docling 2.51.0__tar.gz → 2.53.0__tar.gz

{docling-2.51.0 → docling-2.53.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: docling
-Version: 2.51.0
+Version: 2.53.0
 Summary: SDK and CLI for parsing PDF, DOCX, HTML, and more, to a unified document representation for powering downstream workflows such as gen AI applications.
 Author-email: 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>, Panos Vagenas <pva@zurich.ibm.com>, Peter Staar <taa@zurich.ibm.com>
 License-Expression: MIT
@@ -26,7 +26,7 @@ Requires-Python: <4.0,>=3.9
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: pydantic<3.0.0,>=2.0.0
-Requires-Dist: docling-core[chunking]<3.0.0,>=2.42.0
+Requires-Dist: docling-core[chunking]<3.0.0,>=2.48.0
 Requires-Dist: docling-parse<5.0.0,>=4.4.0
 Requires-Dist: docling-ibm-models<4,>=3.9.1
 Requires-Dist: filetype<2.0.0,>=1.2.0
@@ -108,18 +108,22 @@ Docling simplifies document processing, parsing diverse formats — including ad
 * 🔒 Local execution capabilities for sensitive data and air-gapped environments
 * 🤖 Plug-and-play [integrations][integrations] incl. LangChain, LlamaIndex, Crew AI & Haystack for agentic AI
 * 🔍 Extensive OCR support for scanned PDFs and images
-* 👓 Support of several Visual Language Models ([SmolDocling](https://huggingface.co/ds4sd/SmolDocling-256M-preview))
+* 👓 Support of several Visual Language Models ([GraniteDocling](https://huggingface.co/ibm-granite/granite-docling-258M))
 * 🎙️ Audio support with Automatic Speech Recognition (ASR) models
+* 🔌 Connect to any agent using the [MCP server](https://docling-project.github.io/docling/usage/mcp/)
 * 💻 Simple and convenient CLI
 
 ### What's new
 * 📤 Structured [information extraction][extraction] \[🧪 beta\]
+* 📑 New layout model (**Heron**) by default, for faster PDF parsing
+* 🔌 [MCP server](https://docling-project.github.io/docling/usage/mcp/) for agentic applications
 
 ### Coming soon
 
 * 📝 Metadata extraction, including title, authors, references & language
 * 📝 Chart understanding (Barchart, Piechart, LinePlot, etc)
 * 📝 Complex chemistry understanding (Molecular structures)
+* 📝 Parsing of Web Video Text Tracks (WebVTT) files
 
 ## Installation
 
@@ -145,7 +149,7 @@ result = converter.convert(source)
 print(result.document.export_to_markdown())  # output: "## Docling Technical Report[...]"
 ```
 
-More [advanced usage options](https://docling-project.github.io/docling/usage/) are available in
+More [advanced usage options](https://docling-project.github.io/docling/usage/advanced_options/) are available in
 the docs.
 
 ## CLI
@@ -156,9 +160,9 @@ Docling has a built-in CLI to run conversions.
 docling https://arxiv.org/pdf/2206.01062
 ```
 
-You can also use 🥚[SmolDocling](https://huggingface.co/ds4sd/SmolDocling-256M-preview) and other VLMs via Docling CLI:
+You can also use 🥚[GraniteDocling](https://huggingface.co/ibm-granite/granite-docling-258M) and other VLMs via Docling CLI:
 ```bash
-docling --pipeline vlm --vlm-model smoldocling https://arxiv.org/pdf/2206.01062
+docling --pipeline vlm --vlm-model granite_docling https://arxiv.org/pdf/2206.01062
 ```
 This will use MLX acceleration on supported Apple Silicon hardware.
 

{docling-2.51.0 → docling-2.53.0}/README.md

@@ -36,18 +36,22 @@ Docling simplifies document processing, parsing diverse formats — including ad
 * 🔒 Local execution capabilities for sensitive data and air-gapped environments
 * 🤖 Plug-and-play [integrations][integrations] incl. LangChain, LlamaIndex, Crew AI & Haystack for agentic AI
 * 🔍 Extensive OCR support for scanned PDFs and images
-* 👓 Support of several Visual Language Models ([SmolDocling](https://huggingface.co/ds4sd/SmolDocling-256M-preview))
+* 👓 Support of several Visual Language Models ([GraniteDocling](https://huggingface.co/ibm-granite/granite-docling-258M))
 * 🎙️ Audio support with Automatic Speech Recognition (ASR) models
+* 🔌 Connect to any agent using the [MCP server](https://docling-project.github.io/docling/usage/mcp/)
 * 💻 Simple and convenient CLI
 
 ### What's new
 * 📤 Structured [information extraction][extraction] \[🧪 beta\]
+* 📑 New layout model (**Heron**) by default, for faster PDF parsing
+* 🔌 [MCP server](https://docling-project.github.io/docling/usage/mcp/) for agentic applications
 
 ### Coming soon
 
 * 📝 Metadata extraction, including title, authors, references & language
 * 📝 Chart understanding (Barchart, Piechart, LinePlot, etc)
 * 📝 Complex chemistry understanding (Molecular structures)
+* 📝 Parsing of Web Video Text Tracks (WebVTT) files
 
 ## Installation
 
@@ -73,7 +77,7 @@ result = converter.convert(source)
 print(result.document.export_to_markdown())  # output: "## Docling Technical Report[...]"
 ```
 
-More [advanced usage options](https://docling-project.github.io/docling/usage/) are available in
+More [advanced usage options](https://docling-project.github.io/docling/usage/advanced_options/) are available in
 the docs.
 
 ## CLI
@@ -84,9 +88,9 @@ Docling has a built-in CLI to run conversions.
 docling https://arxiv.org/pdf/2206.01062
 ```
 
-You can also use 🥚[SmolDocling](https://huggingface.co/ds4sd/SmolDocling-256M-preview) and other VLMs via Docling CLI:
+You can also use 🥚[GraniteDocling](https://huggingface.co/ibm-granite/granite-docling-258M) and other VLMs via Docling CLI:
 ```bash
-docling --pipeline vlm --vlm-model smoldocling https://arxiv.org/pdf/2206.01062
+docling --pipeline vlm --vlm-model granite_docling https://arxiv.org/pdf/2206.01062
 ```
 This will use MLX acceleration on supported Apple Silicon hardware.
 

{docling-2.51.0 → docling-2.53.0}/docling/cli/main.py

@@ -48,6 +48,7 @@ from docling.datamodel.base_models import (
 from docling.datamodel.document import ConversionResult
 from docling.datamodel.pipeline_options import (
     AsrPipelineOptions,
+    ConvertPipelineOptions,
     EasyOcrOptions,
     OcrOptions,
     PaginatedPipelineOptions,
@@ -63,6 +64,8 @@ from docling.datamodel.vlm_model_specs import (
     GOT2_TRANSFORMERS,
     GRANITE_VISION_OLLAMA,
     GRANITE_VISION_TRANSFORMERS,
+    GRANITEDOCLING_MLX,
+    GRANITEDOCLING_TRANSFORMERS,
     SMOLDOCLING_MLX,
     SMOLDOCLING_TRANSFORMERS,
     SMOLDOCLING_VLLM,
@@ -71,8 +74,13 @@ from docling.datamodel.vlm_model_specs import (
 from docling.document_converter import (
     AudioFormatOption,
     DocumentConverter,
+    ExcelFormatOption,
     FormatOption,
+    HTMLFormatOption,
+    MarkdownFormatOption,
     PdfFormatOption,
+    PowerpointFormatOption,
+    WordFormatOption,
 )
 from docling.models.factories import get_ocr_factory
 from docling.pipeline.asr_pipeline import AsrPipeline
@@ -328,7 +336,7 @@ def convert(  # noqa: C901
     vlm_model: Annotated[
         VlmModelType,
         typer.Option(..., help="Choose the VLM model to use with PDF or image files."),
-    ] = VlmModelType.SMOLDOCLING,
+    ] = VlmModelType.GRANITEDOCLING,
     asr_model: Annotated[
         AsrModelType,
         typer.Option(..., help="Choose the ASR model to use with audio/video files."),
@@ -626,10 +634,33 @@ def convert(  # noqa: C901
                 backend=MetsGbsDocumentBackend,
             )
 
+            # SimplePipeline options
+            simple_format_option = ConvertPipelineOptions(
+                do_picture_description=enrich_picture_description,
+                do_picture_classification=enrich_picture_classes,
+            )
+            if artifacts_path is not None:
+                simple_format_option.artifacts_path = artifacts_path
+
             format_options = {
                 InputFormat.PDF: pdf_format_option,
                 InputFormat.IMAGE: pdf_format_option,
                 InputFormat.METS_GBS: mets_gbs_format_option,
+                InputFormat.DOCX: WordFormatOption(
+                    pipeline_options=simple_format_option
+                ),
+                InputFormat.PPTX: PowerpointFormatOption(
+                    pipeline_options=simple_format_option
+                ),
+                InputFormat.XLSX: ExcelFormatOption(
+                    pipeline_options=simple_format_option
+                ),
+                InputFormat.HTML: HTMLFormatOption(
+                    pipeline_options=simple_format_option
+                ),
+                InputFormat.MD: MarkdownFormatOption(
+                    pipeline_options=simple_format_option
+                ),
             }
 
         elif pipeline == ProcessingPipeline.VLM:
@@ -655,6 +686,18 @@ def convert(  # noqa: C901
                             "To run SmolDocling faster, please install mlx-vlm:\n"
                             "pip install mlx-vlm"
                         )
+            elif vlm_model == VlmModelType.GRANITEDOCLING:
+                pipeline_options.vlm_options = GRANITEDOCLING_TRANSFORMERS
+                if sys.platform == "darwin":
+                    try:
+                        import mlx_vlm
+
+                        pipeline_options.vlm_options = GRANITEDOCLING_MLX
+                    except ImportError:
+                        _log.warning(
+                            "To run GraniteDocling faster, please install mlx-vlm:\n"
+                            "pip install mlx-vlm"
+                        )
             elif vlm_model == VlmModelType.SMOLDOCLING_VLLM:
                 pipeline_options.vlm_options = SMOLDOCLING_VLLM
 

{docling-2.51.0 → docling-2.53.0}/docling/cli/models.py

@@ -33,6 +33,8 @@ class _AvailableModels(str, Enum):
     CODE_FORMULA = "code_formula"
     PICTURE_CLASSIFIER = "picture_classifier"
     SMOLVLM = "smolvlm"
+    GRANITEDOCLING = "granitedocling"
+    GRANITEDOCLING_MLX = "granitedocling_mlx"
     SMOLDOCLING = "smoldocling"
     SMOLDOCLING_MLX = "smoldocling_mlx"
     GRANITE_VISION = "granite_vision"
@@ -108,6 +110,8 @@ def download(
         with_code_formula=_AvailableModels.CODE_FORMULA in to_download,
         with_picture_classifier=_AvailableModels.PICTURE_CLASSIFIER in to_download,
         with_smolvlm=_AvailableModels.SMOLVLM in to_download,
+        with_granitedocling=_AvailableModels.GRANITEDOCLING in to_download,
+        with_granitedocling_mlx=_AvailableModels.GRANITEDOCLING_MLX in to_download,
         with_smoldocling=_AvailableModels.SMOLDOCLING in to_download,
         with_smoldocling_mlx=_AvailableModels.SMOLDOCLING_MLX in to_download,
         with_granite_vision=_AvailableModels.GRANITE_VISION in to_download,

{docling-2.51.0 → docling-2.53.0}/docling/datamodel/pipeline_options.py

@@ -12,7 +12,7 @@ from pydantic import (
 )
 from typing_extensions import deprecated
 
-from docling.datamodel import asr_model_specs
+from docling.datamodel import asr_model_specs, vlm_model_specs
 
 # Import the following for backwards compatibility
 from docling.datamodel.accelerator_options import AcceleratorDevice, AcceleratorOptions
@@ -114,7 +114,11 @@ class RapidOcrOptions(OcrOptions):
     cls_model_path: Optional[str] = None  # same default as rapidocr
     rec_model_path: Optional[str] = None  # same default as rapidocr
     rec_keys_path: Optional[str] = None  # same default as rapidocr
-    rec_font_path: Optional[str] = None  # same default as rapidocr
+    rec_font_path: Optional[str] = None  # Deprecated, please use font_path instead
+    font_path: Optional[str] = None  # same default as rapidocr
+
+    # Dictionary to overwrite or pass-through additional parameters
+    rapidocr_params: Dict[str, Any] = Field(default_factory=dict)
 
     model_config = ConfigDict(
         extra="forbid",
@@ -135,6 +139,8 @@ class EasyOcrOptions(OcrOptions):
     recog_network: Optional[str] = "standard"
     download_enabled: bool = True
 
+    suppress_mps_warnings: bool = True
+
     model_config = ConfigDict(
         extra="forbid",
         protected_namespaces=(),
@@ -257,11 +263,21 @@ class PipelineOptions(BaseOptions):
     accelerator_options: AcceleratorOptions = AcceleratorOptions()
     enable_remote_services: bool = False
     allow_external_plugins: bool = False
+    artifacts_path: Optional[Union[Path, str]] = None
 
 
-class PaginatedPipelineOptions(PipelineOptions):
-    artifacts_path: Optional[Union[Path, str]] = None
+class ConvertPipelineOptions(PipelineOptions):
+    """Base convert pipeline options."""
+
+    do_picture_classification: bool = False  # True: classify pictures in documents
 
+    do_picture_description: bool = False  # True: run describe pictures in documents
+    picture_description_options: PictureDescriptionBaseOptions = (
+        smolvlm_picture_description
+    )
+
+
+class PaginatedPipelineOptions(ConvertPipelineOptions):
     images_scale: float = 1.0
     generate_page_images: bool = False
     generate_picture_images: bool = False
@@ -274,7 +290,7 @@ class VlmPipelineOptions(PaginatedPipelineOptions):
     )
     # If True, text from backend will be used instead of generated text
     vlm_options: Union[InlineVlmOptions, ApiVlmOptions] = (
-        smoldocling_vlm_conversion_options
+        vlm_model_specs.GRANITEDOCLING_TRANSFORMERS
     )
 
 
@@ -293,13 +309,11 @@ class LayoutOptions(BaseModel):
 
 class AsrPipelineOptions(PipelineOptions):
     asr_options: Union[InlineAsrOptions] = asr_model_specs.WHISPER_TINY
-    artifacts_path: Optional[Union[Path, str]] = None
 
 
 class VlmExtractionPipelineOptions(PipelineOptions):
     """Options for extraction pipeline."""
 
-    artifacts_path: Optional[Union[Path, str]] = None
     vlm_options: Union[InlineVlmOptions] = NU_EXTRACT_2B_TRANSFORMERS
 
 
@@ -310,8 +324,6 @@ class PdfPipelineOptions(PaginatedPipelineOptions):
     do_ocr: bool = True  # True: perform OCR, replace programmatic PDF text
     do_code_enrichment: bool = False  # True: perform code OCR
     do_formula_enrichment: bool = False  # True: perform formula OCR, return Latex code
-    do_picture_classification: bool = False  # True: classify pictures in documents
-    do_picture_description: bool = False  # True: run describe pictures in documents
     force_backend_text: bool = (
         False  # (To be used with vlms, or other generative models)
     )
@@ -319,9 +331,6 @@ class PdfPipelineOptions(PaginatedPipelineOptions):
 
     table_structure_options: TableStructureOptions = TableStructureOptions()
     ocr_options: OcrOptions = EasyOcrOptions()
-    picture_description_options: PictureDescriptionBaseOptions = (
-        smolvlm_picture_description
-    )
     layout_options: LayoutOptions = LayoutOptions()
 
     images_scale: float = 1.0

{docling-2.51.0 → docling-2.53.0}/docling/datamodel/vlm_model_specs.py

@@ -18,6 +18,35 @@ from docling.datamodel.pipeline_options_vlm_model import (
 _log = logging.getLogger(__name__)
 
 
+# Granite-Docling
+GRANITEDOCLING_TRANSFORMERS = InlineVlmOptions(
+    repo_id="ibm-granite/granite-docling-258M",
+    prompt="Convert this page to docling.",
+    response_format=ResponseFormat.DOCTAGS,
+    inference_framework=InferenceFramework.TRANSFORMERS,
+    transformers_model_type=TransformersModelType.AUTOMODEL_IMAGETEXTTOTEXT,
+    supported_devices=[
+        AcceleratorDevice.CPU,
+        AcceleratorDevice.CUDA,
+    ],
+    scale=2.0,
+    temperature=0.0,
+    max_new_tokens=8192,
+    stop_strings=["</doctag>", "<|end_of_text|>"],
+)
+
+GRANITEDOCLING_MLX = InlineVlmOptions(
+    repo_id="ibm-granite/granite-docling-258M-mlx",
+    prompt="Convert this page to docling.",
+    response_format=ResponseFormat.DOCTAGS,
+    inference_framework=InferenceFramework.MLX,
+    supported_devices=[AcceleratorDevice.MPS],
+    scale=2.0,
+    temperature=0.0,
+    max_new_tokens=8192,
+    stop_strings=["</doctag>", "<|end_of_text|>"],
+)
+
 # SmolDocling
 SMOLDOCLING_MLX = InlineVlmOptions(
     repo_id="ds4sd/SmolDocling-256M-preview-mlx-bf16",
@@ -272,3 +301,4 @@ class VlmModelType(str, Enum):
     GRANITE_VISION_VLLM = "granite_vision_vllm"
     GRANITE_VISION_OLLAMA = "granite_vision_ollama"
     GOT_OCR_2 = "got_ocr_2"
+    GRANITEDOCLING = "granite_docling"

{docling-2.51.0 → docling-2.53.0}/docling/models/base_model.py

@@ -4,7 +4,13 @@ from collections.abc import Iterable
 from typing import Any, Generic, Optional, Protocol, Type, Union
 
 import numpy as np
-from docling_core.types.doc import BoundingBox, DocItem, DoclingDocument, NodeItem
+from docling_core.types.doc import (
+    BoundingBox,
+    DocItem,
+    DoclingDocument,
+    NodeItem,
+    PictureItem,
+)
 from PIL.Image import Image
 from typing_extensions import TypeVar
 
@@ -164,8 +170,17 @@ class BaseItemAndImageEnrichmentModel(
             return None
 
         assert isinstance(element, DocItem)
-        element_prov = element.prov[0]
 
+        # Allow the case of documents without page images but embedded images (e.g. Word and HTML docs)
+        if len(element.prov) == 0 and isinstance(element, PictureItem):
+            embedded_im = element.get_image(conv_res.document)
+            if embedded_im is not None:
+                return ItemAndImageEnrichmentElement(item=element, image=embedded_im)
+            else:
+                return None
+
+        # Crop the image form the page
+        element_prov = element.prov[0]
         bbox = element_prov.bbox
         width = bbox.r - bbox.l
         height = bbox.t - bbox.b
@@ -183,4 +198,14 @@ class BaseItemAndImageEnrichmentModel(
         cropped_image = conv_res.pages[page_ix].get_image(
             scale=self.images_scale, cropbox=expanded_bbox
         )
+
+        # Allow for images being embedded without the page backend or page images
+        if cropped_image is None and isinstance(element, PictureItem):
+            embedded_im = element.get_image(conv_res.document)
+            if embedded_im is not None:
+                return ItemAndImageEnrichmentElement(item=element, image=embedded_im)
+            else:
+                return None
+
+        # Return the proper cropped image
         return ItemAndImageEnrichmentElement(item=element, image=cropped_image)

{docling-2.51.0 → docling-2.53.0}/docling/models/easyocr_model.py

@@ -78,14 +78,17 @@ class EasyOcrModel(BaseOcrModel):
                 download_enabled = False
                 model_storage_directory = str(artifacts_path / self._model_repo_folder)
 
-            self.reader = easyocr.Reader(
-                lang_list=self.options.lang,
-                gpu=use_gpu,
-                model_storage_directory=model_storage_directory,
-                recog_network=self.options.recog_network,
-                download_enabled=download_enabled,
-                verbose=False,
-            )
+            with warnings.catch_warnings():
+                if self.options.suppress_mps_warnings:
+                    warnings.filterwarnings("ignore", message=".*pin_memory.*MPS.*")
+                self.reader = easyocr.Reader(
+                    lang_list=self.options.lang,
+                    gpu=use_gpu,
+                    model_storage_directory=model_storage_directory,
+                    recog_network=self.options.recog_network,
+                    download_enabled=download_enabled,
+                    verbose=False,
+                )
 
     @staticmethod
     def download_models(
@@ -147,7 +150,14 @@ class EasyOcrModel(BaseOcrModel):
                             scale=self.scale, cropbox=ocr_rect
                         )
                         im = numpy.array(high_res_image)
-                        result = self.reader.readtext(im)
+
+                        with warnings.catch_warnings():
+                            if self.options.suppress_mps_warnings:
+                                warnings.filterwarnings(
+                                    "ignore", message=".*pin_memory.*MPS.*"
+                                )
+
+                            result = self.reader.readtext(im)
 
                         del high_res_image
                         del im

{docling-2.51.0 → docling-2.53.0}/docling/models/picture_description_vlm_model.py

@@ -67,7 +67,7 @@ class PictureDescriptionVlmModel(
                 self.model = AutoModelForImageTextToText.from_pretrained(
                     artifacts_path,
                     device_map=self.device,
-                    torch_dtype=torch.bfloat16,
+                    dtype=torch.bfloat16,
                     _attn_implementation=(
                         "flash_attention_2"
                         if self.device.startswith("cuda")

{docling-2.51.0 → docling-2.53.0}/docling/models/rapid_ocr_model.py

@@ -62,32 +62,44 @@ class RapidOcrModel(BaseOcrModel):
             }
             backend_enum = _ALIASES.get(self.options.backend, EngineType.ONNXRUNTIME)
 
+            params = {
+                # Global settings (these are still correct)
+                "Global.text_score": self.options.text_score,
+                "Global.font_path": self.options.font_path,
+                # "Global.verbose": self.options.print_verbose,
+                # Detection model settings
+                "Det.model_path": self.options.det_model_path,
+                "Det.use_cuda": use_cuda,
+                "Det.use_dml": use_dml,
+                "Det.intra_op_num_threads": intra_op_num_threads,
+                # Classification model settings
+                "Cls.model_path": self.options.cls_model_path,
+                "Cls.use_cuda": use_cuda,
+                "Cls.use_dml": use_dml,
+                "Cls.intra_op_num_threads": intra_op_num_threads,
+                # Recognition model settings
+                "Rec.model_path": self.options.rec_model_path,
+                "Rec.font_path": self.options.rec_font_path,
+                "Rec.keys_path": self.options.rec_keys_path,
+                "Rec.use_cuda": use_cuda,
+                "Rec.use_dml": use_dml,
+                "Rec.intra_op_num_threads": intra_op_num_threads,
+                "Det.engine_type": backend_enum,
+                "Cls.engine_type": backend_enum,
+                "Rec.engine_type": backend_enum,
+            }
+
+            if self.options.rec_font_path is not None:
+                _log.warning(
+                    "The 'rec_font_path' option for RapidOCR is deprecated. Please use 'font_path' instead."
+                )
+            user_params = self.options.rapidocr_params
+            if user_params:
+                _log.debug("Overwriting RapidOCR params with user-provided values.")
+                params.update(user_params)
+
             self.reader = RapidOCR(
-                params={
-                    # Global settings (these are still correct)
-                    "Global.text_score": self.options.text_score,
-                    # "Global.verbose": self.options.print_verbose,
-                    # Detection model settings
-                    "Det.model_path": self.options.det_model_path,
-                    "Det.use_cuda": use_cuda,
-                    "Det.use_dml": use_dml,
-                    "Det.intra_op_num_threads": intra_op_num_threads,
-                    # Classification model settings
-                    "Cls.model_path": self.options.cls_model_path,
-                    "Cls.use_cuda": use_cuda,
-                    "Cls.use_dml": use_dml,
-                    "Cls.intra_op_num_threads": intra_op_num_threads,
-                    # Recognition model settings
-                    "Rec.model_path": self.options.rec_model_path,
-                    "Rec.font_path": self.options.rec_font_path,
-                    "Rec.keys_path": self.options.rec_keys_path,
-                    "Rec.use_cuda": use_cuda,
-                    "Rec.use_dml": use_dml,
-                    "Rec.intra_op_num_threads": intra_op_num_threads,
-                    "Det.engine_type": backend_enum,
-                    "Cls.engine_type": backend_enum,
-                    "Rec.engine_type": backend_enum,
-                }
+                params=params,
             )
 
     def __call__(
@@ -120,6 +132,9 @@ class RapidOcrModel(BaseOcrModel):
                             use_cls=self.options.use_cls,
                             use_rec=self.options.use_rec,
                         )
+                        if result is None or result.boxes is None:
+                            _log.warning("RapidOCR returned empty result!")
+                            continue
                         result = list(
                             zip(result.boxes.tolist(), result.txts, result.scores)
                         )

{docling-2.51.0 → docling-2.53.0}/docling/models/vlm_models_inline/hf_transformers_model.py

@@ -112,7 +112,7 @@ class HuggingFaceTransformersVlmModel(BaseVlmPageModel, HuggingFaceModelDownload
             self.vlm_model = model_cls.from_pretrained(
                 artifacts_path,
                 device_map=self.device,
-                torch_dtype=self.vlm_options.torch_dtype,
+                dtype=self.vlm_options.torch_dtype,
                 _attn_implementation=(
                     "flash_attention_2"
                     if self.device.startswith("cuda")

{docling-2.51.0 → docling-2.53.0}/docling/models/vlm_models_inline/nuextract_transformers_model.py

@@ -144,7 +144,7 @@ class NuExtractTransformersModel(BaseVlmModel, HuggingFaceModelDownloadMixin):
             self.vlm_model = AutoModelForImageTextToText.from_pretrained(
                 artifacts_path,
                 device_map=self.device,
-                torch_dtype=self.vlm_options.torch_dtype,
+                dtype=self.vlm_options.torch_dtype,
                 _attn_implementation=(
                     "flash_attention_2"
                     if self.device.startswith("cuda")

{docling-2.51.0 → docling-2.53.0}/docling/pipeline/asr_pipeline.py

@@ -208,25 +208,13 @@ class AsrPipeline(BasePipeline):
 
         self.pipeline_options: AsrPipelineOptions = pipeline_options
 
-        artifacts_path: Optional[Path] = None
-        if pipeline_options.artifacts_path is not None:
-            artifacts_path = Path(pipeline_options.artifacts_path).expanduser()
-        elif settings.artifacts_path is not None:
-            artifacts_path = Path(settings.artifacts_path).expanduser()
-
-        if artifacts_path is not None and not artifacts_path.is_dir():
-            raise RuntimeError(
-                f"The value of {artifacts_path=} is not valid. "
-                "When defined, it must point to a folder containing all models required by the pipeline."
-            )
-
         if isinstance(self.pipeline_options.asr_options, InlineAsrNativeWhisperOptions):
             asr_options: InlineAsrNativeWhisperOptions = (
                 self.pipeline_options.asr_options
             )
             self._model = _NativeWhisperModel(
                 enabled=True,  # must be always enabled for this pipeline to make sense.
-                artifacts_path=artifacts_path,
+                artifacts_path=self.artifacts_path,
                 accelerator_options=pipeline_options.accelerator_options,
                 asr_options=asr_options,
             )

{docling-2.51.0 → docling-2.53.0}/docling/pipeline/base_extraction_pipeline.py

@@ -1,19 +1,33 @@
 import logging
 from abc import ABC, abstractmethod
+from pathlib import Path
 from typing import Optional
 
 from docling.datamodel.base_models import ConversionStatus, ErrorItem
 from docling.datamodel.document import InputDocument
 from docling.datamodel.extraction import ExtractionResult, ExtractionTemplateType
-from docling.datamodel.pipeline_options import BaseOptions
+from docling.datamodel.pipeline_options import BaseOptions, PipelineOptions
+from docling.datamodel.settings import settings
 
 _log = logging.getLogger(__name__)
 
 
 class BaseExtractionPipeline(ABC):
-    def __init__(self, pipeline_options: BaseOptions):
+    def __init__(self, pipeline_options: PipelineOptions):
         self.pipeline_options = pipeline_options
 
+        self.artifacts_path: Optional[Path] = None
+        if pipeline_options.artifacts_path is not None:
+            self.artifacts_path = Path(pipeline_options.artifacts_path).expanduser()
+        elif settings.artifacts_path is not None:
+            self.artifacts_path = Path(settings.artifacts_path).expanduser()
+
+        if self.artifacts_path is not None and not self.artifacts_path.is_dir():
+            raise RuntimeError(
+                f"The value of {self.artifacts_path=} is not valid. "
+                "When defined, it must point to a folder containing all models required by the pipeline."
+            )
+
     def execute(
         self,
         in_doc: InputDocument,
@@ -54,5 +68,5 @@ class BaseExtractionPipeline(ABC):
 
     @classmethod
     @abstractmethod
-    def get_default_options(cls) -> BaseOptions:
+    def get_default_options(cls) -> PipelineOptions:
         pass