is cleared during assembling.
 
+    @property
+    def cells(self) -> List[TextCell]:
+        """Return text cells as a read-only view of parsed_page.textline_cells."""
+        if self.parsed_page is not None:
+            return self.parsed_page.textline_cells
+        else:
+            return []
+
     def get_image(
         self, scale: float = 1.0, cropbox: Optional[BoundingBox] = None
     ) -> Optional[Image]:

docling/datamodel/pipeline_options.py

@@ -292,7 +292,9 @@ class PdfPipelineOptions(PaginatedPipelineOptions):
         ),
     )
 
-    generate_parsed_pages: bool = False
+    generate_parsed_pages: Literal[True] = (
+        True  # Always True since parsed_page is now mandatory
+    )
 
 
 class PdfPipeline(str, Enum):

docling/models/base_ocr_model.py

@@ -7,6 +7,7 @@ from typing import List, Optional, Type
 
 import numpy as np
 from docling_core.types.doc import BoundingBox, CoordOrigin
+from docling_core.types.doc.page import TextCell
 from PIL import Image, ImageDraw
 from rtree import index
 from scipy.ndimage import binary_dilation, find_objects, label
@@ -107,7 +108,9 @@ class BaseOcrModel(BasePageModel, BaseModelWithOptions):
             return []
 
     # Filters OCR cells by dropping any OCR cell that intersects with an existing programmatic cell.
-    def _filter_ocr_cells(self, ocr_cells, programmatic_cells):
+    def _filter_ocr_cells(
+        self, ocr_cells: List[TextCell], programmatic_cells: List[TextCell]
+    ) -> List[TextCell]:
         # Create R-tree index for programmatic cells
         p = index.Property()
         p.dimension = 2
@@ -130,19 +133,38 @@ class BaseOcrModel(BasePageModel, BaseModelWithOptions):
         ]
         return filtered_ocr_cells
 
-    def post_process_cells(self, ocr_cells, programmatic_cells):
+    def post_process_cells(self, ocr_cells: List[TextCell], page: Page) -> None:
         r"""
-        Post-process the ocr and programmatic cells and return the final list of of cells
+        Post-process the OCR cells and update the page object.
+        Updates parsed_page.textline_cells directly since page.cells is now read-only.
         """
-        if self.options.force_full_page_ocr:
+        # Get existing cells from the read-only property
-            # If a full page OCR is forced, use only the OCR cells
+        existing_cells = page.cells
-            cells = ocr_cells
-            return cells
 
-        ## Remove OCR cells which overlap with programmatic cells.
+        # Combine existing and OCR cells with overlap filtering
-        filtered_ocr_cells = self._filter_ocr_cells(ocr_cells, programmatic_cells)
+        final_cells = self._combine_cells(existing_cells, ocr_cells)
-        programmatic_cells.extend(filtered_ocr_cells)
+
-        return programmatic_cells
+        assert page.parsed_page is not None
+
+        # Update parsed_page.textline_cells directly
+        page.parsed_page.textline_cells = final_cells
+        page.parsed_page.has_lines = len(final_cells) > 0
+
+    def _combine_cells(
+        self, existing_cells: List[TextCell], ocr_cells: List[TextCell]
+    ) -> List[TextCell]:
+        """Combine existing and OCR cells with filtering and re-indexing."""
+        if self.options.force_full_page_ocr:
+            combined = ocr_cells
+        else:
+            filtered_ocr_cells = self._filter_ocr_cells(ocr_cells, existing_cells)
+            combined = list(existing_cells) + filtered_ocr_cells
+
+        # Re-index in-place
+        for i, cell in enumerate(combined):
+            cell.index = i
+
+        return combined
 
     def draw_ocr_rects_and_cells(self, conv_res, page, ocr_rects, show: bool = False):
         image = copy.deepcopy(page.image)

docling/models/easyocr_model.py

@@ -177,7 +177,7 @@ class EasyOcrModel(BaseOcrModel):
                         all_ocr_cells.extend(cells)
 
                     # Post-process the cells
-                    page.cells = self.post_process_cells(all_ocr_cells, page.cells)
+                    self.post_process_cells(all_ocr_cells, page)
 
                 # DEBUG code:
                 if settings.debug.visualize_ocr:

docling/models/layout_model.py

@@ -176,9 +176,9 @@ class LayoutModel(BasePageModel):
                     # Apply postprocessing
 
                     processed_clusters, processed_cells = LayoutPostprocessor(
-                        page.cells, clusters, page.size
+                        page, clusters
                     ).postprocess()
-                    # processed_clusters, processed_cells = clusters, page.cells
+                    # Note: LayoutPostprocessor updates page.cells and page.parsed_page internally
 
                     with warnings.catch_warnings():
                         warnings.filterwarnings(
@@ -198,7 +198,6 @@ class LayoutModel(BasePageModel):
                             )
                         )
 
-                    page.cells = processed_cells
                     page.predictions.layout = LayoutPrediction(
                         clusters=processed_clusters
                     )

docling/models/ocr_mac_model.py

@@ -132,7 +132,7 @@ class OcrMacModel(BaseOcrModel):
                         all_ocr_cells.extend(cells)
 
                     # Post-process the cells
-                    page.cells = self.post_process_cells(all_ocr_cells, page.cells)
+                    self.post_process_cells(all_ocr_cells, page)
 
                 # DEBUG code:
                 if settings.debug.visualize_ocr:

docling/models/page_preprocessing_model.py

@@ -2,7 +2,7 @@ import re
 import warnings
 from collections.abc import Iterable
 from pathlib import Path
-from typing import Optional
+from typing import Literal, Optional
 
 import numpy as np
 from PIL import ImageDraw
@@ -17,7 +17,6 @@ from docling.utils.profiling import TimeRecorder
 
 class PagePreprocessingOptions(BaseModel):
     images_scale: Optional[float]
-    create_parsed_page: bool
 
 
 class PagePreprocessingModel(BasePageModel):
@@ -66,10 +65,8 @@ class PagePreprocessingModel(BasePageModel):
     def _parse_page_cells(self, conv_res: ConversionResult, page: Page) -> Page:
         assert page._backend is not None
 
-        page.cells = list(page._backend.get_text_cells())
+        page.parsed_page = page._backend.get_segmented_page()
-
+        assert page.parsed_page is not None
-        if self.options.create_parsed_page:
-            page.parsed_page = page._backend.get_segmented_page()
 
         # Rate the text quality from the PDF parser, and aggregate on page
         text_scores = []

docling/models/rapid_ocr_model.py

@@ -134,7 +134,7 @@ class RapidOcrModel(BaseOcrModel):
                             all_ocr_cells.extend(cells)
 
                     # Post-process the cells
-                    page.cells = self.post_process_cells(all_ocr_cells, page.cells)
+                    self.post_process_cells(all_ocr_cells, page)
 
                 # DEBUG code:
                 if settings.debug.visualize_ocr:

docling/models/tesseract_ocr_cli_model.py

@@ -306,7 +306,7 @@ class TesseractOcrCliModel(BaseOcrModel):
                             all_ocr_cells.append(cell)
 
                     # Post-process the cells
-                    page.cells = self.post_process_cells(all_ocr_cells, page.cells)
+                    self.post_process_cells(all_ocr_cells, page)
 
                 # DEBUG code:
                 if settings.debug.visualize_ocr:

docling/models/tesseract_ocr_model.py

@@ -235,7 +235,7 @@ class TesseractOcrModel(BaseOcrModel):
                         all_ocr_cells.extend(cells)
 
                     # Post-process the cells
-                    page.cells = self.post_process_cells(all_ocr_cells, page.cells)
+                    self.post_process_cells(all_ocr_cells, page)
 
                 # DEBUG code:
                 if settings.debug.visualize_ocr:

docling/pipeline/standard_pdf_pipeline.py

@@ -72,7 +72,6 @@ class StandardPdfPipeline(PaginatedPipeline):
             PagePreprocessingModel(
                 options=PagePreprocessingOptions(
                     images_scale=pipeline_options.images_scale,
-                    create_parsed_page=pipeline_options.generate_parsed_pages,
                 )
             ),
             # OCR

docling/utils/layout_postprocessor.py

@@ -8,7 +8,7 @@ from docling_core.types.doc import DocItemLabel, Size
 from docling_core.types.doc.page import TextCell
 from rtree import index
 
-from docling.datamodel.base_models import BoundingBox, Cluster
+from docling.datamodel.base_models import BoundingBox, Cluster, Page
 
 _log = logging.getLogger(__name__)
 
@@ -194,11 +194,11 @@ class LayoutPostprocessor:
         DocItemLabel.TITLE: DocItemLabel.SECTION_HEADER,
     }
 
-    def __init__(self, cells: List[TextCell], clusters: List[Cluster], page_size: Size):
+    def __init__(self, page: Page, clusters: List[Cluster]) -> None:
-        """Initialize processor with cells and clusters."""
+        """Initialize processor with page and clusters."""
-        """Initialize processor with cells and spatial indices."""
+        self.cells = page.cells
-        self.cells = cells
+        self.page = page
-        self.page_size = page_size
+        self.page_size = page.size
         self.all_clusters = clusters
         self.regular_clusters = [
             c for c in clusters if c.label not in self.SPECIAL_TYPES
@@ -240,6 +240,10 @@ class LayoutPostprocessor:
             for child in cluster.children:
                 child.cells = self._sort_cells(child.cells)
 
+        assert self.page.parsed_page is not None
+        self.page.parsed_page.textline_cells = self.cells
+        self.page.parsed_page.has_lines = len(self.cells) > 0
+
         return final_clusters, self.cells
 
     def _process_regular_clusters(self) -> List[Cluster]:
@@ -301,6 +305,7 @@ class LayoutPostprocessor:
         special_clusters = self._handle_cross_type_overlaps(special_clusters)
 
         # Calculate page area from known page size
+        assert self.page_size is not None
         page_area = self.page_size.width * self.page_size.height
         if page_area > 0:
             # Filter out full-page pictures

tests/data/webp/groundtruth/docling_v2/webp-test.pages.json

@@ -5,84 +5,159 @@
       "width": 2000.0,
       "height": 2829.0
     },
-    "cells": [
+    "parsed_page": {
-      {
+      "dimension": {
-        "index": 0,
+        "angle": 0.0,
-        "rgba": {
-          "r": 0,
-          "g": 0,
-          "b": 0,
-          "a": 255
-        },
         "rect": {
-          "r_x0": 246.4065456254215,
+          "r_x0": 0.0,
-          "r_y0": 329.06770715202435,
+          "r_y0": 0.0,
-          "r_x1": 1691.991797818404,
+          "r_x1": 2000.0,
-          "r_y1": 329.06770715202435,
+          "r_y1": 0.0,
-          "r_x2": 1691.991797818404,
+          "r_x2": 2000.0,
-          "r_y2": 258.9040166758338,
+          "r_y2": 2829.0,
-          "r_x3": 246.4065456254215,
+          "r_x3": 0.0,
-          "r_y3": 258.9040166758338,
+          "r_y3": 2829.0,
-          "coord_origin": "TOPLEFT"
+          "coord_origin": "BOTTOMLEFT"
         },
-        "text": "Docling bundles PDF document conversion to",
+        "boundary_type": "crop_box",
-        "orig": "Docling bundles PDF document conversion to",
+        "art_bbox": {
-        "text_direction": "left_to_right",
+          "l": 0.0,
-        "confidence": 1.0,
+          "t": 2829.0,
-        "from_ocr": true
+          "r": 2000.0,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "bleed_bbox": {
+          "l": 0.0,
+          "t": 2829.0,
+          "r": 2000.0,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "crop_bbox": {
+          "l": 0.0,
+          "t": 2829.0,
+          "r": 2000.0,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "media_bbox": {
+          "l": 0.0,
+          "t": 2829.0,
+          "r": 2000.0,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "trim_bbox": {
+          "l": 0.0,
+          "t": 2829.0,
+          "r": 2000.0,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        }
       },
-      {
+      "bitmap_resources": [
-        "index": 1,
+        {
-        "rgba": {
+          "index": 0,
-          "r": 0,
+          "rect": {
-          "g": 0,
+            "r_x0": 0.0,
-          "b": 0,
+            "r_y0": 0.0,
-          "a": 255
+            "r_x1": 2000.0,
+            "r_y1": 0.0,
+            "r_x2": 2000.0,
+            "r_y2": 2829.0,
+            "r_x3": 0.0,
+            "r_y3": 2829.0,
+            "coord_origin": "BOTTOMLEFT"
+          },
+          "uri": null
+        }
+      ],
+      "char_cells": [],
+      "word_cells": [],
+      "textline_cells": [
+        {
+          "index": 0,
+          "rgba": {
+            "r": 0,
+            "g": 0,
+            "b": 0,
+            "a": 255
+          },
+          "rect": {
+            "r_x0": 246.4065456254215,
+            "r_y0": 329.06770715202435,
+            "r_x1": 1691.991797818404,
+            "r_y1": 329.06770715202435,
+            "r_x2": 1691.991797818404,
+            "r_y2": 258.9040166758338,
+            "r_x3": 246.4065456254215,
+            "r_y3": 258.9040166758338,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "Docling bundles PDF document conversion to",
+          "orig": "Docling bundles PDF document conversion to",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "rect": {
+        {
-          "r_x0": 234.08627147881114,
+          "index": 1,
-          "r_y0": 419.5788697734327,
+          "rgba": {
-          "r_x1": 1696.0985042090742,
+            "r": 0,
-          "r_y1": 419.5788697734327,
+            "g": 0,
-          "r_x2": 1696.0985042090742,
+            "b": 0,
-          "r_y2": 349.4151792972422,
+            "a": 255
-          "r_x3": 234.08627147881114,
+          },
-          "r_y3": 349.4151792972422,
+          "rect": {
-          "coord_origin": "TOPLEFT"
+            "r_x0": 234.08627147881114,
+            "r_y0": 419.5788697734327,
+            "r_x1": 1696.0985042090742,
+            "r_y1": 419.5788697734327,
+            "r_x2": 1696.0985042090742,
+            "r_y2": 349.4151792972422,
+            "r_x3": 234.08627147881114,
+            "r_y3": 349.4151792972422,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "JSON and Markdown in an easy self contained",
+          "orig": "JSON and Markdown in an easy self contained",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "text": "JSON and Markdown in an easy self contained",
+        {
-        "orig": "JSON and Markdown in an easy self contained",
+          "index": 2,
-        "text_direction": "left_to_right",
+          "rgba": {
-        "confidence": 1.0,
+            "r": 0,
-        "from_ocr": true
+            "g": 0,
-      },
+            "b": 0,
-      {
+            "a": 255
-        "index": 2,
+          },
-        "rgba": {
+          "rect": {
-          "r": 0,
+            "r_x0": 242.29979922858777,
-          "g": 0,
+            "r_y0": 509.8779072023336,
-          "b": 0,
+            "r_x1": 513.3470125989277,
-          "a": 255
+            "r_y1": 509.8779072023336,
-        },
+            "r_x2": 513.3470125989277,
-        "rect": {
+            "r_y2": 439.9752910477536,
-          "r_x0": 242.29979922858777,
+            "r_x3": 242.29979922858777,
-          "r_y0": 509.8779072023336,
+            "r_y3": 439.9752910477536,
-          "r_x1": 513.3470125989277,
+            "coord_origin": "TOPLEFT"
-          "r_y1": 509.8779072023336,
+          },
-          "r_x2": 513.3470125989277,
+          "text": "package",
-          "r_y2": 439.9752910477536,
+          "orig": "package",
-          "r_x3": 242.29979922858777,
+          "text_direction": "left_to_right",
-          "r_y3": 439.9752910477536,
+          "confidence": 1.0,
-          "coord_origin": "TOPLEFT"
+          "from_ocr": true
-        },
+        }
-        "text": "package",
+      ],
-        "orig": "package",
+      "has_chars": false,
-        "text_direction": "left_to_right",
+      "has_words": false,
-        "confidence": 1.0,
+      "has_lines": true,
-        "from_ocr": true
+      "image": null,
-      }
+      "lines": []
-    ],
+    },
-    "parsed_page": null,
     "predictions": {
       "layout": {
         "clusters": [

tests/data_scanned/groundtruth/docling_v1/ocr_test.pages.json

@@ -5,84 +5,143 @@
       "width": 595.201171875,
       "height": 841.9216918945312
     },
-    "cells": [
+    "parsed_page": {
-      {
+      "dimension": {
-        "index": 0,
+        "angle": 0.0,
-        "rgba": {
-          "r": 0,
-          "g": 0,
-          "b": 0,
-          "a": 255
-        },
         "rect": {
-          "r_x0": 73.34702132031646,
+          "r_x0": 0.0,
-          "r_y0": 97.99999977896755,
+          "r_y0": 0.0,
-          "r_x1": 503.64955224479564,
+          "r_x1": 595.201171875,
-          "r_y1": 97.99999977896755,
+          "r_y1": 0.0,
-          "r_x2": 503.64955224479564,
+          "r_x2": 595.201171875,
-          "r_y2": 76.99999977896756,
+          "r_y2": 841.9216918945312,
-          "r_x3": 73.34702132031646,
+          "r_x3": 0.0,
-          "r_y3": 76.99999977896756,
+          "r_y3": 841.9216918945312,
-          "coord_origin": "TOPLEFT"
+          "coord_origin": "BOTTOMLEFT"
         },
-        "text": "Docling bundles PDF document conversion to",
+        "boundary_type": "crop_box",
-        "orig": "Docling bundles PDF document conversion to",
+        "art_bbox": {
-        "text_direction": "left_to_right",
+          "l": 0.0,
-        "confidence": 1.0,
+          "t": 841.9216918945312,
-        "from_ocr": true
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "bleed_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "crop_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "media_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "trim_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        }
       },
-      {
+      "bitmap_resources": [],
-        "index": 1,
+      "char_cells": [],
-        "rgba": {
+      "word_cells": [],
-          "r": 0,
+      "textline_cells": [
-          "g": 0,
+        {
-          "b": 0,
+          "index": 0,
-          "a": 255
+          "rgba": {
+            "r": 0,
+            "g": 0,
+            "b": 0,
+            "a": 255
+          },
+          "rect": {
+            "r_x0": 73.34702132031646,
+            "r_y0": 97.99999977896755,
+            "r_x1": 503.64955224479564,
+            "r_y1": 97.99999977896755,
+            "r_x2": 503.64955224479564,
+            "r_y2": 76.99999977896756,
+            "r_x3": 73.34702132031646,
+            "r_y3": 76.99999977896756,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "Docling bundles PDF document conversion to",
+          "orig": "Docling bundles PDF document conversion to",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "rect": {
+        {
-          "r_x0": 69.6796630536824,
+          "index": 1,
-          "r_y0": 124.83139494707741,
+          "rgba": {
-          "r_x1": 504.8720051760782,
+            "r": 0,
-          "r_y1": 124.83139494707741,
+            "g": 0,
-          "r_x2": 504.8720051760782,
+            "b": 0,
-          "r_y2": 104.00000011573796,
+            "a": 255
-          "r_x3": 69.6796630536824,
+          },
-          "r_y3": 104.00000011573796,
+          "rect": {
-          "coord_origin": "TOPLEFT"
+            "r_x0": 69.6796630536824,
+            "r_y0": 124.83139494707741,
+            "r_x1": 504.8720051760782,
+            "r_y1": 124.83139494707741,
+            "r_x2": 504.8720051760782,
+            "r_y2": 104.00000011573796,
+            "r_x3": 69.6796630536824,
+            "r_y3": 104.00000011573796,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "JSON and Markdown in an easy self contained",
+          "orig": "JSON and Markdown in an easy self contained",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "text": "JSON and Markdown in an easy self contained",
+        {
-        "orig": "JSON and Markdown in an easy self contained",
+          "index": 2,
-        "text_direction": "left_to_right",
+          "rgba": {
-        "confidence": 1.0,
+            "r": 0,
-        "from_ocr": true
+            "g": 0,
-      },
+            "b": 0,
-      {
+            "a": 255
-        "index": 2,
+          },
-        "rgba": {
+          "rect": {
-          "r": 0,
+            "r_x0": 71.84193505100733,
-          "g": 0,
+            "r_y0": 152.90926970226084,
-          "b": 0,
+            "r_x1": 153.088934155825,
-          "a": 255
+            "r_y1": 152.90926970226084,
-        },
+            "r_x2": 153.088934155825,
-        "rect": {
+            "r_y2": 129.797125232046,
-          "r_x0": 71.84193505100733,
+            "r_x3": 71.84193505100733,
-          "r_y0": 152.90926970226084,
+            "r_y3": 129.797125232046,
-          "r_x1": 153.088934155825,
+            "coord_origin": "TOPLEFT"
-          "r_y1": 152.90926970226084,
+          },
-          "r_x2": 153.088934155825,
+          "text": "package",
-          "r_y2": 129.797125232046,
+          "orig": "package",
-          "r_x3": 71.84193505100733,
+          "text_direction": "left_to_right",
-          "r_y3": 129.797125232046,
+          "confidence": 1.0,
-          "coord_origin": "TOPLEFT"
+          "from_ocr": true
-        },
+        }
-        "text": "package",
+      ],
-        "orig": "package",
+      "has_chars": false,
-        "text_direction": "left_to_right",
+      "has_words": false,
-        "confidence": 1.0,
+      "has_lines": true,
-        "from_ocr": true
+      "image": null,
-      }
+      "lines": []
-    ],
+    },
-    "parsed_page": null,
     "predictions": {
       "layout": {
         "clusters": [

tests/data_scanned/groundtruth/docling_v1/ocr_test_rotated_180.pages.json

@@ -5,84 +5,143 @@
       "width": 595.201171875,
       "height": 841.9216918945312
     },
-    "cells": [
+    "parsed_page": {
-      {
+      "dimension": {
-        "index": 0,
+        "angle": 0.0,
-        "rgba": {
-          "r": 0,
-          "g": 0,
-          "b": 0,
-          "a": 255
-        },
         "rect": {
-          "r_x0": 89.2388782764286,
+          "r_x0": 0.0,
-          "r_y0": 764.898293373551,
+          "r_y0": 0.0,
-          "r_x1": 521.9863147998661,
+          "r_x1": 595.201171875,
-          "r_y1": 764.898293373551,
+          "r_y1": 0.0,
-          "r_x2": 521.9863147998661,
+          "r_x2": 595.201171875,
-          "r_y2": 744.0929853494625,
+          "r_y2": 841.9216918945312,
-          "r_x3": 89.2388782764286,
+          "r_x3": 0.0,
-          "r_y3": 744.0929853494625,
+          "r_y3": 841.9216918945312,
-          "coord_origin": "TOPLEFT"
+          "coord_origin": "BOTTOMLEFT"
         },
-        "text": "Docling bundles PDF document conversion to",
+        "boundary_type": "crop_box",
-        "orig": "Docling bundles PDF document conversion to",
+        "art_bbox": {
-        "text_direction": "left_to_right",
+          "l": 0.0,
-        "confidence": 1.0,
+          "t": 841.9216918945312,
-        "from_ocr": true
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "bleed_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "crop_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "media_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "trim_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        }
       },
-      {
+      "bitmap_resources": [],
-        "index": 1,
+      "char_cells": [],
-        "rgba": {
+      "word_cells": [],
-          "r": 0,
+      "textline_cells": [
-          "g": 0,
+        {
-          "b": 0,
+          "index": 0,
-          "a": 255
+          "rgba": {
+            "r": 0,
+            "g": 0,
+            "b": 0,
+            "a": 255
+          },
+          "rect": {
+            "r_x0": 89.2388782764286,
+            "r_y0": 764.898293373551,
+            "r_x1": 521.9863147998661,
+            "r_y1": 764.898293373551,
+            "r_x2": 521.9863147998661,
+            "r_y2": 744.0929853494625,
+            "r_x3": 89.2388782764286,
+            "r_y3": 744.0929853494625,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "Docling bundles PDF document conversion to",
+          "orig": "Docling bundles PDF document conversion to",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "rect": {
+        {
-          "r_x0": 89.23887497045128,
+          "index": 1,
-          "r_y0": 739.1977118987292,
+          "rgba": {
-          "r_x1": 523.208764293368,
+            "r": 0,
-          "r_y1": 739.1977118987292,
+            "g": 0,
-          "r_x2": 523.208764293368,
+            "b": 0,
-          "r_y2": 717.1685676116198,
+            "a": 255
-          "r_x3": 89.23887497045128,
+          },
-          "r_y3": 717.1685676116198,
+          "rect": {
-          "coord_origin": "TOPLEFT"
+            "r_x0": 89.23887497045128,
+            "r_y0": 739.1977118987292,
+            "r_x1": 523.208764293368,
+            "r_y1": 739.1977118987292,
+            "r_x2": 523.208764293368,
+            "r_y2": 717.1685676116198,
+            "r_x3": 89.23887497045128,
+            "r_y3": 717.1685676116198,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "JSON and Markdown in an easy self contained",
+          "orig": "JSON and Markdown in an easy self contained",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "text": "JSON and Markdown in an easy self contained",
+        {
-        "orig": "JSON and Markdown in an easy self contained",
+          "index": 2,
-        "text_direction": "left_to_right",
+          "rgba": {
-        "confidence": 1.0,
+            "r": 0,
-        "from_ocr": true
+            "g": 0,
-      },
+            "b": 0,
-      {
+            "a": 255
-        "index": 2,
+          },
-        "rgba": {
+          "rect": {
-          "r": 0,
+            "r_x0": 441.2561096985719,
-          "g": 0,
+            "r_y0": 710.0268078458798,
-          "b": 0,
+            "r_x1": 522.0347860494834,
-          "a": 255
+            "r_y1": 710.0268078458798,
-        },
+            "r_x2": 522.0347860494834,
-        "rect": {
+            "r_y2": 690.0429592741025,
-          "r_x0": 441.2561096985719,
+            "r_x3": 441.2561096985719,
-          "r_y0": 710.0268078458798,
+            "r_y3": 690.0429592741025,
-          "r_x1": 522.0347860494834,
+            "coord_origin": "TOPLEFT"
-          "r_y1": 710.0268078458798,
+          },
-          "r_x2": 522.0347860494834,
+          "text": "package",
-          "r_y2": 690.0429592741025,
+          "orig": "package",
-          "r_x3": 441.2561096985719,
+          "text_direction": "left_to_right",
-          "r_y3": 690.0429592741025,
+          "confidence": 1.0,
-          "coord_origin": "TOPLEFT"
+          "from_ocr": true
-        },
+        }
-        "text": "package",
+      ],
-        "orig": "package",
+      "has_chars": false,
-        "text_direction": "left_to_right",
+      "has_words": false,
-        "confidence": 1.0,
+      "has_lines": true,
-        "from_ocr": true
+      "image": null,
-      }
+      "lines": []
-    ],
+    },
-    "parsed_page": null,
     "predictions": {
       "layout": {
         "clusters": [

tests/data_scanned/groundtruth/docling_v1/ocr_test_rotated_270.pages.json

@@ -5,84 +5,143 @@
       "width": 841.9216918945312,
       "height": 595.201171875
     },
-    "cells": [
+    "parsed_page": {
-      {
+      "dimension": {
-        "index": 0,
+        "angle": 0.0,
-        "rgba": {
-          "r": 0,
-          "g": 0,
-          "b": 0,
-          "a": 255
-        },
         "rect": {
-          "r_x0": 744.0930045534915,
+          "r_x0": 0.0,
-          "r_y0": 504.87200373583954,
+          "r_y0": 0.0,
-          "r_x1": 764.8982839673505,
+          "r_x1": 595.201171875,
-          "r_y1": 504.87200373583954,
+          "r_y1": 0.0,
-          "r_x2": 764.8982839673505,
+          "r_x2": 595.201171875,
-          "r_y2": 73.34702001188118,
+          "r_y2": 841.9216918945312,
-          "r_x3": 744.0930045534915,
+          "r_x3": 0.0,
-          "r_y3": 73.34702001188118,
+          "r_y3": 841.9216918945312,
-          "coord_origin": "TOPLEFT"
+          "coord_origin": "BOTTOMLEFT"
         },
-        "text": "Docling bundles PDF document conversion to",
+        "boundary_type": "crop_box",
-        "orig": "Docling bundles PDF document conversion to",
+        "art_bbox": {
-        "text_direction": "left_to_right",
+          "l": 0.0,
-        "confidence": 1.0,
+          "t": 841.9216918945312,
-        "from_ocr": true
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "bleed_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "crop_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "media_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "trim_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        }
       },
-      {
+      "bitmap_resources": [],
-        "index": 1,
+      "char_cells": [],
-        "rgba": {
+      "word_cells": [],
-          "r": 0,
+      "textline_cells": [
-          "g": 0,
+        {
-          "b": 0,
+          "index": 0,
-          "a": 255
+          "rgba": {
+            "r": 0,
+            "g": 0,
+            "b": 0,
+            "a": 255
+          },
+          "rect": {
+            "r_x0": 744.0930045534915,
+            "r_y0": 504.87200373583954,
+            "r_x1": 764.8982839673505,
+            "r_y1": 504.87200373583954,
+            "r_x2": 764.8982839673505,
+            "r_y2": 73.34702001188118,
+            "r_x3": 744.0930045534915,
+            "r_y3": 73.34702001188118,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "Docling bundles PDF document conversion to",
+          "orig": "Docling bundles PDF document conversion to",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "rect": {
+        {
-          "r_x0": 717.168585936602,
+          "index": 1,
-          "r_y0": 504.8720061466397,
+          "rgba": {
-          "r_x1": 737.9738558137178,
+            "r": 0,
-          "r_y1": 504.8720061466397,
+            "g": 0,
-          "r_x2": 737.9738558137178,
+            "b": 0,
-          "r_y2": 70.90211682372312,
+            "a": 255
-          "r_x3": 717.168585936602,
+          },
-          "r_y3": 70.90211682372312,
+          "rect": {
-          "coord_origin": "TOPLEFT"
+            "r_x0": 717.168585936602,
+            "r_y0": 504.8720061466397,
+            "r_x1": 737.9738558137178,
+            "r_y1": 504.8720061466397,
+            "r_x2": 737.9738558137178,
+            "r_y2": 70.90211682372312,
+            "r_x3": 717.168585936602,
+            "r_y3": 70.90211682372312,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "JSON and Markdown in an easy self contained",
+          "orig": "JSON and Markdown in an easy self contained",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "text": "JSON and Markdown in an easy self contained",
+        {
-        "orig": "JSON and Markdown in an easy self contained",
+          "index": 2,
-        "text_direction": "left_to_right",
+          "rgba": {
-        "confidence": 1.0,
+            "r": 0,
-        "from_ocr": true
+            "g": 0,
-      },
+            "b": 0,
-      {
+            "a": 255
-        "index": 2,
+          },
-        "rgba": {
+          "rect": {
-          "r": 0,
+            "r_x0": 690.2441821046808,
-          "g": 0,
+            "r_y0": 152.80629773131633,
-          "b": 0,
+            "r_x1": 709.8255852011977,
-          "a": 255
+            "r_y1": 152.80629773131633,
-        },
+            "r_x2": 709.8255852011977,
-        "rect": {
+            "r_y2": 72.124570639845,
-          "r_x0": 690.2441821046808,
+            "r_x3": 690.2441821046808,
-          "r_y0": 152.80629773131633,
+            "r_y3": 72.124570639845,
-          "r_x1": 709.8255852011977,
+            "coord_origin": "TOPLEFT"
-          "r_y1": 152.80629773131633,
+          },
-          "r_x2": 709.8255852011977,
+          "text": "package",
-          "r_y2": 72.124570639845,
+          "orig": "package",
-          "r_x3": 690.2441821046808,
+          "text_direction": "left_to_right",
-          "r_y3": 72.124570639845,
+          "confidence": 1.0,
-          "coord_origin": "TOPLEFT"
+          "from_ocr": true
-        },
+        }
-        "text": "package",
+      ],
-        "orig": "package",
+      "has_chars": false,
-        "text_direction": "left_to_right",
+      "has_words": false,
-        "confidence": 1.0,
+      "has_lines": true,
-        "from_ocr": true
+      "image": null,
-      }
+      "lines": []
-    ],
+    },
-    "parsed_page": null,
     "predictions": {
       "layout": {
         "clusters": [

tests/data_scanned/groundtruth/docling_v1/ocr_test_rotated_90.pages.json

@@ -5,84 +5,143 @@
       "width": 841.9216918945312,
       "height": 595.201171875
     },
-    "cells": [
+    "parsed_page": {
-      {
+      "dimension": {
-        "index": 0,
+        "angle": 0.0,
-        "rgba": {
-          "r": 0,
-          "g": 0,
-          "b": 0,
-          "a": 255
-        },
         "rect": {
-          "r_x0": 77.10171545548258,
+          "r_x0": 0.0,
-          "r_y0": 520.7638571913312,
+          "r_y0": 0.0,
-          "r_x1": 96.68315797053792,
+          "r_x1": 595.201171875,
-          "r_y1": 520.7638571913312,
+          "r_y1": 0.0,
-          "r_x2": 96.68315797053792,
+          "r_x2": 595.201171875,
-          "r_y2": 89.2388734673729,
+          "r_y2": 841.9216918945312,
-          "r_x3": 77.10171545548258,
+          "r_x3": 0.0,
-          "r_y3": 89.2388734673729,
+          "r_y3": 841.9216918945312,
-          "coord_origin": "TOPLEFT"
+          "coord_origin": "BOTTOMLEFT"
         },
-        "text": "Docling bundles PDF document conversion to",
+        "boundary_type": "crop_box",
-        "orig": "Docling bundles PDF document conversion to",
+        "art_bbox": {
-        "text_direction": "left_to_right",
+          "l": 0.0,
-        "confidence": 1.0,
+          "t": 841.9216918945312,
-        "from_ocr": true
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "bleed_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "crop_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "media_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "trim_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        }
       },
-      {
+      "bitmap_resources": [],
-        "index": 1,
+      "char_cells": [],
-        "rgba": {
+      "word_cells": [],
-          "r": 0,
+      "textline_cells": [
-          "g": 0,
+        {
-          "b": 0,
+          "index": 0,
-          "a": 255
+          "rgba": {
+            "r": 0,
+            "g": 0,
+            "b": 0,
+            "a": 255
+          },
+          "rect": {
+            "r_x0": 77.10171545548258,
+            "r_y0": 520.7638571913312,
+            "r_x1": 96.68315797053792,
+            "r_y1": 520.7638571913312,
+            "r_x2": 96.68315797053792,
+            "r_y2": 89.2388734673729,
+            "r_x3": 77.10171545548258,
+            "r_y3": 89.2388734673729,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "Docling bundles PDF document conversion to",
+          "orig": "Docling bundles PDF document conversion to",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "rect": {
+        {
-          "r_x0": 100.64168123325977,
+          "index": 1,
-          "r_y0": 523.3236155182395,
+          "rgba": {
-          "r_x1": 126.08064862014129,
+            "r": 0,
-          "r_y1": 523.3236155182395,
+            "g": 0,
-          "r_x2": 126.08064862014129,
+            "b": 0,
-          "r_y2": 89.1266754140729,
+            "a": 255
-          "r_x3": 100.64168123325977,
+          },
-          "r_y3": 89.1266754140729,
+          "rect": {
-          "coord_origin": "TOPLEFT"
+            "r_x0": 100.64168123325977,
+            "r_y0": 523.3236155182395,
+            "r_x1": 126.08064862014129,
+            "r_y1": 523.3236155182395,
+            "r_x2": 126.08064862014129,
+            "r_y2": 89.1266754140729,
+            "r_x3": 100.64168123325977,
+            "r_y3": 89.1266754140729,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "JSON and Markdown in an easy self contained",
+          "orig": "JSON and Markdown in an easy self contained",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "text": "JSON and Markdown in an easy self contained",
+        {
-        "orig": "JSON and Markdown in an easy self contained",
+          "index": 2,
-        "text_direction": "left_to_right",
+          "rgba": {
-        "confidence": 1.0,
+            "r": 0,
-        "from_ocr": true
+            "g": 0,
-      },
+            "b": 0,
-      {
+            "a": 255
-        "index": 2,
+          },
-        "rgba": {
+          "rect": {
-          "r": 0,
+            "r_x0": 131.21306574279092,
-          "g": 0,
+            "r_y0": 521.0762158417759,
-          "b": 0,
+            "r_x1": 152.19606490864376,
-          "a": 255
+            "r_y1": 521.0762158417759,
-        },
+            "r_x2": 152.19606490864376,
-        "rect": {
+            "r_y2": 441.0071698212682,
-          "r_x0": 131.21306574279092,
+            "r_x3": 131.21306574279092,
-          "r_y0": 521.0762158417759,
+            "r_y3": 441.0071698212682,
-          "r_x1": 152.19606490864376,
+            "coord_origin": "TOPLEFT"
-          "r_y1": 521.0762158417759,
+          },
-          "r_x2": 152.19606490864376,
+          "text": "package",
-          "r_y2": 441.0071698212682,
+          "orig": "package",
-          "r_x3": 131.21306574279092,
+          "text_direction": "left_to_right",
-          "r_y3": 441.0071698212682,
+          "confidence": 1.0,
-          "coord_origin": "TOPLEFT"
+          "from_ocr": true
-        },
+        }
-        "text": "package",
+      ],
-        "orig": "package",
+      "has_chars": false,
-        "text_direction": "left_to_right",
+      "has_words": false,
-        "confidence": 1.0,
+      "has_lines": true,
-        "from_ocr": true
+      "image": null,
-      }
+      "lines": []
-    ],
+    },
-    "parsed_page": null,
     "predictions": {
       "layout": {
         "clusters": [

tests/data_scanned/groundtruth/docling_v2/ocr_test.pages.json

@@ -5,84 +5,143 @@
       "width": 595.201171875,
       "height": 841.9216918945312
     },
-    "cells": [
+    "parsed_page": {
-      {
+      "dimension": {
-        "index": 0,
+        "angle": 0.0,
-        "rgba": {
-          "r": 0,
-          "g": 0,
-          "b": 0,
-          "a": 255
-        },
         "rect": {
-          "r_x0": 73.34702132031646,
+          "r_x0": 0.0,
-          "r_y0": 97.99999977896755,
+          "r_y0": 0.0,
-          "r_x1": 503.64955224479564,
+          "r_x1": 595.201171875,
-          "r_y1": 97.99999977896755,
+          "r_y1": 0.0,
-          "r_x2": 503.64955224479564,
+          "r_x2": 595.201171875,
-          "r_y2": 76.99999977896756,
+          "r_y2": 841.9216918945312,
-          "r_x3": 73.34702132031646,
+          "r_x3": 0.0,
-          "r_y3": 76.99999977896756,
+          "r_y3": 841.9216918945312,
-          "coord_origin": "TOPLEFT"
+          "coord_origin": "BOTTOMLEFT"
         },
-        "text": "Docling bundles PDF document conversion to",
+        "boundary_type": "crop_box",
-        "orig": "Docling bundles PDF document conversion to",
+        "art_bbox": {
-        "text_direction": "left_to_right",
+          "l": 0.0,
-        "confidence": 1.0,
+          "t": 841.9216918945312,
-        "from_ocr": true
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "bleed_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "crop_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "media_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "trim_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        }
       },
-      {
+      "bitmap_resources": [],
-        "index": 1,
+      "char_cells": [],
-        "rgba": {
+      "word_cells": [],
-          "r": 0,
+      "textline_cells": [
-          "g": 0,
+        {
-          "b": 0,
+          "index": 0,
-          "a": 255
+          "rgba": {
+            "r": 0,
+            "g": 0,
+            "b": 0,
+            "a": 255
+          },
+          "rect": {
+            "r_x0": 73.34702132031646,
+            "r_y0": 97.99999977896755,
+            "r_x1": 503.64955224479564,
+            "r_y1": 97.99999977896755,
+            "r_x2": 503.64955224479564,
+            "r_y2": 76.99999977896756,
+            "r_x3": 73.34702132031646,
+            "r_y3": 76.99999977896756,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "Docling bundles PDF document conversion to",
+          "orig": "Docling bundles PDF document conversion to",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "rect": {
+        {
-          "r_x0": 69.6796630536824,
+          "index": 1,
-          "r_y0": 124.83139494707741,
+          "rgba": {
-          "r_x1": 504.8720051760782,
+            "r": 0,
-          "r_y1": 124.83139494707741,
+            "g": 0,
-          "r_x2": 504.8720051760782,
+            "b": 0,
-          "r_y2": 104.00000011573796,
+            "a": 255
-          "r_x3": 69.6796630536824,
+          },
-          "r_y3": 104.00000011573796,
+          "rect": {
-          "coord_origin": "TOPLEFT"
+            "r_x0": 69.6796630536824,
+            "r_y0": 124.83139494707741,
+            "r_x1": 504.8720051760782,
+            "r_y1": 124.83139494707741,
+            "r_x2": 504.8720051760782,
+            "r_y2": 104.00000011573796,
+            "r_x3": 69.6796630536824,
+            "r_y3": 104.00000011573796,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "JSON and Markdown in an easy self contained",
+          "orig": "JSON and Markdown in an easy self contained",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "text": "JSON and Markdown in an easy self contained",
+        {
-        "orig": "JSON and Markdown in an easy self contained",
+          "index": 2,
-        "text_direction": "left_to_right",
+          "rgba": {
-        "confidence": 1.0,
+            "r": 0,
-        "from_ocr": true
+            "g": 0,
-      },
+            "b": 0,
-      {
+            "a": 255
-        "index": 2,
+          },
-        "rgba": {
+          "rect": {
-          "r": 0,
+            "r_x0": 71.84193505100733,
-          "g": 0,
+            "r_y0": 152.90926970226084,
-          "b": 0,
+            "r_x1": 153.088934155825,
-          "a": 255
+            "r_y1": 152.90926970226084,
-        },
+            "r_x2": 153.088934155825,
-        "rect": {
+            "r_y2": 129.797125232046,
-          "r_x0": 71.84193505100733,
+            "r_x3": 71.84193505100733,
-          "r_y0": 152.90926970226084,
+            "r_y3": 129.797125232046,
-          "r_x1": 153.088934155825,
+            "coord_origin": "TOPLEFT"
-          "r_y1": 152.90926970226084,
+          },
-          "r_x2": 153.088934155825,
+          "text": "package",
-          "r_y2": 129.797125232046,
+          "orig": "package",
-          "r_x3": 71.84193505100733,
+          "text_direction": "left_to_right",
-          "r_y3": 129.797125232046,
+          "confidence": 1.0,
-          "coord_origin": "TOPLEFT"
+          "from_ocr": true
-        },
+        }
-        "text": "package",
+      ],
-        "orig": "package",
+      "has_chars": false,
-        "text_direction": "left_to_right",
+      "has_words": false,
-        "confidence": 1.0,
+      "has_lines": true,
-        "from_ocr": true
+      "image": null,
-      }
+      "lines": []
-    ],
+    },
-    "parsed_page": null,
     "predictions": {
       "layout": {
         "clusters": [

tests/data_scanned/groundtruth/docling_v2/ocr_test_rotated_180.pages.json

@@ -5,84 +5,143 @@
       "width": 595.201171875,
       "height": 841.9216918945312
     },
-    "cells": [
+    "parsed_page": {
-      {
+      "dimension": {
-        "index": 0,
+        "angle": 0.0,
-        "rgba": {
-          "r": 0,
-          "g": 0,
-          "b": 0,
-          "a": 255
-        },
         "rect": {
-          "r_x0": 89.2388782764286,
+          "r_x0": 0.0,
-          "r_y0": 764.898293373551,
+          "r_y0": 0.0,
-          "r_x1": 521.9863147998661,
+          "r_x1": 595.201171875,
-          "r_y1": 764.898293373551,
+          "r_y1": 0.0,
-          "r_x2": 521.9863147998661,
+          "r_x2": 595.201171875,
-          "r_y2": 744.0929853494625,
+          "r_y2": 841.9216918945312,
-          "r_x3": 89.2388782764286,
+          "r_x3": 0.0,
-          "r_y3": 744.0929853494625,
+          "r_y3": 841.9216918945312,
-          "coord_origin": "TOPLEFT"
+          "coord_origin": "BOTTOMLEFT"
         },
-        "text": "Docling bundles PDF document conversion to",
+        "boundary_type": "crop_box",
-        "orig": "Docling bundles PDF document conversion to",
+        "art_bbox": {
-        "text_direction": "left_to_right",
+          "l": 0.0,
-        "confidence": 1.0,
+          "t": 841.9216918945312,
-        "from_ocr": true
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "bleed_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "crop_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "media_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "trim_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        }
       },
-      {
+      "bitmap_resources": [],
-        "index": 1,
+      "char_cells": [],
-        "rgba": {
+      "word_cells": [],
-          "r": 0,
+      "textline_cells": [
-          "g": 0,
+        {
-          "b": 0,
+          "index": 0,
-          "a": 255
+          "rgba": {
+            "r": 0,
+            "g": 0,
+            "b": 0,
+            "a": 255
+          },
+          "rect": {
+            "r_x0": 89.2388782764286,
+            "r_y0": 764.898293373551,
+            "r_x1": 521.9863147998661,
+            "r_y1": 764.898293373551,
+            "r_x2": 521.9863147998661,
+            "r_y2": 744.0929853494625,
+            "r_x3": 89.2388782764286,
+            "r_y3": 744.0929853494625,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "Docling bundles PDF document conversion to",
+          "orig": "Docling bundles PDF document conversion to",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "rect": {
+        {
-          "r_x0": 89.23887497045128,
+          "index": 1,
-          "r_y0": 739.1977118987292,
+          "rgba": {
-          "r_x1": 523.208764293368,
+            "r": 0,
-          "r_y1": 739.1977118987292,
+            "g": 0,
-          "r_x2": 523.208764293368,
+            "b": 0,
-          "r_y2": 717.1685676116198,
+            "a": 255
-          "r_x3": 89.23887497045128,
+          },
-          "r_y3": 717.1685676116198,
+          "rect": {
-          "coord_origin": "TOPLEFT"
+            "r_x0": 89.23887497045128,
+            "r_y0": 739.1977118987292,
+            "r_x1": 523.208764293368,
+            "r_y1": 739.1977118987292,
+            "r_x2": 523.208764293368,
+            "r_y2": 717.1685676116198,
+            "r_x3": 89.23887497045128,
+            "r_y3": 717.1685676116198,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "JSON and Markdown in an easy self contained",
+          "orig": "JSON and Markdown in an easy self contained",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "text": "JSON and Markdown in an easy self contained",
+        {
-        "orig": "JSON and Markdown in an easy self contained",
+          "index": 2,
-        "text_direction": "left_to_right",
+          "rgba": {
-        "confidence": 1.0,
+            "r": 0,
-        "from_ocr": true
+            "g": 0,
-      },
+            "b": 0,
-      {
+            "a": 255
-        "index": 2,
+          },
-        "rgba": {
+          "rect": {
-          "r": 0,
+            "r_x0": 441.2561096985719,
-          "g": 0,
+            "r_y0": 710.0268078458798,
-          "b": 0,
+            "r_x1": 522.0347860494834,
-          "a": 255
+            "r_y1": 710.0268078458798,
-        },
+            "r_x2": 522.0347860494834,
-        "rect": {
+            "r_y2": 690.0429592741025,
-          "r_x0": 441.2561096985719,
+            "r_x3": 441.2561096985719,
-          "r_y0": 710.0268078458798,
+            "r_y3": 690.0429592741025,
-          "r_x1": 522.0347860494834,
+            "coord_origin": "TOPLEFT"
-          "r_y1": 710.0268078458798,
+          },
-          "r_x2": 522.0347860494834,
+          "text": "package",
-          "r_y2": 690.0429592741025,
+          "orig": "package",
-          "r_x3": 441.2561096985719,
+          "text_direction": "left_to_right",
-          "r_y3": 690.0429592741025,
+          "confidence": 1.0,
-          "coord_origin": "TOPLEFT"
+          "from_ocr": true
-        },
+        }
-        "text": "package",
+      ],
-        "orig": "package",
+      "has_chars": false,
-        "text_direction": "left_to_right",
+      "has_words": false,
-        "confidence": 1.0,
+      "has_lines": true,
-        "from_ocr": true
+      "image": null,
-      }
+      "lines": []
-    ],
+    },
-    "parsed_page": null,
     "predictions": {
       "layout": {
         "clusters": [

tests/data_scanned/groundtruth/docling_v2/ocr_test_rotated_270.pages.json

@@ -5,84 +5,143 @@
       "width": 841.9216918945312,
       "height": 595.201171875
     },
-    "cells": [
+    "parsed_page": {
-      {
+      "dimension": {
-        "index": 0,
+        "angle": 0.0,
-        "rgba": {
-          "r": 0,
-          "g": 0,
-          "b": 0,
-          "a": 255
-        },
         "rect": {
-          "r_x0": 744.0930045534915,
+          "r_x0": 0.0,
-          "r_y0": 504.87200373583954,
+          "r_y0": 0.0,
-          "r_x1": 764.8982839673505,
+          "r_x1": 595.201171875,
-          "r_y1": 504.87200373583954,
+          "r_y1": 0.0,
-          "r_x2": 764.8982839673505,
+          "r_x2": 595.201171875,
-          "r_y2": 73.34702001188118,
+          "r_y2": 841.9216918945312,
-          "r_x3": 744.0930045534915,
+          "r_x3": 0.0,
-          "r_y3": 73.34702001188118,
+          "r_y3": 841.9216918945312,
-          "coord_origin": "TOPLEFT"
+          "coord_origin": "BOTTOMLEFT"
         },
-        "text": "Docling bundles PDF document conversion to",
+        "boundary_type": "crop_box",
-        "orig": "Docling bundles PDF document conversion to",
+        "art_bbox": {
-        "text_direction": "left_to_right",
+          "l": 0.0,
-        "confidence": 1.0,
+          "t": 841.9216918945312,
-        "from_ocr": true
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "bleed_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "crop_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "media_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "trim_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        }
       },
-      {
+      "bitmap_resources": [],
-        "index": 1,
+      "char_cells": [],
-        "rgba": {
+      "word_cells": [],
-          "r": 0,
+      "textline_cells": [
-          "g": 0,
+        {
-          "b": 0,
+          "index": 0,
-          "a": 255
+          "rgba": {
+            "r": 0,
+            "g": 0,
+            "b": 0,
+            "a": 255
+          },
+          "rect": {
+            "r_x0": 744.0930045534915,
+            "r_y0": 504.87200373583954,
+            "r_x1": 764.8982839673505,
+            "r_y1": 504.87200373583954,
+            "r_x2": 764.8982839673505,
+            "r_y2": 73.34702001188118,
+            "r_x3": 744.0930045534915,
+            "r_y3": 73.34702001188118,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "Docling bundles PDF document conversion to",
+          "orig": "Docling bundles PDF document conversion to",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "rect": {
+        {
-          "r_x0": 717.168585936602,
+          "index": 1,
-          "r_y0": 504.8720061466397,
+          "rgba": {
-          "r_x1": 737.9738558137178,
+            "r": 0,
-          "r_y1": 504.8720061466397,
+            "g": 0,
-          "r_x2": 737.9738558137178,
+            "b": 0,
-          "r_y2": 70.90211682372312,
+            "a": 255
-          "r_x3": 717.168585936602,
+          },
-          "r_y3": 70.90211682372312,
+          "rect": {
-          "coord_origin": "TOPLEFT"
+            "r_x0": 717.168585936602,
+            "r_y0": 504.8720061466397,
+            "r_x1": 737.9738558137178,
+            "r_y1": 504.8720061466397,
+            "r_x2": 737.9738558137178,
+            "r_y2": 70.90211682372312,
+            "r_x3": 717.168585936602,
+            "r_y3": 70.90211682372312,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "JSON and Markdown in an easy self contained",
+          "orig": "JSON and Markdown in an easy self contained",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "text": "JSON and Markdown in an easy self contained",
+        {
-        "orig": "JSON and Markdown in an easy self contained",
+          "index": 2,
-        "text_direction": "left_to_right",
+          "rgba": {
-        "confidence": 1.0,
+            "r": 0,
-        "from_ocr": true
+            "g": 0,
-      },
+            "b": 0,
-      {
+            "a": 255
-        "index": 2,
+          },
-        "rgba": {
+          "rect": {
-          "r": 0,
+            "r_x0": 690.2441821046808,
-          "g": 0,
+            "r_y0": 152.80629773131633,
-          "b": 0,
+            "r_x1": 709.8255852011977,
-          "a": 255
+            "r_y1": 152.80629773131633,
-        },
+            "r_x2": 709.8255852011977,
-        "rect": {
+            "r_y2": 72.124570639845,
-          "r_x0": 690.2441821046808,
+            "r_x3": 690.2441821046808,
-          "r_y0": 152.80629773131633,
+            "r_y3": 72.124570639845,
-          "r_x1": 709.8255852011977,
+            "coord_origin": "TOPLEFT"
-          "r_y1": 152.80629773131633,
+          },
-          "r_x2": 709.8255852011977,
+          "text": "package",
-          "r_y2": 72.124570639845,
+          "orig": "package",
-          "r_x3": 690.2441821046808,
+          "text_direction": "left_to_right",
-          "r_y3": 72.124570639845,
+          "confidence": 1.0,
-          "coord_origin": "TOPLEFT"
+          "from_ocr": true
-        },
+        }
-        "text": "package",
+      ],
-        "orig": "package",
+      "has_chars": false,
-        "text_direction": "left_to_right",
+      "has_words": false,
-        "confidence": 1.0,
+      "has_lines": true,
-        "from_ocr": true
+      "image": null,
-      }
+      "lines": []
-    ],
+    },
-    "parsed_page": null,
     "predictions": {
       "layout": {
         "clusters": [

tests/data_scanned/groundtruth/docling_v2/ocr_test_rotated_90.pages.json

@@ -5,84 +5,143 @@
       "width": 841.9216918945312,
       "height": 595.201171875
     },
-    "cells": [
+    "parsed_page": {
-      {
+      "dimension": {
-        "index": 0,
+        "angle": 0.0,
-        "rgba": {
-          "r": 0,
-          "g": 0,
-          "b": 0,
-          "a": 255
-        },
         "rect": {
-          "r_x0": 77.10171545548258,
+          "r_x0": 0.0,
-          "r_y0": 520.7638571913312,
+          "r_y0": 0.0,
-          "r_x1": 96.68315797053792,
+          "r_x1": 595.201171875,
-          "r_y1": 520.7638571913312,
+          "r_y1": 0.0,
-          "r_x2": 96.68315797053792,
+          "r_x2": 595.201171875,
-          "r_y2": 89.2388734673729,
+          "r_y2": 841.9216918945312,
-          "r_x3": 77.10171545548258,
+          "r_x3": 0.0,
-          "r_y3": 89.2388734673729,
+          "r_y3": 841.9216918945312,
-          "coord_origin": "TOPLEFT"
+          "coord_origin": "BOTTOMLEFT"
         },
-        "text": "Docling bundles PDF document conversion to",
+        "boundary_type": "crop_box",
-        "orig": "Docling bundles PDF document conversion to",
+        "art_bbox": {
-        "text_direction": "left_to_right",
+          "l": 0.0,
-        "confidence": 1.0,
+          "t": 841.9216918945312,
-        "from_ocr": true
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "bleed_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "crop_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "media_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        },
+        "trim_bbox": {
+          "l": 0.0,
+          "t": 841.9216918945312,
+          "r": 595.201171875,
+          "b": 0.0,
+          "coord_origin": "BOTTOMLEFT"
+        }
       },
-      {
+      "bitmap_resources": [],
-        "index": 1,
+      "char_cells": [],
-        "rgba": {
+      "word_cells": [],
-          "r": 0,
+      "textline_cells": [
-          "g": 0,
+        {
-          "b": 0,
+          "index": 0,
-          "a": 255
+          "rgba": {
+            "r": 0,
+            "g": 0,
+            "b": 0,
+            "a": 255
+          },
+          "rect": {
+            "r_x0": 77.10171545548258,
+            "r_y0": 520.7638571913312,
+            "r_x1": 96.68315797053792,
+            "r_y1": 520.7638571913312,
+            "r_x2": 96.68315797053792,
+            "r_y2": 89.2388734673729,
+            "r_x3": 77.10171545548258,
+            "r_y3": 89.2388734673729,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "Docling bundles PDF document conversion to",
+          "orig": "Docling bundles PDF document conversion to",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "rect": {
+        {
-          "r_x0": 100.64168123325977,
+          "index": 1,
-          "r_y0": 523.3236155182395,
+          "rgba": {
-          "r_x1": 126.08064862014129,
+            "r": 0,
-          "r_y1": 523.3236155182395,
+            "g": 0,
-          "r_x2": 126.08064862014129,
+            "b": 0,
-          "r_y2": 89.1266754140729,
+            "a": 255
-          "r_x3": 100.64168123325977,
+          },
-          "r_y3": 89.1266754140729,
+          "rect": {
-          "coord_origin": "TOPLEFT"
+            "r_x0": 100.64168123325977,
+            "r_y0": 523.3236155182395,
+            "r_x1": 126.08064862014129,
+            "r_y1": 523.3236155182395,
+            "r_x2": 126.08064862014129,
+            "r_y2": 89.1266754140729,
+            "r_x3": 100.64168123325977,
+            "r_y3": 89.1266754140729,
+            "coord_origin": "TOPLEFT"
+          },
+          "text": "JSON and Markdown in an easy self contained",
+          "orig": "JSON and Markdown in an easy self contained",
+          "text_direction": "left_to_right",
+          "confidence": 1.0,
+          "from_ocr": true
         },
-        "text": "JSON and Markdown in an easy self contained",
+        {
-        "orig": "JSON and Markdown in an easy self contained",
+          "index": 2,
-        "text_direction": "left_to_right",
+          "rgba": {
-        "confidence": 1.0,
+            "r": 0,
-        "from_ocr": true
+            "g": 0,
-      },
+            "b": 0,
-      {
+            "a": 255
-        "index": 2,
+          },
-        "rgba": {
+          "rect": {
-          "r": 0,
+            "r_x0": 131.21306574279092,
-          "g": 0,
+            "r_y0": 521.0762158417759,
-          "b": 0,
+            "r_x1": 152.19606490864376,
-          "a": 255
+            "r_y1": 521.0762158417759,
-        },
+            "r_x2": 152.19606490864376,
-        "rect": {
+            "r_y2": 441.0071698212682,
-          "r_x0": 131.21306574279092,
+            "r_x3": 131.21306574279092,
-          "r_y0": 521.0762158417759,
+            "r_y3": 441.0071698212682,
-          "r_x1": 152.19606490864376,
+            "coord_origin": "TOPLEFT"
-          "r_y1": 521.0762158417759,
+          },
-          "r_x2": 152.19606490864376,
+          "text": "package",
-          "r_y2": 441.0071698212682,
+          "orig": "package",
-          "r_x3": 131.21306574279092,
+          "text_direction": "left_to_right",
-          "r_y3": 441.0071698212682,
+          "confidence": 1.0,
-          "coord_origin": "TOPLEFT"
+          "from_ocr": true
-        },
+        }
-        "text": "package",
+      ],
-        "orig": "package",
+      "has_chars": false,
-        "text_direction": "left_to_right",
+      "has_words": false,
-        "confidence": 1.0,
+      "has_lines": true,
-        "from_ocr": true
+      "image": null,
-      }
+      "lines": []
-    ],
+    },
-    "parsed_page": null,
     "predictions": {
       "layout": {
         "clusters": [

tests/test_e2e_ocr_conversion.py

@@ -57,14 +57,14 @@ def test_e2e_conversions():
     pdf_paths = get_pdf_paths()
 
     engines: List[Tuple[OcrOptions, bool]] = [
-        (EasyOcrOptions(), False),
         (TesseractOcrOptions(), True),
         (TesseractCliOcrOptions(), True),
-        (EasyOcrOptions(force_full_page_ocr=True), False),
+        (EasyOcrOptions(), False),
         (TesseractOcrOptions(force_full_page_ocr=True), True),
         (TesseractOcrOptions(force_full_page_ocr=True, lang=["auto"]), True),
         (TesseractCliOcrOptions(force_full_page_ocr=True), True),
         (TesseractCliOcrOptions(force_full_page_ocr=True, lang=["auto"]), True),
+        (EasyOcrOptions(force_full_page_ocr=True), False),
     ]
 
     # rapidocr is only available for Python >=3.6,<3.13