deepdoctection 0.39.7__tar.gz → 0.41.0__tar.gz

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: deepdoctection
-Version: 0.39.7
+Version: 0.41.0
 Summary: Repository for Document AI
 Home-page: https://github.com/deepdoctection/deepdoctection
 Author: Dr. Janis Meyer

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/__init__.py

@@ -25,7 +25,7 @@ from .utils.logger import LoggingRecord, logger
 
 # pylint: enable=wrong-import-position
 
-__version__ = "0.39.7"
+__version__ = "0.41.0"
 
 _IMPORT_STRUCTURE = {
     "analyzer": ["config_sanity_checks", "get_dd_analyzer", "ServiceFactory"],
@@ -90,8 +90,6 @@ _IMPORT_STRUCTURE = {
         "convert_np_array_to_b64_b",
         "convert_bytes_to_np_array",
         "convert_pdf_bytes_to_np_array_v2",
-        "box_to_point4",
-        "point4_to_box",
         "as_dict",
         "ImageAnnotationBaseView",
         "Image",
@@ -164,6 +162,7 @@ _IMPORT_STRUCTURE = {
         "LMSequenceClassifier",
         "LanguageDetector",
         "ImageTransformer",
+        "DeterministicImageTransformer",
         "InferenceResize",
         "D2FrcnnDetector",
         "D2FrcnnTracingDetector",
@@ -260,6 +259,7 @@ _IMPORT_STRUCTURE = {
         "ImageCroppingService",
         "IntersectionMatcher",
         "NeighbourMatcher",
+        "FamilyCompound",
         "MatchingService",
         "PageParsingService",
         "AnnotationNmsService",
@@ -400,11 +400,14 @@ _IMPORT_STRUCTURE = {
         "get_type",
         "get_tqdm",
         "get_tqdm_default_kwargs",
+        "box_to_point4",
+        "point4_to_box",
         "ResizeTransform",
         "InferenceResize",
         "normalize_image",
         "pad_image",
         "PadTransform",
+        "RotationTransform",
         "delete_keys_from_dict",
         "split_string",
         "string_to_dict",

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/analyzer/_config.py

@@ -72,7 +72,6 @@ cfg.SEGMENTATION.THRESHOLD_COLS = 0.4
 cfg.SEGMENTATION.FULL_TABLE_TILING = True
 cfg.SEGMENTATION.REMOVE_IOU_THRESHOLD_ROWS = 0.001
 cfg.SEGMENTATION.REMOVE_IOU_THRESHOLD_COLS = 0.001
-cfg.SEGMENTATION.CELL_CATEGORY_ID = 12
 cfg.SEGMENTATION.TABLE_NAME = LayoutType.TABLE
 cfg.SEGMENTATION.PUBTABLES_CELL_NAMES = [
     CellType.SPANNING,

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/analyzer/factory.py

@@ -35,13 +35,14 @@ from ..extern.tpdetect import TPFrcnnDetector
 from ..pipe.base import PipelineComponent
 from ..pipe.common import (
     AnnotationNmsService,
+    FamilyCompound,
     IntersectionMatcher,
     MatchingService,
     NeighbourMatcher,
     PageParsingService,
 )
 from ..pipe.doctectionpipe import DoctectionPipe
-from ..pipe.layout import ImageLayoutService
+from ..pipe.layout import ImageLayoutService, skip_if_category_or_service_extracted
 from ..pipe.order import TextOrderService
 from ..pipe.refine import TableSegmentationRefinementService
 from ..pipe.segment import PubtablesSegmentationService, TableSegmentationService
@@ -196,7 +197,7 @@ class ServiceFactory:
             getattr(config.PT, mode).PAD.BOTTOM,
             getattr(config.PT, mode).PAD.LEFT,
         )
-        return PadTransform(top=top, right=right, bottom=bottom, left=left)  #
+        return PadTransform(pad_top=top, pad_right=right, pad_bottom=bottom, pad_left=left)  #
 
     @staticmethod
     def build_padder(config: AttrDict, mode: str) -> PadTransform:
@@ -284,7 +285,6 @@ class ServiceFactory:
         return SubImageLayoutService(
             sub_image_detector=detector,
             sub_image_names=[LayoutType.TABLE, LayoutType.TABLE_ROTATED],
-            category_id_mapping=None,
             detect_result_generator=detect_result_generator,
             padder=padder,
         )
@@ -405,7 +405,6 @@ class ServiceFactory:
                 tile_table_with_items=config.SEGMENTATION.FULL_TABLE_TILING,
                 remove_iou_threshold_rows=config.SEGMENTATION.REMOVE_IOU_THRESHOLD_ROWS,
                 remove_iou_threshold_cols=config.SEGMENTATION.REMOVE_IOU_THRESHOLD_COLS,
-                cell_class_id=config.SEGMENTATION.CELL_CATEGORY_ID,
                 table_name=config.SEGMENTATION.TABLE_NAME,
                 cell_names=config.SEGMENTATION.PUBTABLES_CELL_NAMES,
                 spanning_cell_names=config.SEGMENTATION.PUBTABLES_SPANNING_CELL_NAMES,
@@ -516,6 +515,15 @@ class ServiceFactory:
         """
         return ServiceFactory._build_pdf_miner_text_service(detector)
 
+    @staticmethod
+    def _build_doctr_word_detector_service(detector: DoctrTextlineDetector) -> ImageLayoutService:
+        """Building a Doctr word detector service
+
+        :param detector: DoctrTextlineDetector
+        :return: ImageLayoutService
+        """
+        return ImageLayoutService(layout_detector=detector, to_image=True, crop_image=True)
+
     @staticmethod
     def build_doctr_word_detector_service(detector: DoctrTextlineDetector) -> ImageLayoutService:
         """Building a Doctr word detector service
@@ -523,9 +531,7 @@ class ServiceFactory:
         :param detector: DoctrTextlineDetector
         :return: ImageLayoutService
         """
-        return ImageLayoutService(
-            layout_detector=detector, to_image=True, crop_image=True, skip_if_layout_extracted=True
-        )
+        return ServiceFactory._build_doctr_word_detector_service(detector)
 
     @staticmethod
     def _build_text_extraction_service(
@@ -539,7 +545,6 @@ class ServiceFactory:
         """
         return TextExtractionService(
             detector,
-            skip_if_text_extracted=config.USE_PDF_MINER,
             extract_from_roi=config.TEXT_CONTAINER if config.OCR.USE_DOCTR else None,
         )
 
@@ -567,11 +572,16 @@ class ServiceFactory:
             threshold=config.WORD_MATCHING.THRESHOLD,
             max_parent_only=config.WORD_MATCHING.MAX_PARENT_ONLY,
         )
+        family_compounds = [
+            FamilyCompound(
+                parent_categories=config.WORD_MATCHING.PARENTAL_CATEGORIES,
+                child_categories=config.TEXT_CONTAINER,
+                relationship_key=Relationships.CHILD,
+            )
+        ]
         return MatchingService(
-            parent_categories=config.WORD_MATCHING.PARENTAL_CATEGORIES,
-            child_categories=config.TEXT_CONTAINER,
+            family_compounds=family_compounds,
             matcher=matcher,
-            relationship_key=Relationships.CHILD,
         )
 
     @staticmethod
@@ -591,11 +601,16 @@ class ServiceFactory:
         :return: MatchingService
         """
         neighbor_matcher = NeighbourMatcher()
+        family_compounds = [
+            FamilyCompound(
+                parent_categories=config.LAYOUT_LINK.PARENTAL_CATEGORIES,
+                child_categories=config.LAYOUT_LINK.CHILD_CATEGORIES,
+                relationship_key=Relationships.LAYOUT_LINK,
+            )
+        ]
         return MatchingService(
-            parent_categories=config.LAYOUT_LINK.PARENTAL_CATEGORIES,
-            child_categories=config.LAYOUT_LINK.CHILD_CATEGORIES,
+            family_compounds=family_compounds,
             matcher=neighbor_matcher,
-            relationship_key=Relationships.LAYOUT_LINK,
         )
 
     @staticmethod
@@ -699,9 +714,11 @@ class ServiceFactory:
                 table_refinement_service = ServiceFactory.build_table_refinement_service(config)
                 pipe_component_list.append(table_refinement_service)
 
+        d_text_service_id = ""
         if config.USE_PDF_MINER:
             pdf_miner = ServiceFactory.build_pdf_text_detector(config)
             d_text = ServiceFactory.build_pdf_miner_text_service(pdf_miner)
+            d_text_service_id = d_text.service_id
             pipe_component_list.append(d_text)
 
         # setup ocr
@@ -710,10 +727,14 @@ class ServiceFactory:
             if config.OCR.USE_DOCTR:
                 word_detector = ServiceFactory.build_doctr_word_detector(config)
                 word_service = ServiceFactory.build_doctr_word_detector_service(word_detector)
+                word_service.set_inbound_filter(skip_if_category_or_service_extracted(service_ids=d_text_service_id))
                 pipe_component_list.append(word_service)
 
             ocr_detector = ServiceFactory.build_ocr_detector(config)
             text_extraction_service = ServiceFactory.build_text_extraction_service(config, ocr_detector)
+            text_extraction_service.set_inbound_filter(
+                skip_if_category_or_service_extracted(service_ids=d_text_service_id)
+            )
             pipe_component_list.append(text_extraction_service)
 
         if config.USE_PDF_MINER or config.USE_OCR:

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/datapoint/convert.py

@@ -27,7 +27,6 @@ from typing import Any, Optional, Union, no_type_check
 
 import numpy as np
 from numpy import uint8
-from numpy.typing import NDArray
 from pypdf import PdfReader
 
 from ..utils.develop import deprecated
@@ -42,8 +41,6 @@ __all__ = [
     "convert_np_array_to_b64_b",
     "convert_bytes_to_np_array",
     "convert_pdf_bytes_to_np_array_v2",
-    "box_to_point4",
-    "point4_to_box",
     "as_dict",
 ]
 
@@ -187,24 +184,3 @@ def convert_pdf_bytes_to_np_array_v2(
             width = shape[2] - shape[0]
         return pdf_to_np_array(pdf_bytes, size=(int(width), int(height)))  # type: ignore
     return pdf_to_np_array(pdf_bytes, dpi=dpi)
-
-
-def box_to_point4(boxes: NDArray[np.float32]) -> NDArray[np.float32]:
-    """
-    :param boxes: nx4
-    :return: (nx4)x2
-    """
-    box = boxes[:, [0, 1, 2, 3, 0, 3, 2, 1]]
-    box = box.reshape((-1, 2))
-    return box
-
-
-def point4_to_box(points: NDArray[np.float32]) -> NDArray[np.float32]:
-    """
-    :param points: (nx4)x2
-    :return: nx4 boxes (x1y1x2y2)
-    """
-    points = points.reshape((-1, 4, 2))
-    min_xy = points.min(axis=1)  # nx2
-    max_xy = points.max(axis=1)  # nx2
-    return np.concatenate((min_xy, max_xy), axis=1)

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/datapoint/image.py

@@ -342,7 +342,7 @@ class Image:
         self,
         category_names: Optional[Union[str, ObjectTypes, Sequence[Union[str, ObjectTypes]]]] = None,
         annotation_ids: Optional[Union[str, Sequence[str]]] = None,
-        service_id: Optional[Union[str, Sequence[str]]] = None,
+        service_ids: Optional[Union[str, Sequence[str]]] = None,
         model_id: Optional[Union[str, Sequence[str]]] = None,
         session_ids: Optional[Union[str, Sequence[str]]] = None,
         ignore_inactive: bool = True,
@@ -356,7 +356,7 @@ class Image:
 
         :param category_names: A single name or list of names
         :param annotation_ids: A single id or list of ids
-        :param service_id: A single service name or list of service names
+        :param service_ids: A single service name or list of service names
         :param model_id: A single model name or list of model names
         :param session_ids: A single session id or list of session ids
         :param ignore_inactive: If set to `True` only active annotations are returned.
@@ -372,7 +372,7 @@ class Image:
             )
 
         ann_ids = [annotation_ids] if isinstance(annotation_ids, str) else annotation_ids
-        service_id = [service_id] if isinstance(service_id, str) else service_id
+        service_ids = [service_ids] if isinstance(service_ids, str) else service_ids
         model_id = [model_id] if isinstance(model_id, str) else model_id
         session_id = [session_ids] if isinstance(session_ids, str) else session_ids
 
@@ -387,8 +387,8 @@ class Image:
         if ann_ids is not None:
             anns = filter(lambda x: x.annotation_id in ann_ids, anns)
 
-        if service_id is not None:
-            anns = filter(lambda x: x.service_id in service_id, anns)
+        if service_ids is not None:
+            anns = filter(lambda x: x.service_id in service_ids, anns)
 
         if model_id is not None:
             anns = filter(lambda x: x.model_id in model_id, anns)

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/datapoint/view.py

@@ -41,12 +41,11 @@ from ..utils.settings import (
     WordType,
     get_type,
 )
-from ..utils.transform import ResizeTransform
+from ..utils.transform import ResizeTransform, box_to_point4, point4_to_box
 from ..utils.types import HTML, AnnotationDict, Chunks, ImageDict, PathLikeOrStr, PixelValues, Text_, csv
 from ..utils.viz import draw_boxes, interactive_imshow, viz_handler
 from .annotation import CategoryAnnotation, ContainerAnnotation, ImageAnnotation, ann_from_dict
 from .box import BoundingBox, crop_box_from_image
-from .convert import box_to_point4, point4_to_box
 from .image import Image
 
 
@@ -659,7 +658,7 @@ class Page(Image):
         self,
         category_names: Optional[Union[str, ObjectTypes, Sequence[Union[str, ObjectTypes]]]] = None,
         annotation_ids: Optional[Union[str, Sequence[str]]] = None,
-        service_id: Optional[Union[str, Sequence[str]]] = None,
+        service_ids: Optional[Union[str, Sequence[str]]] = None,
         model_id: Optional[Union[str, Sequence[str]]] = None,
         session_ids: Optional[Union[str, Sequence[str]]] = None,
         ignore_inactive: bool = True,
@@ -676,7 +675,7 @@ class Page(Image):
 
         :param category_names: A single name or list of names
         :param annotation_ids: A single id or list of ids
-        :param service_id: A single service name or list of service names
+        :param service_ids: A single service name or list of service names
         :param model_id: A single model name or list of model names
         :param session_ids: A single session id or list of session ids
         :param ignore_inactive: If set to `True` only active annotations are returned.
@@ -691,7 +690,7 @@ class Page(Image):
                 else tuple(get_type(cat_name) for cat_name in category_names)
             )
         ann_ids = [annotation_ids] if isinstance(annotation_ids, str) else annotation_ids
-        service_id = [service_id] if isinstance(service_id, str) else service_id
+        service_ids = [service_ids] if isinstance(service_ids, str) else service_ids
         model_id = [model_id] if isinstance(model_id, str) else model_id
         session_id = [session_ids] if isinstance(session_ids, str) else session_ids
 
@@ -706,8 +705,8 @@ class Page(Image):
         if ann_ids is not None:
             anns = filter(lambda x: x.annotation_id in ann_ids, anns)
 
-        if service_id is not None:
-            anns = filter(lambda x: x.generating_service in service_id, anns)
+        if service_ids is not None:
+            anns = filter(lambda x: x.generating_service in service_ids, anns)
 
         if model_id is not None:
             anns = filter(lambda x: x.generating_model in model_id, anns)

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/datasets/base.py

@@ -369,7 +369,9 @@ class MergeDataset(DatasetBase):
         self.buffer_datasets(**dataflow_build_kwargs)
         split_defaultdict = defaultdict(list)
         for image in self.datapoint_list:  # type: ignore
-            split_defaultdict[ann_id_to_split[image.image_id]].append(image)
+            maybe_image_id = ann_id_to_split.get(image.image_id)
+            if maybe_image_id is not None:
+                split_defaultdict[maybe_image_id].append(image)
         train_dataset = split_defaultdict["train"]
         val_dataset = split_defaultdict["val"]
         test_dataset = split_defaultdict["test"]

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/extern/base.py

@@ -26,6 +26,7 @@ from dataclasses import dataclass, field
 from types import MappingProxyType
 from typing import TYPE_CHECKING, Any, Literal, Mapping, Optional, Sequence, Union, overload
 
+import numpy as np
 from lazy_imports import try_import
 
 from ..utils.identifier import get_uuid_from_str
@@ -38,6 +39,7 @@ from ..utils.settings import (
     token_class_tag_to_token_class_with_tag,
     token_class_with_tag_to_token_class_and_tag,
 )
+from ..utils.transform import BaseTransform, box_to_point4, point4_to_box
 from ..utils.types import JsonDict, PixelValues, Requirement
 
 if TYPE_CHECKING:
@@ -621,7 +623,7 @@ class ImageTransformer(PredictorBase, ABC):
     """
 
     @abstractmethod
-    def transform(self, np_img: PixelValues, specification: DetectionResult) -> PixelValues:
+    def transform_image(self, np_img: PixelValues, specification: DetectionResult) -> PixelValues:
         """
         Abstract method transform
         """
@@ -641,3 +643,108 @@ class ImageTransformer(PredictorBase, ABC):
     def get_category_names(self) -> tuple[ObjectTypes, ...]:
         """returns category names"""
         raise NotImplementedError()
+
+    def transform_coords(self, detect_results: Sequence[DetectionResult]) -> Sequence[DetectionResult]:
+        """
+        Transform coordinates aligned with the transform_image method.
+
+        :param detect_results: List of DetectionResults
+        :return: List of DetectionResults. If you pass uuid it is possible to track the transformed bounding boxes.
+        """
+
+        raise NotImplementedError()
+
+    def inverse_transform_coords(self, detect_results: Sequence[DetectionResult]) -> Sequence[DetectionResult]:
+        """
+        Inverse transform coordinates aligned with the transform_image method. Composing transform_coords with
+        inverse_transform_coords should return the original coordinates.
+
+        :param detect_results: List of DetectionResults
+        :return: List of DetectionResults. If you pass uuid it is possible to track the transformed bounding boxes.
+        """
+
+        raise NotImplementedError()
+
+
+class DeterministicImageTransformer(ImageTransformer):
+    """
+    A wrapper for BaseTransform classes that implements the ImageTransformer interface.
+
+    This class provides a bridge between the BaseTransform system (which handles image and coordinate
+    transformations like rotation, padding, etc.) and the predictors framework by implementing the
+    ImageTransformer interface. It allows BaseTransform objects to be used within pipelines that
+    expect ImageTransformer components.
+
+    The transformer performs deterministic transformations on images and their associated coordinates,
+    enabling operations like padding, rotation, and other geometric transformations while maintaining
+    the relationship between image content and annotation coordinates.
+
+    :param base_transform: A BaseTransform instance that defines the actual transformation operations
+                          to be applied to images and coordinates.
+    """
+
+    def __init__(self, base_transform: BaseTransform):
+        """
+        Initialize the DeterministicImageTransformer with a BaseTransform instance.
+
+        :param base_transform: A BaseTransform instance that defines the actual transformation operations
+        """
+        self.base_transform = base_transform
+        self.name = base_transform.__class__.__name__
+        self.model_id = self.get_model_id()
+
+    def transform_image(self, np_img: PixelValues, specification: DetectionResult) -> PixelValues:
+        return self.base_transform.apply_image(np_img)
+
+    def transform_coords(self, detect_results: Sequence[DetectionResult]) -> Sequence[DetectionResult]:
+        boxes = np.array([detect_result.box for detect_result in detect_results])
+        # boxes = box_to_point4(boxes)
+        boxes = self.base_transform.apply_coords(boxes)
+        # boxes = point4_to_box(boxes)
+        detection_results = []
+        for idx, detect_result in enumerate(detect_results):
+            detection_results.append(
+                DetectionResult(
+                    box=boxes[idx, :].tolist(),
+                    class_name=detect_result.class_name,
+                    class_id=detect_result.class_id,
+                    score=detect_result.score,
+                    absolute_coords=detect_result.absolute_coords,
+                    uuid=detect_result.uuid,
+                )
+            )
+        return detection_results
+
+    def inverse_transform_coords(self, detect_results: Sequence[DetectionResult]) -> Sequence[DetectionResult]:
+        boxes = np.array([detect_result.box for detect_result in detect_results])
+        boxes = box_to_point4(boxes)
+        boxes = self.base_transform.inverse_apply_coords(boxes)
+        boxes = point4_to_box(boxes)
+        detection_results = []
+        for idx, detect_result in enumerate(detect_results):
+            detection_results.append(
+                DetectionResult(
+                    box=boxes[idx, :].tolist(),
+                    class_id=detect_result.class_id,
+                    score=detect_result.score,
+                    absolute_coords=detect_result.absolute_coords,
+                    uuid=detect_result.uuid,
+                )
+            )
+        return detection_results
+
+    def clone(self) -> DeterministicImageTransformer:
+        return self.__class__(self.base_transform)
+
+    def predict(self, np_img: PixelValues) -> DetectionResult:
+        detect_result = DetectionResult()
+        for init_arg in self.base_transform.get_init_args():
+            setattr(detect_result, init_arg, getattr(self.base_transform, init_arg))
+        return detect_result
+
+    def get_category_names(self) -> tuple[ObjectTypes, ...]:
+        return self.base_transform.get_category_names()
+
+    @classmethod
+    def get_requirements(cls) -> list[Requirement]:
+        return []

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/extern/deskew.py

@@ -43,7 +43,7 @@ class Jdeskewer(ImageTransformer):
         self.model_id = self.get_model_id()
         self.min_angle_rotation = min_angle_rotation
 
-    def transform(self, np_img: PixelValues, specification: DetectionResult) -> PixelValues:
+    def transform_image(self, np_img: PixelValues, specification: DetectionResult) -> PixelValues:
         """
         Rotation of the image according to the angle determined by the jdeskew estimator.
 

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/extern/doctrocr.py

@@ -514,8 +514,9 @@ class DocTrRotationTransformer(ImageTransformer):
         self.number_contours = number_contours
         self.ratio_threshold_for_lines = ratio_threshold_for_lines
         self.name = "doctr_rotation_transformer"
+        self.model_id = self.get_model_id()
 
-    def transform(self, np_img: PixelValues, specification: DetectionResult) -> PixelValues:
+    def transform_image(self, np_img: PixelValues, specification: DetectionResult) -> PixelValues:
         """
         Applies the predicted rotation to the image, effectively rotating the image backwards.
         This method uses either the Pillow library or OpenCV for the rotation operation, depending on the configuration.

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/extern/tessocr.py

@@ -423,7 +423,7 @@ class TesseractRotationTransformer(ImageTransformer):
         self.categories = ModelCategories(init_categories={1: PageType.ANGLE})
         self.model_id = self.get_model_id()
 
-    def transform(self, np_img: PixelValues, specification: DetectionResult) -> PixelValues:
+    def transform_image(self, np_img: PixelValues, specification: DetectionResult) -> PixelValues:
         """
         Applies the predicted rotation to the image, effectively rotating the image backwards.
         This method uses either the Pillow library or OpenCV for the rotation operation, depending on the configuration.

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/extern/tp/tpfrcnn/preproc.py

@@ -15,9 +15,9 @@ from typing import Any, List, Optional, Tuple
 import numpy as np
 from lazy_imports import try_import
 
-from ....datapoint.convert import box_to_point4, point4_to_box
 from ....utils.error import MalformedData
 from ....utils.logger import log_once
+from ....utils.transform import box_to_point4, point4_to_box
 from ....utils.types import JsonDict, PixelValues
 from .common import filter_boxes_inside_shape, np_iou
 from .modeling.model_fpn import get_all_anchors_fpn

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/mapper/laylmstruct.py

@@ -31,11 +31,10 @@ import numpy.typing as npt
 from lazy_imports import try_import
 
 from ..datapoint.annotation import ContainerAnnotation
-from ..datapoint.convert import box_to_point4, point4_to_box
 from ..datapoint.image import Image
 from ..datapoint.view import Page
 from ..utils.settings import DatasetType, LayoutType, PageType, Relationships, WordType
-from ..utils.transform import ResizeTransform, normalize_image
+from ..utils.transform import ResizeTransform, box_to_point4, normalize_image, point4_to_box
 from ..utils.types import JsonDict
 from .maputils import curry
 

{deepdoctection-0.39.7 → deepdoctection-0.41.0}/deepdoctection/mapper/match.py

@@ -34,13 +34,15 @@ from ..utils.settings import TypeOrStr
 
 def match_anns_by_intersection(
     dp: Image,
-    parent_ann_category_names: Union[TypeOrStr, Sequence[TypeOrStr]],
-    child_ann_category_names: Union[TypeOrStr, Sequence[TypeOrStr]],
     matching_rule: Literal["iou", "ioa"],
     threshold: float,
     use_weighted_intersections: bool = False,
+    parent_ann_category_names: Optional[Union[TypeOrStr, Sequence[TypeOrStr]]] = None,
+    child_ann_category_names: Optional[Union[TypeOrStr, Sequence[TypeOrStr]]] = None,
     parent_ann_ids: Optional[Union[Sequence[str], str]] = None,
     child_ann_ids: Optional[Union[str, Sequence[str]]] = None,
+    parent_ann_service_ids: Optional[Union[str, Sequence[str]]] = None,
+    child_ann_service_ids: Optional[Union[str, Sequence[str]]] = None,
     max_parent_only: bool = False,
 ) -> tuple[Any, Any, Sequence[ImageAnnotation], Sequence[ImageAnnotation]]:
     """
@@ -87,13 +89,19 @@ def match_anns_by_intersection(
                            dates which are not in the list.
     :param child_ann_ids: Additional filter condition. If some ids are selected, it will ignore all other children
                           candidates which are not in the list.
+    :param parent_ann_service_ids: Additional filter condition. If some ids are selected, it will ignore all other
+                                   parent candidates which are not in the list.
+    :param child_ann_service_ids: Additional filter condition. If some ids are selected, it will ignore all other
+                                  children candidates which are not in the list.
     :param max_parent_only: Will assign to each child at most one parent with maximum ioa
     :return: child indices, parent indices (see Example), list of parent ids and list of children ids.
     """
 
     assert matching_rule in ["iou", "ioa"], "matching rule must be either iou or ioa"
 
-    child_anns = dp.get_annotation(annotation_ids=child_ann_ids, category_names=child_ann_category_names)
+    child_anns = dp.get_annotation(
+        annotation_ids=child_ann_ids, category_names=child_ann_category_names, service_ids=child_ann_service_ids
+    )
     child_ann_boxes = np.array(
         [
             ann.get_bounding_box(dp.image_id).transform(dp.width, dp.height, absolute_coords=True).to_list(mode="xyxy")
@@ -101,7 +109,9 @@ def match_anns_by_intersection(
         ]
     )
 
-    parent_anns = dp.get_annotation(annotation_ids=parent_ann_ids, category_names=parent_ann_category_names)
+    parent_anns = dp.get_annotation(
+        annotation_ids=parent_ann_ids, category_names=parent_ann_category_names, service_ids=parent_ann_service_ids
+    )
     parent_ann_boxes = np.array(
         [
             ann.get_bounding_box(dp.image_id).transform(dp.width, dp.height, absolute_coords=True).to_list(mode="xyxy")
@@ -147,10 +157,12 @@ def match_anns_by_intersection(
 
 def match_anns_by_distance(
     dp: Image,
-    parent_ann_category_names: Union[TypeOrStr, Sequence[TypeOrStr]],
-    child_ann_category_names: Union[TypeOrStr, Sequence[TypeOrStr]],
+    parent_ann_category_names: Optional[Union[TypeOrStr, Sequence[TypeOrStr]]] = None,
+    child_ann_category_names: Optional[Union[TypeOrStr, Sequence[TypeOrStr]]] = None,
     parent_ann_ids: Optional[Union[Sequence[str], str]] = None,
     child_ann_ids: Optional[Union[str, Sequence[str]]] = None,
+    parent_ann_service_ids: Optional[Union[str, Sequence[str]]] = None,
+    child_ann_service_ids: Optional[Union[str, Sequence[str]]] = None,
 ) -> list[tuple[ImageAnnotation, ImageAnnotation]]:
     """
     Generates pairs of parent and child annotations by calculating the euclidean distance between the centers of the
@@ -164,11 +176,19 @@ def match_anns_by_distance(
                            dates which are not in the list.
     :param child_ann_ids: Additional filter condition. If some ids are selected, it will ignore all other children
                           candidates which are not in the list.
+    :param parent_ann_service_ids: Additional filter condition. If some ids are selected, it will ignore all other
+                                   parent candidates which are not in the list.
+    :param child_ann_service_ids: Additional filter condition. If some ids are selected, it will ignore all other
+                                  children candidates which are not in the list.
     :return:
     """
 
-    parent_anns = dp.get_annotation(annotation_ids=parent_ann_ids, category_names=parent_ann_category_names)
-    child_anns = dp.get_annotation(annotation_ids=child_ann_ids, category_names=child_ann_category_names)
+    parent_anns = dp.get_annotation(
+        annotation_ids=parent_ann_ids, category_names=parent_ann_category_names, service_ids=parent_ann_service_ids
+    )
+    child_anns = dp.get_annotation(
+        annotation_ids=child_ann_ids, category_names=child_ann_category_names, service_ids=child_ann_service_ids
+    )
     child_centers = [block.get_bounding_box(dp.image_id).center for block in child_anns]
     parent_centers = [block.get_bounding_box(dp.image_id).center for block in parent_anns]
     if child_centers and parent_centers: