docling 1.10.0__tar.gz → 1.12.0__tar.gz

{docling-1.10.0 → docling-1.12.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: docling
-Version: 1.10.0
+Version: 1.12.0
 Summary: Docling PDF conversion package
 Home-page: https://github.com/DS4SD/docling
 License: MIT
@@ -19,21 +19,34 @@ Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Provides-Extra: examples
 Requires-Dist: certifi (>=2024.7.4)
 Requires-Dist: deepsearch-glm (>=0.21.0,<0.22.0)
-Requires-Dist: docling-core (>=1.1.3,<2.0.0)
-Requires-Dist: docling-ibm-models (>=1.1.3,<2.0.0)
+Requires-Dist: docling-core (>=1.3.0,<2.0.0)
+Requires-Dist: docling-ibm-models (>=1.1.7,<2.0.0)
 Requires-Dist: docling-parse (>=1.2.0,<2.0.0)
 Requires-Dist: easyocr (>=1.7,<2.0)
 Requires-Dist: filetype (>=1.2.0,<2.0.0)
 Requires-Dist: huggingface_hub (>=0.23,<1)
+Requires-Dist: langchain-huggingface (>=0.0.3,<0.0.4) ; extra == "examples"
+Requires-Dist: langchain-milvus (>=0.1.4,<0.2.0) ; extra == "examples"
+Requires-Dist: langchain-text-splitters (>=0.2.4,<0.3.0) ; extra == "examples"
+Requires-Dist: llama-index-embeddings-huggingface (>=0.3.1,<0.4.0) ; extra == "examples"
+Requires-Dist: llama-index-llms-huggingface-api (>=0.2.0,<0.3.0) ; extra == "examples"
+Requires-Dist: llama-index-vector-stores-milvus (>=0.2.1,<0.3.0) ; extra == "examples"
 Requires-Dist: pyarrow (>=16.1.0,<17.0.0)
 Requires-Dist: pydantic (>=2.0.0,<3.0.0)
 Requires-Dist: pydantic-settings (>=2.3.0,<3.0.0)
 Requires-Dist: pypdfium2 (>=4.30.0,<5.0.0)
+Requires-Dist: python-dotenv (>=1.0.1,<2.0.0) ; extra == "examples"
 Requires-Dist: requests (>=2.32.3,<3.0.0)
 Requires-Dist: rtree (>=1.3.0,<2.0.0)
 Requires-Dist: scipy (>=1.14.1,<2.0.0)
+Requires-Dist: torch (>=2.2.2,<2.3.0) ; sys_platform == "darwin" and platform_machine == "x86_64"
+Requires-Dist: torch (>=2.2.2,<3.0.0) ; sys_platform != "darwin" or platform_machine != "x86_64"
+Requires-Dist: torchvision (>=0,<1) ; sys_platform != "darwin" or platform_machine != "x86_64"
+Requires-Dist: torchvision (>=0.17.2,<0.18.0) ; sys_platform == "darwin" and platform_machine == "x86_64"
+Requires-Dist: typer (>=0.12.5,<0.13.0)
 Project-URL: Repository, https://github.com/DS4SD/docling
 Description-Content-Type: text/markdown
 
@@ -62,8 +75,7 @@ Docling bundles PDF document conversion to JSON and Markdown in an easy, self-co
 * 📑 Understands detailed page layout, reading order and recovers table structures
 * 📝 Extracts metadata from the document, such as title, authors, references and language
 * 🔍 Optionally applies OCR (use with scanned PDFs)
-
-For RAG, check out [Quackling](https://github.com/DS4SD/quackling) to get the most out of your docs, be it using LlamaIndex, LangChain or your pipeline.
+* 🤖 Integrates easily with LLM app / RAG frameworks like 🦙 LlamaIndex and 🦜🔗 LangChain
 
 ## Installation
 
@@ -182,6 +194,10 @@ results = doc_converter.convert(conv_input)
 
 You can limit the CPU threads used by Docling by setting the environment variable `OMP_NUM_THREADS` accordingly. The default setting is using 4 CPU threads.
 
+### RAG
+Check out the following examples showcasing RAG using Docling with standard LLM application frameworks:
+- [Basic RAG pipeline with 🦙 LlamaIndex](https://github.com/DS4SD/docling/tree/main/examples/rag_llamaindex.ipynb)
+- [Basic RAG pipeline with 🦜🔗 LangChain](https://github.com/DS4SD/docling/tree/main/examples/rag_langchain.ipynb)
 
 ## Technical report
 

{docling-1.10.0 → docling-1.12.0}/README.md

@@ -23,8 +23,7 @@ Docling bundles PDF document conversion to JSON and Markdown in an easy, self-co
 * 📑 Understands detailed page layout, reading order and recovers table structures
 * 📝 Extracts metadata from the document, such as title, authors, references and language
 * 🔍 Optionally applies OCR (use with scanned PDFs)
-
-For RAG, check out [Quackling](https://github.com/DS4SD/quackling) to get the most out of your docs, be it using LlamaIndex, LangChain or your pipeline.
+* 🤖 Integrates easily with LLM app / RAG frameworks like 🦙 LlamaIndex and 🦜🔗 LangChain
 
 ## Installation
 
@@ -143,6 +142,10 @@ results = doc_converter.convert(conv_input)
 
 You can limit the CPU threads used by Docling by setting the environment variable `OMP_NUM_THREADS` accordingly. The default setting is using 4 CPU threads.
 
+### RAG
+Check out the following examples showcasing RAG using Docling with standard LLM application frameworks:
+- [Basic RAG pipeline with 🦙 LlamaIndex](https://github.com/DS4SD/docling/tree/main/examples/rag_llamaindex.ipynb)
+- [Basic RAG pipeline with 🦜🔗 LangChain](https://github.com/DS4SD/docling/tree/main/examples/rag_langchain.ipynb)
 
 ## Technical report
 

{docling-1.10.0 → docling-1.12.0}/docling/backend/abstract_backend.py

@@ -1,10 +1,13 @@
 from abc import ABC, abstractmethod
 from io import BytesIO
 from pathlib import Path
-from typing import Any, Iterable, Optional, Union
+from typing import TYPE_CHECKING, Any, Iterable, Optional, Union
 
 from PIL import Image
 
+if TYPE_CHECKING:
+    from docling.datamodel.base_models import BoundingBox, Cell, PageSize
+
 
 class PdfPageBackend(ABC):
 
@@ -17,12 +20,12 @@ class PdfPageBackend(ABC):
         pass
 
     @abstractmethod
-    def get_bitmap_rects(self, scale: int = 1) -> Iterable["BoundingBox"]:
+    def get_bitmap_rects(self, float: int = 1) -> Iterable["BoundingBox"]:
         pass
 
     @abstractmethod
     def get_page_image(
-        self, scale: int = 1, cropbox: Optional["BoundingBox"] = None
+        self, scale: float = 1, cropbox: Optional["BoundingBox"] = None
     ) -> Image.Image:
         pass
 

{docling-1.10.0 → docling-1.12.0}/docling/backend/docling_parse_backend.py

@@ -2,7 +2,7 @@ import logging
 import random
 from io import BytesIO
 from pathlib import Path
-from typing import Iterable, Optional, Union
+from typing import Iterable, List, Optional, Union
 
 import pypdfium2 as pdfium
 from docling_parse.docling_parse import pdf_parser
@@ -22,7 +22,6 @@ class DoclingParsePageBackend(PdfPageBackend):
         self._ppage = page_obj
         parsed_page = parser.parse_pdf_from_key_on_page(document_hash, page_no)
 
-        self._dpage = None
         self.valid = "pages" in parsed_page
         if self.valid:
             self._dpage = parsed_page["pages"][0]
@@ -68,7 +67,7 @@ class DoclingParsePageBackend(PdfPageBackend):
         return text_piece
 
     def get_text_cells(self) -> Iterable[Cell]:
-        cells = []
+        cells: List[Cell] = []
         cell_counter = 0
 
         if not self.valid:
@@ -130,7 +129,7 @@ class DoclingParsePageBackend(PdfPageBackend):
 
         return cells
 
-    def get_bitmap_rects(self, scale: int = 1) -> Iterable[BoundingBox]:
+    def get_bitmap_rects(self, scale: float = 1) -> Iterable[BoundingBox]:
         AREA_THRESHOLD = 32 * 32
 
         for i in range(len(self._dpage["images"])):
@@ -145,7 +144,7 @@ class DoclingParsePageBackend(PdfPageBackend):
                 yield cropbox
 
     def get_page_image(
-        self, scale: int = 1, cropbox: Optional[BoundingBox] = None
+        self, scale: float = 1, cropbox: Optional[BoundingBox] = None
     ) -> Image.Image:
 
         page_size = self.get_size()

{docling-1.10.0 → docling-1.12.0}/docling/backend/pypdfium2_backend.py

@@ -7,7 +7,7 @@ from typing import Iterable, List, Optional, Union
 import pypdfium2 as pdfium
 import pypdfium2.raw as pdfium_c
 from PIL import Image, ImageDraw
-from pypdfium2 import PdfPage
+from pypdfium2 import PdfPage, PdfTextPage
 from pypdfium2._helpers.misc import PdfiumError
 
 from docling.backend.abstract_backend import PdfDocumentBackend, PdfPageBackend
@@ -29,12 +29,12 @@ class PyPdfiumPageBackend(PdfPageBackend):
                 exc_info=True,
             )
             self.valid = False
-        self.text_page = None
+        self.text_page: Optional[PdfTextPage] = None
 
     def is_valid(self) -> bool:
         return self.valid
 
-    def get_bitmap_rects(self, scale: int = 1) -> Iterable[BoundingBox]:
+    def get_bitmap_rects(self, scale: float = 1) -> Iterable[BoundingBox]:
         AREA_THRESHOLD = 32 * 32
         for obj in self._ppage.get_objects(filter=[pdfium_c.FPDF_PAGEOBJ_IMAGE]):
             pos = obj.get_pos()
@@ -189,7 +189,7 @@ class PyPdfiumPageBackend(PdfPageBackend):
         return cells
 
     def get_page_image(
-        self, scale: int = 1, cropbox: Optional[BoundingBox] = None
+        self, scale: float = 1, cropbox: Optional[BoundingBox] = None
     ) -> Image.Image:
 
         page_size = self.get_size()

docling-1.12.0/docling/cli/main.py

@@ -0,0 +1,257 @@
+import importlib
+import json
+import logging
+import time
+import warnings
+from enum import Enum
+from pathlib import Path
+from typing import Annotated, Iterable, List, Optional
+
+import typer
+from pydantic import AnyUrl
+
+from docling.backend.docling_parse_backend import DoclingParseDocumentBackend
+from docling.backend.pypdfium2_backend import PyPdfiumDocumentBackend
+from docling.datamodel.base_models import ConversionStatus, PipelineOptions
+from docling.datamodel.document import ConversionResult, DocumentConversionInput
+from docling.document_converter import DocumentConverter
+
+warnings.filterwarnings(action="ignore", category=UserWarning, module="pydantic|torch")
+warnings.filterwarnings(action="ignore", category=FutureWarning, module="easyocr")
+
+_log = logging.getLogger(__name__)
+from rich.console import Console
+
+err_console = Console(stderr=True)
+
+
+app = typer.Typer(
+    name="Docling",
+    no_args_is_help=True,
+    add_completion=False,
+    pretty_exceptions_enable=False,
+)
+
+
+def version_callback(value: bool):
+    if value:
+        docling_version = importlib.metadata.version("docling")
+        docling_core_version = importlib.metadata.version("docling-core")
+        docling_ibm_models_version = importlib.metadata.version("docling-ibm-models")
+        docling_parse_version = importlib.metadata.version("docling-parse")
+        print(f"Docling version: {docling_version}")
+        print(f"Docling Core version: {docling_core_version}")
+        print(f"Docling IBM Models version: {docling_ibm_models_version}")
+        print(f"Docling Parse version: {docling_parse_version}")
+        raise typer.Exit()
+
+
+# Define an enum for the backend options
+class Backend(str, Enum):
+    PYPDFIUM2 = "pypdfium2"
+    DOCLING = "docling"
+
+
+def export_documents(
+    conv_results: Iterable[ConversionResult],
+    output_dir: Path,
+    export_json: bool,
+    export_md: bool,
+    export_txt: bool,
+    export_doctags: bool,
+):
+
+    success_count = 0
+    failure_count = 0
+
+    for conv_res in conv_results:
+        if conv_res.status == ConversionStatus.SUCCESS:
+            success_count += 1
+            doc_filename = conv_res.input.file.stem
+
+            # Export Deep Search document JSON format:
+            if export_json:
+                fname = output_dir / f"{doc_filename}.json"
+                with fname.open("w") as fp:
+                    _log.info(f"writing JSON output to {fname}")
+                    fp.write(json.dumps(conv_res.render_as_dict()))
+
+            # Export Text format:
+            if export_txt:
+                fname = output_dir / f"{doc_filename}.txt"
+                with fname.open("w") as fp:
+                    _log.info(f"writing Text output to {fname}")
+                    fp.write(conv_res.render_as_text())
+
+            # Export Markdown format:
+            if export_md:
+                fname = output_dir / f"{doc_filename}.md"
+                with fname.open("w") as fp:
+                    _log.info(f"writing Markdown output to {fname}")
+                    fp.write(conv_res.render_as_markdown())
+
+            # Export Document Tags format:
+            if export_doctags:
+                fname = output_dir / f"{doc_filename}.doctags"
+                with fname.open("w") as fp:
+                    _log.info(f"writing Doc Tags output to {fname}")
+                    fp.write(conv_res.render_as_doctags())
+
+        else:
+            _log.warning(f"Document {conv_res.input.file} failed to convert.")
+            failure_count += 1
+
+    _log.info(
+        f"Processed {success_count + failure_count} docs, of which {failure_count} failed"
+    )
+
+
+@app.command(no_args_is_help=True)
+def convert(
+    input_sources: Annotated[
+        List[Path],
+        typer.Argument(
+            ...,
+            metavar="source",
+            help="PDF files to convert. Directories are also accepted.",
+        ),
+    ],
+    export_json: Annotated[
+        bool,
+        typer.Option(
+            ..., "--json/--no-json", help="If enabled the document is exported as JSON."
+        ),
+    ] = False,
+    export_md: Annotated[
+        bool,
+        typer.Option(
+            ..., "--md/--no-md", help="If enabled the document is exported as Markdown."
+        ),
+    ] = True,
+    export_txt: Annotated[
+        bool,
+        typer.Option(
+            ..., "--txt/--no-txt", help="If enabled the document is exported as Text."
+        ),
+    ] = False,
+    export_doctags: Annotated[
+        bool,
+        typer.Option(
+            ...,
+            "--doctags/--no-doctags",
+            help="If enabled the document is exported as Doc Tags.",
+        ),
+    ] = False,
+    ocr: Annotated[
+        bool,
+        typer.Option(
+            ..., help="If enabled, the bitmap content will be processed using OCR."
+        ),
+    ] = True,
+    backend: Annotated[
+        Backend, typer.Option(..., help="The PDF backend to use.")
+    ] = Backend.DOCLING,
+    output: Annotated[
+        Path, typer.Option(..., help="Output directory where results are saved.")
+    ] = Path("."),
+    version: Annotated[
+        Optional[bool],
+        typer.Option(
+            "--version",
+            callback=version_callback,
+            is_eager=True,
+            help="Show version information.",
+        ),
+    ] = None,
+):
+    logging.basicConfig(level=logging.INFO)
+
+    input_doc_paths: List[Path] = []
+    for source in input_sources:
+        if not source.exists():
+            err_console.print(
+                f"[red]Error: The input file {source} does not exist.[/red]"
+            )
+            raise typer.Abort()
+        elif source.is_dir():
+            input_doc_paths.extend(list(source.glob("**/*.pdf", case_sensitive=False)))
+        else:
+            input_doc_paths.append(source)
+
+    ###########################################################################
+
+    # The following sections contain a combination of PipelineOptions
+    # and PDF Backends for various configurations.
+    # Uncomment one section at the time to see the differences in the output.
+
+    doc_converter = None
+    if backend == Backend.PYPDFIUM2 and not ocr:  # PyPdfium without OCR
+        pipeline_options = PipelineOptions()
+        pipeline_options.do_ocr = False
+        pipeline_options.do_table_structure = True
+        pipeline_options.table_structure_options.do_cell_matching = False
+
+        doc_converter = DocumentConverter(
+            pipeline_options=pipeline_options,
+            pdf_backend=PyPdfiumDocumentBackend,
+        )
+
+    elif backend == Backend.PYPDFIUM2.value and ocr:  # PyPdfium with OCR
+        pipeline_options = PipelineOptions()
+        pipeline_options.do_ocr = False
+        pipeline_options.do_table_structure = True
+        pipeline_options.table_structure_options.do_cell_matching = True
+
+        doc_converter = DocumentConverter(
+            pipeline_options=pipeline_options,
+            pdf_backend=PyPdfiumDocumentBackend,
+        )
+
+    elif backend == Backend.DOCLING.value and not ocr:  # Docling Parse without OCR
+        pipeline_options = PipelineOptions()
+        pipeline_options.do_ocr = False
+        pipeline_options.do_table_structure = True
+        pipeline_options.table_structure_options.do_cell_matching = True
+
+        doc_converter = DocumentConverter(
+            pipeline_options=pipeline_options,
+            pdf_backend=DoclingParseDocumentBackend,
+        )
+
+    elif backend == Backend.DOCLING.value and ocr:  # Docling Parse with OCR
+        pipeline_options = PipelineOptions()
+        pipeline_options.do_ocr = True
+        pipeline_options.do_table_structure = True
+        pipeline_options.table_structure_options.do_cell_matching = True
+
+        doc_converter = DocumentConverter(
+            pipeline_options=pipeline_options,
+            pdf_backend=DoclingParseDocumentBackend,
+        )
+
+    ###########################################################################
+
+    # Define input files
+    input = DocumentConversionInput.from_paths(input_doc_paths)
+
+    start_time = time.time()
+
+    conv_results = doc_converter.convert(input)
+
+    output.mkdir(parents=True, exist_ok=True)
+    export_documents(
+        conv_results,
+        output_dir=output,
+        export_json=export_json,
+        export_md=export_md,
+        export_txt=export_txt,
+        export_doctags=export_doctags,
+    )
+
+    end_time = time.time() - start_time
+
+    _log.info(f"All documents were converted in {end_time:.2f} seconds.")
+
+
+if __name__ == "__main__":
+    app()

{docling-1.10.0 → docling-1.12.0}/docling/datamodel/base_models.py

@@ -87,7 +87,7 @@ class BoundingBox(BaseModel):
             return (self.l, self.b, self.r, self.t)
 
     @classmethod
-    def from_tuple(cls, coord: Tuple[float], origin: CoordOrigin):
+    def from_tuple(cls, coord: Tuple[float, ...], origin: CoordOrigin):
         if origin == CoordOrigin.TOPLEFT:
             l, t, r, b = coord[0], coord[1], coord[2], coord[3]
             if r < l:
@@ -246,7 +246,7 @@ class EquationPrediction(BaseModel):
 
 
 class PagePredictions(BaseModel):
-    layout: LayoutPrediction = None
+    layout: Optional[LayoutPrediction] = None
     tablestructure: Optional[TableStructurePrediction] = None
     figures_classification: Optional[FigureClassificationPrediction] = None
     equations_prediction: Optional[EquationPrediction] = None
@@ -267,7 +267,7 @@ class Page(BaseModel):
     page_no: int
     page_hash: Optional[str] = None
     size: Optional[PageSize] = None
-    cells: List[Cell] = None
+    cells: List[Cell] = []
     predictions: PagePredictions = PagePredictions()
     assembled: Optional[AssembledUnit] = None
 

{docling-1.10.0 → docling-1.12.0}/docling/datamodel/document.py

@@ -11,6 +11,7 @@ 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 import TableCell
+from docling_core.types.doc.base import Figure
 from pydantic import BaseModel
 from typing_extensions import deprecated
 
@@ -279,7 +280,7 @@ class ConvertedDocument(BaseModel):
                     ),
                 )
                 figures.append(
-                    BaseCell(
+                    Figure(
                         prov=[
                             Prov(
                                 bbox=target_bbox,
@@ -312,8 +313,76 @@ class ConvertedDocument(BaseModel):
     def render_as_dict(self):
         return self.output.model_dump(by_alias=True, exclude_none=True)
 
-    def render_as_markdown(self):
-        return self.output.export_to_markdown()
+    def render_as_markdown(
+        self,
+        delim: str = "\n\n",
+        main_text_start: int = 0,
+        main_text_stop: Optional[int] = None,
+        main_text_labels: list[str] = [
+            "title",
+            "subtitle-level-1",
+            "paragraph",
+            "caption",
+            "table",
+        ],
+        strict_text: bool = False,
+    ):
+        return self.output.export_to_markdown(
+            delim=delim,
+            main_text_start=main_text_start,
+            main_text_stop=main_text_stop,
+            main_text_labels=main_text_labels,
+            strict_text=strict_text,
+        )
+
+    def render_as_text(
+        self,
+        delim: str = "\n\n",
+        main_text_start: int = 0,
+        main_text_stop: Optional[int] = None,
+        main_text_labels: list[str] = [
+            "title",
+            "subtitle-level-1",
+            "paragraph",
+            "caption",
+        ],
+    ):
+        return self.output.export_to_markdown(
+            delim=delim,
+            main_text_start=main_text_start,
+            main_text_stop=main_text_stop,
+            main_text_labels=main_text_labels,
+            strict_text=True,
+        )
+
+    def render_as_doctags(
+        self,
+        delim: str = "\n\n",
+        main_text_start: int = 0,
+        main_text_stop: Optional[int] = None,
+        main_text_labels: list[str] = [
+            "title",
+            "subtitle-level-1",
+            "paragraph",
+            "caption",
+            "table",
+            "figure",
+        ],
+        page_tagging: bool = True,
+        location_tagging: bool = True,
+        location_dimensions: Tuple[int, int] = (100, 100),
+        add_new_line: bool = True,
+    ) -> str:
+        return self.output.export_to_document_tokens(
+            delim=delim,
+            main_text_start=main_text_start,
+            main_text_stop=main_text_stop,
+            main_text_labels=main_text_labels,
+            page_tagging=page_tagging,
+            location_tagging=location_tagging,
+            location_dimensions=location_dimensions,
+            add_new_line=add_new_line,
+        )
 
     def render_element_images(
         self, element_types: Tuple[PageElement] = (FigureElement,)

{docling-1.10.0 → docling-1.12.0}/docling/pipeline/base_model_pipeline.py

@@ -1,12 +1,12 @@
 from pathlib import Path
-from typing import Iterable
+from typing import Callable, Iterable, List
 
 from docling.datamodel.base_models import Page, PipelineOptions
 
 
 class BaseModelPipeline:
     def __init__(self, artifacts_path: Path, pipeline_options: PipelineOptions):
-        self.model_pipe = []
+        self.model_pipe: List[Callable] = []
         self.artifacts_path = artifacts_path
         self.pipeline_options = pipeline_options
 

{docling-1.10.0 → docling-1.12.0}/docling/utils/export.py

@@ -1,10 +1,10 @@
 import logging
-from typing import Any, Dict, Iterable, List, Tuple
+from typing import Any, Dict, Iterable, List, Tuple, Union
 
-from docling_core.types.doc.base import BaseCell, Ref, Table, TableCell
+from docling_core.types.doc.base import BaseCell, BaseText, Ref, Table, TableCell
 
 from docling.datamodel.base_models import BoundingBox, CoordOrigin, OcrCell
-from docling.datamodel.document import ConvertedDocument, Page
+from docling.datamodel.document import ConversionResult, Page
 
 _log = logging.getLogger(__name__)
 
@@ -15,7 +15,10 @@ def _export_table_to_html(table: Table):
     # to the docling-core package.
 
     def _get_tablecell_span(cell: TableCell, ix):
-        span = set([s[ix] for s in cell.spans])
+        if cell.spans is None:
+            span = set()
+        else:
+            span = set([s[ix] for s in cell.spans])
         if len(span) == 0:
             return 1, None, None
         return len(span), min(span), max(span)
@@ -24,6 +27,8 @@ def _export_table_to_html(table: Table):
     nrows = table.num_rows
     ncols = table.num_cols
 
+    if table.data is None:
+        return ""
     for i in range(nrows):
         body += "<tr>"
         for j in range(ncols):
@@ -66,7 +71,7 @@ def _export_table_to_html(table: Table):
 
 
 def generate_multimodal_pages(
-    doc_result: ConvertedDocument,
+    doc_result: ConversionResult,
 ) -> Iterable[Tuple[str, str, List[Dict[str, Any]], List[Dict[str, Any]], Page]]:
 
     label_to_doclaynet = {
@@ -94,7 +99,7 @@ def generate_multimodal_pages(
     page_no = 0
     start_ix = 0
     end_ix = 0
-    doc_items = []
+    doc_items: List[Tuple[int, Union[BaseCell, BaseText]]] = []
 
     doc = doc_result.output
 
@@ -105,11 +110,11 @@ def generate_multimodal_pages(
             item_type = item.obj_type
             label = label_to_doclaynet.get(item_type, None)
 
-            if label is None:
+            if label is None or item.prov is None or page.size is None:
                 continue
 
             bbox = BoundingBox.from_tuple(
-                item.prov[0].bbox, origin=CoordOrigin.BOTTOMLEFT
+                tuple(item.prov[0].bbox), origin=CoordOrigin.BOTTOMLEFT
             )
             new_bbox = bbox.to_top_left_origin(page_height=page.size.height).normalized(
                 page_size=page.size
@@ -137,13 +142,15 @@ def generate_multimodal_pages(
         return segments
 
     def _process_page_cells(page: Page):
-        cells = []
+        cells: List[dict] = []
+        if page.size is None:
+            return cells
         for cell in page.cells:
             new_bbox = cell.bbox.to_top_left_origin(
                 page_height=page.size.height
             ).normalized(page_size=page.size)
             is_ocr = isinstance(cell, OcrCell)
-            ocr_confidence = cell.confidence if is_ocr else 1.0
+            ocr_confidence = cell.confidence if isinstance(cell, OcrCell) else 1.0
             cells.append(
                 {
                     "text": cell.text,
@@ -163,9 +170,15 @@ def generate_multimodal_pages(
         content_md = doc.export_to_markdown(
             main_text_start=start_ix, main_text_stop=end_ix
         )
+        # No page-tagging since we only do 1 page at the time
+        content_dt = doc.export_to_document_tokens(
+            main_text_start=start_ix, main_text_stop=end_ix, page_tagging=False
+        )
 
-        return content_text, content_md, page_cells, page_segments, page
+        return content_text, content_md, content_dt, page_cells, page_segments, page
 
+    if doc.main_text is None:
+        return
     for ix, orig_item in enumerate(doc.main_text):
 
         item = doc._resolve_ref(orig_item) if isinstance(orig_item, Ref) else orig_item

{docling-1.10.0 → docling-1.12.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "docling"
-version = "1.10.0"  # DO NOT EDIT, updated automatically
+version = "1.12.0"  # DO NOT EDIT, updated automatically
 description = "Docling PDF conversion package"
 authors = ["Christoph Auer <cau@zurich.ibm.com>", "Michele Dolfi <dol@zurich.ibm.com>", "Maxim Lysak <mly@zurich.ibm.com>", "Nikos Livathinos <nli@zurich.ibm.com>", "Ahmed Nassar <ahn@zurich.ibm.com>", "Peter Staar <taa@zurich.ibm.com>"]
 license = "MIT"
@@ -23,8 +23,8 @@ packages = [{include = "docling"}]
 [tool.poetry.dependencies]
 python = "^3.10"
 pydantic = "^2.0.0"
-docling-core = "^1.1.3"
-docling-ibm-models = "^1.1.3"
+docling-core = "^1.3.0"
+docling-ibm-models = "^1.1.7"
 deepsearch-glm = "^0.21.0"
 filetype = "^1.2.0"
 pypdfium2 = "^4.30.0"
@@ -38,6 +38,30 @@ rtree = "^1.3.0"
 scipy = "^1.14.1"
 pyarrow = "^16.1.0"
 
+#########
+# extras:
+#########
+python-dotenv = { version = "^1.0.1", optional = true }
+llama-index-embeddings-huggingface = { version = "^0.3.1", optional = true }
+llama-index-llms-huggingface-api = { version = "^0.2.0", optional = true }
+llama-index-vector-stores-milvus = { version = "^0.2.1", optional = true }
+langchain-huggingface = { version = "^0.0.3", optional = true}
+langchain-milvus = { version = "^0.1.4", optional = true }
+langchain-text-splitters = { version = "^0.2.4", optional = true }
+
+##############
+# constraints:
+##############
+torch = [
+  {version = "^2.2.2", optional = true, markers = "sys_platform != 'darwin' or platform_machine != 'x86_64'"},
+  {version = "~2.2.2", optional = true, markers = "sys_platform == 'darwin' and platform_machine == 'x86_64'"}
+]
+torchvision = [
+  {version = "^0", optional = true, markers = "sys_platform != 'darwin' or platform_machine != 'x86_64'"},
+  {version = "~0.17.2", optional = true, markers = "sys_platform == 'darwin' and platform_machine == 'x86_64'"}
+]
+typer = "^0.12.5"
+
 [tool.poetry.group.dev.dependencies]
 black = {extras = ["jupyter"], version = "^24.4.2"}
 pytest = "^7.2.2"
@@ -51,11 +75,31 @@ pytest-xdist = "^3.3.1"
 types-requests = "^2.31.0.2"
 flake8-pyproject = "^1.2.3"
 pylint = "^2.17.5"
-
+pandas-stubs = "^2.2.2.240909"
+ipykernel = "^6.29.5"
+ipywidgets = "^8.1.5"
+nbqa = "^1.9.0"
 
 [tool.poetry.group.examples.dependencies]
 datasets = "^2.21.0"
 
+[tool.poetry.extras]
+examples = [
+    "python-dotenv",
+    # LlamaIndex examples:
+    "llama-index-embeddings-huggingface",
+    "llama-index-llms-huggingface-api",
+    "llama-index-vector-stores-milvus",
+    # LangChain examples:
+    "langchain-huggingface",
+    "langchain-milvus",
+    "langchain-text-splitters",
+]
+
+
+[tool.poetry.scripts]
+docling = "docling.cli.main:app"
+
 [build-system]
 requires = ["poetry-core"]
 build-backend = "poetry.core.masonry.api"
@@ -76,6 +120,14 @@ pretty = true
 no_implicit_optional = true
 python_version = "3.10"
 
+[[tool.mypy.overrides]]
+module = [
+    "docling_parse.*",
+    "pypdfium2.*",
+    "networkx.*",
+]
+ignore_missing_imports = true
+
 [tool.flake8]
 max-line-length = 88
 extend-ignore = ["E203", "E501"]