docling 2.0.0__py3-none-any.whl → 2.2.0__py3-none-any.whl

docling/document_converter.py

@@ -8,8 +8,10 @@ from typing import Dict, Iterable, Iterator, List, Optional, Type
 from pydantic import BaseModel, ConfigDict, model_validator, validate_call
 
 from docling.backend.abstract_backend import AbstractDocumentBackend
+from docling.backend.asciidoc_backend import AsciiDocBackend
 from docling.backend.docling_parse_backend import DoclingParseDocumentBackend
 from docling.backend.html_backend import HTMLDocumentBackend
+from docling.backend.md_backend import MarkdownDocumentBackend
 from docling.backend.mspowerpoint_backend import MsPowerpointDocumentBackend
 from docling.backend.msword_backend import MsWordDocumentBackend
 from docling.datamodel.base_models import ConversionStatus, DocumentStream, InputFormat
@@ -52,6 +54,16 @@ class PowerpointFormatOption(FormatOption):
     backend: Type[AbstractDocumentBackend] = MsPowerpointDocumentBackend
 
 
+class MarkdownFormatOption(FormatOption):
+    pipeline_cls: Type = SimplePipeline
+    backend: Type[AbstractDocumentBackend] = MarkdownDocumentBackend
+
+
+class AsciiDocFormatOption(FormatOption):
+    pipeline_cls: Type = SimplePipeline
+    backend: Type[AbstractDocumentBackend] = AsciiDocBackend
+
+
 class HTMLFormatOption(FormatOption):
     pipeline_cls: Type = SimplePipeline
     backend: Type[AbstractDocumentBackend] = HTMLDocumentBackend
@@ -74,6 +86,12 @@ _format_to_default_options = {
     InputFormat.PPTX: FormatOption(
         pipeline_cls=SimplePipeline, backend=MsPowerpointDocumentBackend
     ),
+    InputFormat.MD: FormatOption(
+        pipeline_cls=SimplePipeline, backend=MarkdownDocumentBackend
+    ),
+    InputFormat.ASCIIDOC: FormatOption(
+        pipeline_cls=SimplePipeline, backend=AsciiDocBackend
+    ),
     InputFormat.HTML: FormatOption(
         pipeline_cls=SimplePipeline, backend=HTMLDocumentBackend
     ),

docling/models/base_ocr_model.py

@@ -69,7 +69,7 @@ class BaseOcrModel:
         coverage, ocr_rects = find_ocr_rects(page.size, bitmap_rects)
 
         # return full-page rectangle if sufficiently covered with bitmaps
-        if coverage > BITMAP_COVERAGE_TRESHOLD:
+        if coverage > max(BITMAP_COVERAGE_TRESHOLD, self.options.bitmap_area_threshold):
             return [
                 BoundingBox(
                     l=0,
@@ -81,6 +81,14 @@ class BaseOcrModel:
             ]
         # return individual rectangles if the bitmap coverage is smaller
         else:  # coverage <= BITMAP_COVERAGE_TRESHOLD:
+
+            # skip OCR if the bitmap area on the page is smaller than the options threshold
+            ocr_rects = [
+                rect
+                for rect in ocr_rects
+                if rect.area() / (page.size.width * page.size.height)
+                > self.options.bitmap_area_threshold
+            ]
             return ocr_rects
 
     # Filters OCR cells by dropping any OCR cell that intersects with an existing programmatic cell.

docling/models/ds_glm_model.py

@@ -5,15 +5,23 @@ from typing import List, Union
 from deepsearch_glm.nlp_utils import init_nlp_model
 from deepsearch_glm.utils.doc_utils import to_docling_document
 from deepsearch_glm.utils.load_pretrained_models import load_pretrained_nlp_models
-from docling_core.types import BaseText
-from docling_core.types import Document as DsDocument
-from docling_core.types import DocumentDescription as DsDocumentDescription
-from docling_core.types import FileInfoObject as DsFileInfoObject
-from docling_core.types import PageDimensions, PageReference, Prov, Ref
-from docling_core.types import Table as DsSchemaTable
 from docling_core.types.doc import BoundingBox, CoordOrigin, DoclingDocument
 from docling_core.types.legacy_doc.base import BoundingBox as DsBoundingBox
-from docling_core.types.legacy_doc.base import Figure, TableCell
+from docling_core.types.legacy_doc.base import (
+    Figure,
+    PageDimensions,
+    PageReference,
+    Prov,
+    Ref,
+)
+from docling_core.types.legacy_doc.base import Table as DsSchemaTable
+from docling_core.types.legacy_doc.base import TableCell
+from docling_core.types.legacy_doc.document import BaseText
+from docling_core.types.legacy_doc.document import (
+    CCSDocumentDescription as DsDocumentDescription,
+)
+from docling_core.types.legacy_doc.document import CCSFileInfoObject as DsFileInfoObject
+from docling_core.types.legacy_doc.document import ExportedCCSDocument as DsDocument
 from PIL import ImageDraw
 from pydantic import BaseModel, ConfigDict
 
@@ -202,6 +210,7 @@ class GlmModel:
         page_dimensions = [
             PageDimensions(page=p.page_no + 1, height=p.size.height, width=p.size.width)
             for p in conv_res.pages
+            if p.size is not None
         ]
 
         ds_doc: DsDocument = DsDocument(

docling/models/easyocr_model.py

@@ -41,48 +41,50 @@ class EasyOcrModel(BaseOcrModel):
 
         for page in page_batch:
             assert page._backend is not None
-
-            ocr_rects = self.get_ocr_rects(page)
-
-            all_ocr_cells = []
-            for ocr_rect in ocr_rects:
-                # Skip zero area boxes
-                if ocr_rect.area() == 0:
-                    continue
-                high_res_image = page._backend.get_page_image(
-                    scale=self.scale, cropbox=ocr_rect
-                )
-                im = numpy.array(high_res_image)
-                result = self.reader.readtext(im)
-
-                del high_res_image
-                del im
-
-                cells = [
-                    OcrCell(
-                        id=ix,
-                        text=line[1],
-                        confidence=line[2],
-                        bbox=BoundingBox.from_tuple(
-                            coord=(
-                                (line[0][0][0] / self.scale) + ocr_rect.l,
-                                (line[0][0][1] / self.scale) + ocr_rect.t,
-                                (line[0][2][0] / self.scale) + ocr_rect.l,
-                                (line[0][2][1] / self.scale) + ocr_rect.t,
-                            ),
-                            origin=CoordOrigin.TOPLEFT,
-                        ),
+            if not page._backend.is_valid():
+                yield page
+            else:
+                ocr_rects = self.get_ocr_rects(page)
+
+                all_ocr_cells = []
+                for ocr_rect in ocr_rects:
+                    # Skip zero area boxes
+                    if ocr_rect.area() == 0:
+                        continue
+                    high_res_image = page._backend.get_page_image(
+                        scale=self.scale, cropbox=ocr_rect
                     )
-                    for ix, line in enumerate(result)
-                ]
-                all_ocr_cells.extend(cells)
+                    im = numpy.array(high_res_image)
+                    result = self.reader.readtext(im)
+
+                    del high_res_image
+                    del im
+
+                    cells = [
+                        OcrCell(
+                            id=ix,
+                            text=line[1],
+                            confidence=line[2],
+                            bbox=BoundingBox.from_tuple(
+                                coord=(
+                                    (line[0][0][0] / self.scale) + ocr_rect.l,
+                                    (line[0][0][1] / self.scale) + ocr_rect.t,
+                                    (line[0][2][0] / self.scale) + ocr_rect.l,
+                                    (line[0][2][1] / self.scale) + ocr_rect.t,
+                                ),
+                                origin=CoordOrigin.TOPLEFT,
+                            ),
+                        )
+                        for ix, line in enumerate(result)
+                    ]
+                    all_ocr_cells.extend(cells)
 
-            ## Remove OCR cells which overlap with programmatic cells.
-            filtered_ocr_cells = self.filter_ocr_cells(all_ocr_cells, page.cells)
+                ## Remove OCR cells which overlap with programmatic cells.
+                filtered_ocr_cells = self.filter_ocr_cells(all_ocr_cells, page.cells)
 
-            page.cells.extend(filtered_ocr_cells)
+                page.cells.extend(filtered_ocr_cells)
 
-            # DEBUG code:
-            # self.draw_ocr_rects_and_cells(page, ocr_rects)
+                # DEBUG code:
+                # self.draw_ocr_rects_and_cells(page, ocr_rects)
 
-            yield page
+                yield page

docling/models/layout_model.py

@@ -273,68 +273,72 @@ class LayoutModel(BasePageModel):
 
     def __call__(self, page_batch: Iterable[Page]) -> Iterable[Page]:
         for page in page_batch:
-            assert page.size is not None
-
-            clusters = []
-            for ix, pred_item in enumerate(
-                self.layout_predictor.predict(page.get_image(scale=1.0))
-            ):
-                label = DocItemLabel(
-                    pred_item["label"].lower().replace(" ", "_").replace("-", "_")
-                )  # Temporary, until docling-ibm-model uses docling-core types
-                cluster = Cluster(
-                    id=ix,
-                    label=label,
-                    confidence=pred_item["confidence"],
-                    bbox=BoundingBox.model_validate(pred_item),
-                    cells=[],
-                )
-                clusters.append(cluster)
-
-            # Map cells to clusters
-            # TODO: Remove, postprocess should take care of it anyway.
-            for cell in page.cells:
-                for cluster in clusters:
-                    if not cell.bbox.area() > 0:
-                        overlap_frac = 0.0
-                    else:
-                        overlap_frac = (
-                            cell.bbox.intersection_area_with(cluster.bbox)
-                            / cell.bbox.area()
-                        )
-
-                    if overlap_frac > 0.5:
-                        cluster.cells.append(cell)
-
-            # Pre-sort clusters
-            # clusters = self.sort_clusters_by_cell_order(clusters)
-
-            # DEBUG code:
-            def draw_clusters_and_cells():
-                image = copy.deepcopy(page.image)
-                draw = ImageDraw.Draw(image)
-                for c in clusters:
-                    x0, y0, x1, y1 = c.bbox.as_tuple()
-                    draw.rectangle([(x0, y0), (x1, y1)], outline="green")
-
-                    cell_color = (
-                        random.randint(30, 140),
-                        random.randint(30, 140),
-                        random.randint(30, 140),
+            assert page._backend is not None
+            if not page._backend.is_valid():
+                yield page
+            else:
+                assert page.size is not None
+
+                clusters = []
+                for ix, pred_item in enumerate(
+                    self.layout_predictor.predict(page.get_image(scale=1.0))
+                ):
+                    label = DocItemLabel(
+                        pred_item["label"].lower().replace(" ", "_").replace("-", "_")
+                    )  # Temporary, until docling-ibm-model uses docling-core types
+                    cluster = Cluster(
+                        id=ix,
+                        label=label,
+                        confidence=pred_item["confidence"],
+                        bbox=BoundingBox.model_validate(pred_item),
+                        cells=[],
                     )
-                    for tc in c.cells:  # [:1]:
-                        x0, y0, x1, y1 = tc.bbox.as_tuple()
-                        draw.rectangle([(x0, y0), (x1, y1)], outline=cell_color)
-                image.show()
+                    clusters.append(cluster)
+
+                # Map cells to clusters
+                # TODO: Remove, postprocess should take care of it anyway.
+                for cell in page.cells:
+                    for cluster in clusters:
+                        if not cell.bbox.area() > 0:
+                            overlap_frac = 0.0
+                        else:
+                            overlap_frac = (
+                                cell.bbox.intersection_area_with(cluster.bbox)
+                                / cell.bbox.area()
+                            )
+
+                        if overlap_frac > 0.5:
+                            cluster.cells.append(cell)
+
+                # Pre-sort clusters
+                # clusters = self.sort_clusters_by_cell_order(clusters)
+
+                # DEBUG code:
+                def draw_clusters_and_cells():
+                    image = copy.deepcopy(page.image)
+                    draw = ImageDraw.Draw(image)
+                    for c in clusters:
+                        x0, y0, x1, y1 = c.bbox.as_tuple()
+                        draw.rectangle([(x0, y0), (x1, y1)], outline="green")
+
+                        cell_color = (
+                            random.randint(30, 140),
+                            random.randint(30, 140),
+                            random.randint(30, 140),
+                        )
+                        for tc in c.cells:  # [:1]:
+                            x0, y0, x1, y1 = tc.bbox.as_tuple()
+                            draw.rectangle([(x0, y0), (x1, y1)], outline=cell_color)
+                    image.show()
 
-            # draw_clusters_and_cells()
+                # draw_clusters_and_cells()
 
-            clusters, page.cells = self.postprocess(
-                clusters, page.cells, page.size.height
-            )
+                clusters, page.cells = self.postprocess(
+                    clusters, page.cells, page.size.height
+                )
 
-            # draw_clusters_and_cells()
+                # draw_clusters_and_cells()
 
-            page.predictions.layout = LayoutPrediction(clusters=clusters)
+                page.predictions.layout = LayoutPrediction(clusters=clusters)
 
-            yield page
+                yield page

docling/models/page_assemble_model.py

@@ -54,111 +54,119 @@ class PageAssembleModel(BasePageModel):
     def __call__(self, page_batch: Iterable[Page]) -> Iterable[Page]:
         for page in page_batch:
             assert page._backend is not None
-            assert page.predictions.layout is not None
-            # assembles some JSON output page by page.
-
-            elements: List[PageElement] = []
-            headers: List[PageElement] = []
-            body: List[PageElement] = []
-
-            for cluster in page.predictions.layout.clusters:
-                # _log.info("Cluster label seen:", cluster.label)
-                if cluster.label in LayoutModel.TEXT_ELEM_LABELS:
-
-                    textlines = [
-                        cell.text.replace("\x02", "-").strip()
-                        for cell in cluster.cells
-                        if len(cell.text.strip()) > 0
-                    ]
-                    text = self.sanitize_text(textlines)
-                    text_el = TextElement(
-                        label=cluster.label,
-                        id=cluster.id,
-                        text=text,
-                        page_no=page.page_no,
-                        cluster=cluster,
-                    )
-                    elements.append(text_el)
-
-                    if cluster.label in LayoutModel.PAGE_HEADER_LABELS:
-                        headers.append(text_el)
-                    else:
-                        body.append(text_el)
-                elif cluster.label == LayoutModel.TABLE_LABEL:
-                    tbl = None
-                    if page.predictions.tablestructure:
-                        tbl = page.predictions.tablestructure.table_map.get(
-                            cluster.id, None
-                        )
-                    if (
-                        not tbl
-                    ):  # fallback: add table without structure, if it isn't present
-                        tbl = Table(
-                            label=cluster.label,
-                            id=cluster.id,
-                            text="",
-                            otsl_seq=[],
-                            table_cells=[],
-                            cluster=cluster,
-                            page_no=page.page_no,
-                        )
+            if not page._backend.is_valid():
+                yield page
+            else:
+                assert page.predictions.layout is not None
 
-                    elements.append(tbl)
-                    body.append(tbl)
-                elif cluster.label == LayoutModel.FIGURE_LABEL:
-                    fig = None
-                    if page.predictions.figures_classification:
-                        fig = page.predictions.figures_classification.figure_map.get(
-                            cluster.id, None
-                        )
-                    if (
-                        not fig
-                    ):  # fallback: add figure without classification, if it isn't present
-                        fig = FigureElement(
+                # assembles some JSON output page by page.
+
+                elements: List[PageElement] = []
+                headers: List[PageElement] = []
+                body: List[PageElement] = []
+
+                for cluster in page.predictions.layout.clusters:
+                    # _log.info("Cluster label seen:", cluster.label)
+                    if cluster.label in LayoutModel.TEXT_ELEM_LABELS:
+
+                        textlines = [
+                            cell.text.replace("\x02", "-").strip()
+                            for cell in cluster.cells
+                            if len(cell.text.strip()) > 0
+                        ]
+                        text = self.sanitize_text(textlines)
+                        text_el = TextElement(
                             label=cluster.label,
                             id=cluster.id,
-                            text="",
-                            data=None,
-                            cluster=cluster,
+                            text=text,
                             page_no=page.page_no,
+                            cluster=cluster,
                         )
-                    elements.append(fig)
-                    body.append(fig)
-                elif cluster.label == LayoutModel.FORMULA_LABEL:
-                    equation = None
-                    if page.predictions.equations_prediction:
-                        equation = (
-                            page.predictions.equations_prediction.equation_map.get(
+                        elements.append(text_el)
+
+                        if cluster.label in LayoutModel.PAGE_HEADER_LABELS:
+                            headers.append(text_el)
+                        else:
+                            body.append(text_el)
+                    elif cluster.label == LayoutModel.TABLE_LABEL:
+                        tbl = None
+                        if page.predictions.tablestructure:
+                            tbl = page.predictions.tablestructure.table_map.get(
                                 cluster.id, None
                             )
-                        )
-                    if not equation:  # fallback: add empty formula, if it isn't present
-                        text = self.sanitize_text(
-                            [
-                                cell.text.replace("\x02", "-").strip()
-                                for cell in cluster.cells
-                                if len(cell.text.strip()) > 0
-                            ]
-                        )
-                        equation = TextElement(
-                            label=cluster.label,
-                            id=cluster.id,
-                            cluster=cluster,
-                            page_no=page.page_no,
-                            text=text,
-                        )
-                    elements.append(equation)
-                    body.append(equation)
+                        if (
+                            not tbl
+                        ):  # fallback: add table without structure, if it isn't present
+                            tbl = Table(
+                                label=cluster.label,
+                                id=cluster.id,
+                                text="",
+                                otsl_seq=[],
+                                table_cells=[],
+                                cluster=cluster,
+                                page_no=page.page_no,
+                            )
+
+                        elements.append(tbl)
+                        body.append(tbl)
+                    elif cluster.label == LayoutModel.FIGURE_LABEL:
+                        fig = None
+                        if page.predictions.figures_classification:
+                            fig = (
+                                page.predictions.figures_classification.figure_map.get(
+                                    cluster.id, None
+                                )
+                            )
+                        if (
+                            not fig
+                        ):  # fallback: add figure without classification, if it isn't present
+                            fig = FigureElement(
+                                label=cluster.label,
+                                id=cluster.id,
+                                text="",
+                                data=None,
+                                cluster=cluster,
+                                page_no=page.page_no,
+                            )
+                        elements.append(fig)
+                        body.append(fig)
+                    elif cluster.label == LayoutModel.FORMULA_LABEL:
+                        equation = None
+                        if page.predictions.equations_prediction:
+                            equation = (
+                                page.predictions.equations_prediction.equation_map.get(
+                                    cluster.id, None
+                                )
+                            )
+                        if (
+                            not equation
+                        ):  # fallback: add empty formula, if it isn't present
+                            text = self.sanitize_text(
+                                [
+                                    cell.text.replace("\x02", "-").strip()
+                                    for cell in cluster.cells
+                                    if len(cell.text.strip()) > 0
+                                ]
+                            )
+                            equation = TextElement(
+                                label=cluster.label,
+                                id=cluster.id,
+                                cluster=cluster,
+                                page_no=page.page_no,
+                                text=text,
+                            )
+                        elements.append(equation)
+                        body.append(equation)
 
-            page.assembled = AssembledUnit(
-                elements=elements, headers=headers, body=body
-            )
+                page.assembled = AssembledUnit(
+                    elements=elements, headers=headers, body=body
+                )
 
-            # Remove page images (can be disabled)
-            if not self.options.keep_images:
-                page._image_cache = {}
+                # Remove page images (can be disabled)
+                if not self.options.keep_images:
+                    page._image_cache = {}
 
-            # Unload backend
-            page._backend.unload()
+                # Unload backend
+                page._backend.unload()
 
-            yield page
+                yield page

docling/models/page_preprocessing_model.py

@@ -17,9 +17,13 @@ class PagePreprocessingModel(BasePageModel):
 
     def __call__(self, page_batch: Iterable[Page]) -> Iterable[Page]:
         for page in page_batch:
-            page = self._populate_page_images(page)
-            page = self._parse_page_cells(page)
-            yield page
+            assert page._backend is not None
+            if not page._backend.is_valid():
+                yield page
+            else:
+                page = self._populate_page_images(page)
+                page = self._parse_page_cells(page)
+                yield page
 
     # Generate the page image and store it in the page object
     def _populate_page_images(self, page: Page) -> Page: