# SPDX-FileCopyrightText: The Docling Contributors # SPDX-License-Identifier: MIT import logging import sys import warnings from enum import Enum from pathlib import Path from typing import Annotated, Optional # Check for CLI dependencies try: import typer from rich.console import Console from rich.logging import RichHandler except ImportError as e: missing_package = str(e).split("'")[1] if "'" in str(e) else "typer or rich" print( f"Error: Missing required CLI dependency '{missing_package}'", file=sys.stderr ) print("\nThe docling-tools CLI requires additional dependencies.", file=sys.stderr) print("Please install them using one of the following options:\n", file=sys.stderr) print(" 1. Install the full docling package (recommended):", file=sys.stderr) print(" pip install docling\n", file=sys.stderr) print(" 2. Install docling-slim with CLI support:", file=sys.stderr) print(" pip install docling-slim[cli]\n", file=sys.stderr) print(" 3. Install just the missing dependencies:", file=sys.stderr) print(" pip install typer rich\n", file=sys.stderr) sys.exit(1) from docling.datamodel.settings import settings from docling.models.stages.ocr.easyocr_model import ( _resolve_easyocr_recognition_models, ) from docling.models.stages.ocr.rapid_ocr_model import _parse_rapidocr_model_spec from docling.models.utils.hf_model_download import download_hf_model from docling.utils.model_downloader import download_models warnings.filterwarnings(action="ignore", category=UserWarning, module="pydantic|torch") warnings.filterwarnings(action="ignore", category=FutureWarning, module="easyocr") console = Console() err_console = Console(stderr=True) app = typer.Typer( name="Docling models helper", no_args_is_help=True, add_completion=False, pretty_exceptions_enable=False, ) class _AvailableModels(str, Enum): LAYOUT = "layout" TABLEFORMER = "tableformer" TABLEFORMERV2 = "tableformerv2" 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" GRANITE_CHART_EXTRACTION = "granite_chart_extraction" GRANITE_CHART_EXTRACTION_V4 = "granite_chart_extraction_v4" RAPIDOCR = "rapidocr" EASYOCR = "easyocr" NEMOTRON_OCR_V2 = "nemotron_ocr_v2" _default_models = [ _AvailableModels.LAYOUT, _AvailableModels.TABLEFORMER, _AvailableModels.CODE_FORMULA, _AvailableModels.PICTURE_CLASSIFIER, _AvailableModels.RAPIDOCR, ] @app.command("download") def download( output_dir: Annotated[ Path, typer.Option( ..., "-o", "--output-dir", help="The directory where to download the models.", ), ] = (settings.cache_dir / "models"), force: Annotated[ bool, typer.Option(..., help="If true, the download will be forced.") ] = False, models: Annotated[ Optional[list[_AvailableModels]], typer.Argument( help="Models to download (default behavior: a predefined set of models will be downloaded).", ), ] = None, all: Annotated[ bool, typer.Option( ..., "--all", help="If true, all available models will be downloaded (mutually exclusive with passing specific models).", show_default=True, ), ] = False, quiet: Annotated[ bool, typer.Option( ..., "-q", "--quiet", help="No extra output is generated, the CLI prints only the directory with the cached models.", ), ] = False, easyocr_lang: Annotated[ Optional[list[str]], typer.Option( ..., "--easyocr-lang", help="EasyOCR language code to prefetch. Repeat for multiple languages.", ), ] = None, rapidocr_backend_lang: Annotated[ Optional[list[str]], typer.Option( ..., "--rapidocr-backend-lang", help=( "RapidOCR checkpoint set to prefetch, as ':' " "(e.g. 'onnxruntime:el', 'torch:korean'). Repeat for multiple. Replaces the default set." ), ), ] = None, ): if models and all: raise typer.BadParameter( "Cannot simultaneously set 'all' parameter and specify models to download." ) if not quiet: logging.basicConfig( level=logging.INFO, format="[blue]%(message)s[/blue]", datefmt="[%X]", handlers=[RichHandler(show_level=False, show_time=False, markup=True)], ) to_download = models or (list(_AvailableModels) if all else _default_models) if easyocr_lang is not None: if _AvailableModels.EASYOCR not in to_download: raise typer.BadParameter( "--easyocr-lang requires the 'easyocr' model", param_hint="--easyocr-lang", ) try: _resolve_easyocr_recognition_models(easyocr_lang) except ValueError as error: raise typer.BadParameter(str(error), param_hint="--easyocr-lang") from error if rapidocr_backend_lang is not None: if _AvailableModels.RAPIDOCR not in to_download: raise typer.BadParameter( "--rapidocr-backend-lang requires the 'rapidocr' model", param_hint="--rapidocr-backend-lang", ) try: for value in rapidocr_backend_lang: _parse_rapidocr_model_spec(value) except ValueError as error: raise typer.BadParameter( str(error), param_hint="--rapidocr-backend-lang" ) from error output_dir = download_models( output_dir=output_dir, force=force, progress=(not quiet), with_layout=_AvailableModels.LAYOUT in to_download, with_tableformer=_AvailableModels.TABLEFORMER in to_download, with_tableformer_v2=_AvailableModels.TABLEFORMERV2 in to_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, with_granite_chart_extraction=_AvailableModels.GRANITE_CHART_EXTRACTION in to_download, with_granite_chart_extraction_v4=_AvailableModels.GRANITE_CHART_EXTRACTION_V4 in to_download, with_rapidocr=_AvailableModels.RAPIDOCR in to_download, rapidocr_models=rapidocr_backend_lang, with_easyocr=_AvailableModels.EASYOCR in to_download, easyocr_languages=easyocr_lang, with_nemotron_ocr=_AvailableModels.NEMOTRON_OCR_V2 in to_download, ) if quiet: typer.echo(output_dir) else: typer.secho(f"\nModels downloaded into: {output_dir}.", fg="green") console.print( "\n", "Docling can now be configured for running offline using the local artifacts.\n\n", "Using the CLI:", f"`docling --artifacts-path={output_dir} FILE`", "\n", "Using Python: see the documentation at .", ) @app.command("download-hf-repo") def download_hf_repo( models: Annotated[ list[str], typer.Argument( help="Specific models to download from HuggingFace identified by their repo id. For example: docling-project/docling-models .", ), ], output_dir: Annotated[ Path, typer.Option( ..., "-o", "--output-dir", help="The directory where to download the models.", ), ] = (settings.cache_dir / "models"), force: Annotated[ bool, typer.Option(..., help="If true, the download will be forced.") ] = False, quiet: Annotated[ bool, typer.Option( ..., "-q", "--quiet", help="No extra output is generated, the CLI prints only the directory with the cached models.", ), ] = False, ): if not quiet: logging.basicConfig( level=logging.INFO, format="[blue]%(message)s[/blue]", datefmt="[%X]", handlers=[RichHandler(show_level=False, show_time=False, markup=True)], ) for item in models: if not quiet: typer.secho(f"\nDownloading {item} model from HuggingFace...") download_hf_model( repo_id=item, # would be better to reuse "repo_cache_folder" property: https://github.com/docling-project/docling/blob/main/docling/datamodel/pipeline_options_vlm_model.py#L76 # but creating options objects seams like an overkill local_dir=output_dir / item.replace("/", "--"), force=force, progress=(not quiet), ) if quiet: typer.echo(output_dir) else: typer.secho(f"\nModels downloaded into: {output_dir}.", fg="green") click_app = typer.main.get_command(app) if __name__ == "__main__": app()