feat: Introduce support for GPU Accelerators (#593)

* Upgraded Layout Postprocessing, sending old code back to ERZ

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

* Implement hierachical cluster layout processing

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

* Pass nested cluster processing through full pipeline

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

* Pass nested clusters through GLM as payload

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

* Move to_docling_document from ds-glm to this repo

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

* Clean up imports again

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

* feat(Accelerator): Introduce options to control the num_threads and device from API, envvars, CLI.
- Introduce the AcceleratorOptions, AcceleratorDevice and use them to set the device where the models run.
- Introduce the accelerator_utils with function to decide the device and resolve the AUTO setting.
- Refactor the way how the docling-ibm-models are called to match the new init signature of models.
- Translate the accelerator options to the specific inputs for third-party models.
- Extend the docling CLI with parameters to set the num_threads and device.
- Add new unit tests.
- Write new example how to use the accelerator options.

* fix: Improve the pydantic objects in the pipeline_options and imports.

Signed-off-by: Nikos Livathinos <nli@zurich.ibm.com>

* fix: TableStructureModel: Refactor the artifacts path to use the new structure for fast/accurate model

Signed-off-by: Nikos Livathinos <nli@zurich.ibm.com>

* Updated test ground-truth

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

* Updated test ground-truth (again), bugfix for empty layout

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

* fix: Do proper check to set the device in EasyOCR, RapidOCR.

Signed-off-by: Nikos Livathinos <nli@zurich.ibm.com>

* Rollback changes from main

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

* Update test gt

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

* Remove unused debug settings

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

* Review fixes

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

* Nail the accelerator defaults for MPS

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

---------

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>
Signed-off-by: Nikos Livathinos <nli@zurich.ibm.com>
Co-authored-by: Christoph Auer <cau@zurich.ibm.com>
Co-authored-by: Christoph Auer <60343111+cau-git@users.noreply.github.com>

docling/datamodel/pipeline_options.py

@@ -1,8 +1,66 @@
+import logging
+import os
+import warnings
 from enum import Enum
 from pathlib import Path
-from typing import List, Literal, Optional, Union
+from typing import Annotated, Any, Dict, List, Literal, Optional, Tuple, Type, Union
 
-from pydantic import BaseModel, ConfigDict, Field
+from pydantic import BaseModel, ConfigDict, Field, field_validator, model_validator
+from pydantic_settings import (
+    BaseSettings,
+    PydanticBaseSettingsSource,
+    SettingsConfigDict,
+)
+from typing_extensions import deprecated
+
+_log = logging.getLogger(__name__)
+
+
+class AcceleratorDevice(str, Enum):
+    """Devices to run model inference"""
+
+    AUTO = "auto"
+    CPU = "cpu"
+    CUDA = "cuda"
+    MPS = "mps"
+
+
+class AcceleratorOptions(BaseSettings):
+    model_config = SettingsConfigDict(
+        env_prefix="DOCLING_", env_nested_delimiter="_", populate_by_name=True
+    )
+
+    num_threads: int = 4
+    device: AcceleratorDevice = AcceleratorDevice.AUTO
+
+    @model_validator(mode="before")
+    @classmethod
+    def check_alternative_envvars(cls, data: Any) -> Any:
+        r"""
+        Set num_threads from the "alternative" envvar OMP_NUM_THREADS.
+        The alternative envvar is used only if it is valid and the regular envvar is not set.
+
+        Notice: The standard pydantic settings mechanism with parameter "aliases" does not provide
+        the same functionality. In case the alias envvar is set and the user tries to override the
+        parameter in settings initialization, Pydantic treats the parameter provided in __init__()
+        as an extra input instead of simply overwriting the evvar value for that parameter.
+        """
+        if isinstance(data, dict):
+            input_num_threads = data.get("num_threads")
+
+            # Check if to set the num_threads from the alternative envvar
+            if input_num_threads is None:
+                docling_num_threads = os.getenv("DOCLING_NUM_THREADS")
+                omp_num_threads = os.getenv("OMP_NUM_THREADS")
+                if docling_num_threads is None and omp_num_threads is not None:
+                    try:
+                        data["num_threads"] = int(omp_num_threads)
+                    except ValueError:
+                        _log.error(
+                            "Ignoring misformatted envvar OMP_NUM_THREADS '%s'",
+                            omp_num_threads,
+                        )
+        return data
 
 
 class TableFormerMode(str, Enum):
@@ -78,9 +136,11 @@ class EasyOcrOptions(OcrOptions):
 
     kind: Literal["easyocr"] = "easyocr"
     lang: List[str] = ["fr", "de", "es", "en"]
-    use_gpu: bool = True  # same default as easyocr.Reader
+
+    use_gpu: Optional[bool] = None
+
     model_storage_directory: Optional[str] = None
-    download_enabled: bool = True  # same default as easyocr.Reader
+    download_enabled: bool = True
 
     model_config = ConfigDict(
         extra="forbid",
@@ -153,6 +213,7 @@ class PipelineOptions(BaseModel):
         True  # This default will be set to False on a future version of docling
     )
     document_timeout: Optional[float] = None
+    accelerator_options: AcceleratorOptions = AcceleratorOptions()
 
 
 class PdfPipelineOptions(PipelineOptions):