deepdoctection 0.36__py3-none-any.whl → 0.37.1__py3-none-any.whl

deepdoctection/__init__.py

@@ -24,7 +24,7 @@ from .utils.logger import LoggingRecord, logger
 
 # pylint: enable=wrong-import-position
 
-__version__ = 0.36
+__version__ = "0.37.1"
 
 _IMPORT_STRUCTURE = {
     "analyzer": ["config_sanity_checks", "get_dd_analyzer", "ServiceFactory"],
@@ -87,6 +87,7 @@ _IMPORT_STRUCTURE = {
         "convert_b64_to_np_array",
         "convert_np_array_to_b64",
         "convert_np_array_to_b64_b",
+        "convert_bytes_to_np_array",
         "convert_pdf_bytes_to_np_array_v2",
         "box_to_point4",
         "point4_to_box",
@@ -371,6 +372,7 @@ _IMPORT_STRUCTURE = {
         "save_config_to_yaml",
         "config_to_cli_str",
         "decrypt_pdf_document",
+        "decrypt_pdf_document_from_bytes",
         "get_pdf_file_reader",
         "get_pdf_file_writer",
         "PDFStreamer",

deepdoctection/analyzer/factory.py

@@ -327,9 +327,9 @@ class ServiceFactory:
             )
         if config.OCR.USE_TEXTRACT:
             credentials_kwargs = {
-                "aws_access_key_id": environ.get("ACCESS_KEY", None),
-                "aws_secret_access_key": environ.get("SECRET_KEY", None),
-                "config": Config(region_name=environ.get("REGION", None)),
+                "aws_access_key_id": environ.get("AWS_ACCESS_KEY", None),
+                "aws_secret_access_key": environ.get("AWS_SECRET_KEY", None),
+                "config": Config(region_name=environ.get("AWS_REGION", None)),
             }
             return TextractOcrDetector(**credentials_kwargs)
         raise ValueError("You have set USE_OCR=True but any of USE_TESSERACT, USE_DOCTR, USE_TEXTRACT is set to False")

deepdoctection/dataflow/custom_serialize.py

@@ -593,7 +593,7 @@ class SerializerPdfDoc:
         file_name = os.path.split(path)[1]
         prefix, suffix = os.path.splitext(file_name)
         df: DataFlow
-        df = CustomDataFromIterable(PDFStreamer(path=path), max_datapoints=max_datapoints)
+        df = CustomDataFromIterable(PDFStreamer(path_or_bytes=path), max_datapoints=max_datapoints)
         df = MapData(
             df,
             lambda dp: {

deepdoctection/datapoint/convert.py

@@ -40,6 +40,7 @@ __all__ = [
     "convert_b64_to_np_array",
     "convert_np_array_to_b64",
     "convert_np_array_to_b64_b",
+    "convert_bytes_to_np_array",
     "convert_pdf_bytes_to_np_array_v2",
     "box_to_point4",
     "point4_to_box",
@@ -107,6 +108,16 @@ def convert_np_array_to_b64_b(np_image: PixelValues) -> bytes:
     return viz_handler.encode(np_image)
 
 
+def convert_bytes_to_np_array(image_bytes: bytes) -> PixelValues:
+    """
+    Converts an image in bytes to a numpy array
+
+    :param image_bytes: An image as bytes.
+    :return: numpy array.
+    """
+    return viz_handler.convert_bytes_to_np(image_bytes)
+
+
 @deprecated("Use convert_pdf_bytes_to_np_array_v2", "2022-02-23")
 def convert_pdf_bytes_to_np_array(pdf_bytes: bytes, dpi: Optional[int] = None) -> PixelValues:
     """

deepdoctection/datapoint/image.py

@@ -34,6 +34,7 @@ from ..utils.error import AnnotationError, BoundingBoxError, ImageError, UUIDErr
 from ..utils.identifier import get_uuid, is_uuid_like
 from ..utils.settings import ObjectTypes, SummaryType, get_type
 from ..utils.types import ImageDict, PathLikeOrStr, PixelValues
+from ..utils.logger import LoggingRecord, logger
 from .annotation import Annotation, AnnotationMap, BoundingBox, CategoryAnnotation, ImageAnnotation
 from .box import crop_box_from_image, global_to_local_coords, intersection_box
 from .convert import as_dict, convert_b64_to_np_array, convert_np_array_to_b64, convert_pdf_bytes_to_np_array_v2
@@ -474,8 +475,11 @@ class Image:
 
             for service_id in service_ids:
                 if service_id not in service_id_to_annotation_id:
-                    raise ImageError(f"Service id {service_id} not found")
-                annotation_ids = service_id_to_annotation_id[service_id]
+                    logger.info(
+                        LoggingRecord(
+                            f"Service_id {service_id} for image_id: {self.image_id} not found. Skipping removal."))
+
+                annotation_ids = service_id_to_annotation_id.get(service_id, [])
 
                 for ann_id in annotation_ids:
                     if ann_id not in ann_id_to_annotation_maps:
@@ -587,7 +591,7 @@ class Image:
             )
             ann.image.dump(sub_image)
 
-    def remove_image_from_lower_hierachy(self, pixel_values_only: bool = False) -> None:
+    def remove_image_from_lower_hierarchy(self, pixel_values_only: bool = False) -> None:
         """Will remove all images from image annotations."""
         for ann in self.annotations:
             if pixel_values_only:
@@ -717,7 +721,7 @@ class Image:
         else:
             path_json = fspath(path) + ".json"
         if highest_hierarchy_only:
-            self.remove_image_from_lower_hierachy()
+            self.remove_image_from_lower_hierarchy()
         export_dict = self.as_dict()
         export_dict["location"] = fspath(export_dict["location"])
         if not image_to_json:
@@ -747,7 +751,7 @@ class Image:
                 if sub_cat.service_id:
                     service_id_dict[sub_cat.service_id].append(sub_cat.annotation_id)
             if ann.image is not None:
-                for summary_cat_key in ann.image.summary:
+                for summary_cat_key in ann.image.summary.sub_categories:
                     summary_cat = ann.get_summary(summary_cat_key)
                     if summary_cat.service_id:
                         service_id_dict[summary_cat.service_id].append(summary_cat.annotation_id)

deepdoctection/datasets/save.py

@@ -62,7 +62,7 @@ def dataflow_to_json(
     if highest_hierarchy_only:
 
         def _remove_hh(dp: Image) -> Image:
-            dp.remove_image_from_lower_hierachy()
+            dp.remove_image_from_lower_hierarchy()
             return dp
 
         df = MapData(df, _remove_hh)

deepdoctection/extern/base.py

@@ -69,8 +69,7 @@ class ModelCategories:
         if self.init_categories:
             self._init_categories = MappingProxyType({key: get_type(val) for key, val in self.init_categories.items()})
         else:
-            if self._init_categories is None:
-                self._init_categories = MappingProxyType({})
+            self._init_categories = MappingProxyType({})
         self.categories = self._init_categories
 
     @overload
@@ -181,7 +180,7 @@ class NerModelCategories(ModelCategories):
             self._init_categories = self.merge_bio_semantics_categories(
                 self._categories_semantics, self._categories_bio
             )
-        super().__post_init__()
+        self.categories = self._init_categories
 
     @staticmethod
     def merge_bio_semantics_categories(

deepdoctection/extern/hflayoutlm.py

@@ -48,7 +48,7 @@ with try_import() as pt_import_guard:
     import torch.nn.functional as F
 
 with try_import() as tr_import_guard:
-    from timm.data.constants import IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD  # type: ignore
+    from timm.data.constants import IMAGENET_DEFAULT_MEAN, IMAGENET_DEFAULT_STD
     from transformers import (
         LayoutLMForSequenceClassification,
         LayoutLMForTokenClassification,

deepdoctection/mapper/misc.py

@@ -27,7 +27,7 @@ from typing import Mapping, Optional, Sequence, Union
 
 from lazy_imports import try_import
 
-from ..datapoint.convert import convert_pdf_bytes_to_np_array_v2
+from ..datapoint.convert import convert_bytes_to_np_array, convert_pdf_bytes_to_np_array_v2
 from ..datapoint.image import Image
 from ..utils.fs import get_load_image_func, load_image_from_file
 from ..utils.types import JsonDict
@@ -49,6 +49,7 @@ def to_image(dp: Union[str, Mapping[str, Union[str, bytes]]], dpi: Optional[int]
 
     file_name: Optional[str]
     location: Optional[str]
+    image_bytes: Optional[bytes] = None
 
     if isinstance(dp, str):
         _, file_name = os.path.split(dp)
@@ -62,6 +63,7 @@ def to_image(dp: Union[str, Mapping[str, Union[str, bytes]]], dpi: Optional[int]
         document_id = dp.get("document_id")
         if location == "":
             location = str(dp.get("path", ""))
+        image_bytes = dp.get("image_bytes")
     else:
         raise TypeError("datapoint not of expected type for converting to image")
 
@@ -76,6 +78,8 @@ def to_image(dp: Union[str, Mapping[str, Union[str, bytes]]], dpi: Optional[int]
                 if dp_image.pdf_bytes is not None:
                     if isinstance(dp_image.pdf_bytes, bytes):
                         dp_image.image = convert_pdf_bytes_to_np_array_v2(dp_image.pdf_bytes, dpi=dpi)
+            elif image_bytes is not None:
+                dp_image.image = convert_bytes_to_np_array(image_bytes)
             else:
                 dp_image.image = load_image_from_file(location)
 

deepdoctection/pipe/base.py

@@ -29,6 +29,7 @@ from uuid import uuid1
 
 from ..dataflow import DataFlow, MapData
 from ..datapoint.image import Image
+from ..mapper.misc import curry
 from ..utils.context import timed_operation
 from ..utils.identifier import get_uuid_from_str
 from ..utils.settings import ObjectTypes
@@ -247,17 +248,24 @@ class Pipeline(ABC):
         """
         raise NotImplementedError()
 
-    def _build_pipe(self, df: DataFlow, session_id: Optional[str] = None) -> DataFlow:
+    @staticmethod
+    @curry
+    def _undo(dp: Image, service_ids: Optional[list[str]] = None) -> Image:
         """
-        Composition of the backbone
+        Remove annotations from a datapoint
         """
-        if session_id is None and self.set_session_id:
-            session_id = self.get_session_id()
-        for component in self.pipe_component_list:
-            component.timer_on = True
-            component.dp_manager.session_id = session_id
-            df = component.predict_dataflow(df)
-        return df
+        dp.remove(service_ids=service_ids)
+        return dp
+
+    def undo(self, df: DataFlow, service_ids: Optional[set[str]] = None) -> DataFlow:
+        """
+        Mapping a datapoint via `_undo` within a dataflow pipeline
+
+        :param df: An input dataflow of Images
+        :param service_ids: A set of service ids to remove
+        :return: A output dataflow of Images
+        """
+        return MapData(df, self._undo(service_ids=service_ids))
 
     @abstractmethod
     def analyze(self, **kwargs: Any) -> DataFlow:
@@ -273,6 +281,18 @@ class Pipeline(ABC):
         """
         raise NotImplementedError()
 
+    def _build_pipe(self, df: DataFlow, session_id: Optional[str] = None) -> DataFlow:
+        """
+        Composition of the backbone
+        """
+        if session_id is None and self.set_session_id:
+            session_id = self.get_session_id()
+        for component in self.pipe_component_list:
+            component.timer_on = True
+            component.dp_manager.session_id = session_id
+            df = component.predict_dataflow(df)
+        return df
+
     def get_meta_annotation(self) -> MetaAnnotation:
         """
         Collects meta annotations from all pipeline components and summarizes the returned results

deepdoctection/pipe/doctectionpipe.py

@@ -23,31 +23,38 @@ import os
 from pathlib import Path
 from typing import List, Mapping, Optional, Sequence, Tuple, Union
 
-from ..dataflow import DataFlow, MapData
+from ..dataflow import CustomDataFromIterable, DataFlow, DataFromList, MapData
 from ..dataflow.custom_serialize import SerializerFiles, SerializerPdfDoc
 from ..datapoint.image import Image
 from ..datapoint.view import IMAGE_DEFAULTS
 from ..mapper.maputils import curry
 from ..mapper.misc import to_image
 from ..utils.fs import maybe_path_or_pdf
+from ..utils.identifier import get_uuid_from_str
 from ..utils.logger import LoggingRecord, logger
+from ..utils.pdf_utils import PDFStreamer
 from ..utils.types import PathLikeOrStr
+from ..utils.utils import is_file_extension
 from .base import Pipeline, PipelineComponent
 from .common import PageParsingService
 
 
 def _collect_from_kwargs(
-    **kwargs: Union[str, DataFlow, bool, int, PathLikeOrStr, Union[str, List[str]]]
-) -> Tuple[Optional[str], Optional[str], bool, int, str, DataFlow]:
+    **kwargs: Union[Optional[str], bytes, DataFlow, bool, int, PathLikeOrStr, Union[str, List[str]]]
+) -> Tuple[Optional[str], Union[str, Sequence[str]], bool, int, str, DataFlow, Optional[bytes]]:
+    b_bytes = kwargs.get("bytes")
     dataset_dataflow = kwargs.get("dataset_dataflow")
     path = kwargs.get("path")
     if path is None and dataset_dataflow is None:
         raise ValueError("Pass either path or dataset_dataflow as argument")
+    if path is None and b_bytes:
+        raise ValueError("When passing bytes, a path to the source document must be provided")
 
     shuffle = kwargs.get("shuffle", False)
     if not isinstance(shuffle, bool):
         raise TypeError(f"shuffle must be of type bool but is of type {type(shuffle)}")
 
+    file_type = None
     doc_path = None
     if path:
         if not isinstance(path, (str, Path)):
@@ -56,15 +63,27 @@ def _collect_from_kwargs(
         if path_type == 2:
             doc_path = path
             path = None
+            file_type = ".pdf"
+        elif path_type == 3:
+            if is_file_extension(path, ".jpg"):
+                file_type = ".jpg"
+            if is_file_extension(path, ".png"):
+                file_type = ".png"
+            if is_file_extension(path, ".jpeg"):
+                file_type = ".jpeg"
+            if not b_bytes:
+                raise ValueError("When passing a path to a single image, bytes of the image must be passed")
         elif not path_type:
             raise ValueError("Pass only a path to a directory or to a pdf file")
 
-    file_type = kwargs.get("file_type", [".jpg", ".png", ".tif"])
+    file_type = kwargs.get(
+        "file_type", [".jpg", ".png", ".jpeg", ".tif"] if file_type is None else file_type  # type: ignore
+    )
 
     max_datapoints = kwargs.get("max_datapoints")
     if not isinstance(max_datapoints, (int, type(None))):
         raise TypeError(f"max_datapoints must be of type int, but is of type {type(max_datapoints)}")
-    return path, file_type, shuffle, max_datapoints, doc_path, dataset_dataflow  # type: ignore
+    return path, file_type, shuffle, max_datapoints, doc_path, dataset_dataflow, b_bytes  # type: ignore
 
 
 @curry
@@ -142,12 +161,18 @@ class DoctectionPipe(Pipeline):
 
         super().__init__(pipeline_component_list)
 
-    def _entry(self, **kwargs: Union[str, DataFlow, bool, int, PathLikeOrStr, Union[str, List[str]]]) -> DataFlow:
-        path, file_type, shuffle, max_datapoints, doc_path, dataset_dataflow = _collect_from_kwargs(**kwargs)
+    def _entry(self, **kwargs: Union[str, bytes, DataFlow, bool, int, PathLikeOrStr, Union[str, List[str]]]) \
+            -> DataFlow:
+        path, file_type, shuffle, max_datapoints, doc_path, dataset_dataflow, b_bytes = _collect_from_kwargs(**kwargs)
 
         df: DataFlow
 
-        if isinstance(path, (str, Path)):
+        if isinstance(b_bytes, bytes):
+            df = DoctectionPipe.bytes_to_dataflow(path=doc_path if path is None else path,
+                                                  b_bytes=b_bytes,
+                                                  file_type=file_type)
+
+        elif isinstance(path, (str, Path)):
             if not isinstance(file_type, (str, list)):
                 raise TypeError(f"file_type must be of type string or list, but is of type {type(file_type)}")
             df = DoctectionPipe.path_to_dataflow(path=path, file_type=file_type, shuffle=shuffle)
@@ -162,7 +187,7 @@ class DoctectionPipe(Pipeline):
 
         df = MapData(df, _proto_process(path, doc_path))
         if dataset_dataflow is None:
-            df = MapData(df, _to_image(dpi=300))  # pylint: disable=E1120
+            df = MapData(df, _to_image(dpi=os.environ.get("DPI", 300)))  # pylint: disable=E1120
         return df
 
     @staticmethod
@@ -197,6 +222,44 @@ class DoctectionPipe(Pipeline):
         """
         return _doc_to_dataflow(path, max_datapoints)
 
+    @staticmethod
+    def bytes_to_dataflow(
+        path: str, b_bytes: bytes, file_type: Union[str, Sequence[str]], max_datapoints: Optional[int] = None
+    ) -> DataFlow:
+        """
+        Converts a bytes object to a dataflow
+
+        :param path: path to directory or an image file
+        :param b_bytes: bytes object
+        :param file_type: e.g. ".pdf", ".jpg" or [".jpg", ".png", ".jpeg", ".tif"]
+        :param max_datapoints: max number of datapoints to consider
+        :return: DataFlow
+        """
+
+        file_name = os.path.split(path)[1]
+        if isinstance(file_type, str):
+            if file_type == ".pdf":
+                prefix, suffix = os.path.splitext(file_name)
+                df: DataFlow
+                df = CustomDataFromIterable(PDFStreamer(path_or_bytes=b_bytes), max_datapoints=max_datapoints)
+                df = MapData(
+                    df,
+                    lambda dp: {
+                        "path": path,
+                        "file_name": prefix + f"_{dp[1]}" + suffix,
+                        "pdf_bytes": dp[0],
+                        "page_number": dp[1],
+                        "document_id": get_uuid_from_str(prefix),
+                    },
+                )
+            else:
+                df = DataFromList(lst=[{"path": path, "file_name": file_name, "image_bytes": b_bytes}])
+            return df
+        raise ValueError(
+            f"pass: {path}, b_bytes: {b_bytes!r}, file_type: {file_type} and max_datapoints: {max_datapoints} "
+            f"not supported"
+        )
+
     def dataflow_to_page(self, df: DataFlow) -> DataFlow:
         """
         Converts a dataflow of images to a dataflow of pages
@@ -206,7 +269,9 @@ class DoctectionPipe(Pipeline):
         """
         return self.page_parser.predict_dataflow(df)
 
-    def analyze(self, **kwargs: Union[str, DataFlow, bool, int, PathLikeOrStr, Union[str, List[str]]]) -> DataFlow:
+    def analyze(
+        self, **kwargs: Union[str, bytes, DataFlow, bool, int, PathLikeOrStr, Union[str, List[str]]]
+    ) -> DataFlow:
         """
         `kwargs key dataset_dataflow:` Transfer a dataflow of a dataset via its dataflow builder
 
@@ -215,6 +280,8 @@ class DoctectionPipe(Pipeline):
                            only the first page is processed through the pipeline.
                            Alternatively, a path to a pdf document with multiple pages.
 
+        `kwargs key bytes:` A bytes object of an image
+
         `kwargs key file_type:` Selection of the file type, if: args:`file_type` is passed
 
         `kwargs key max_datapoints:` Stops processing as soon as max_datapoints images have been processed

deepdoctection/utils/fs.py

@@ -227,20 +227,21 @@ def get_load_image_func(
 
 def maybe_path_or_pdf(path: PathLikeOrStr) -> int:
     """
-    Checks if the path points to a directory or a pdf document. Returns 1 if the path points to a directory, 2
-    if the path points to a pdf doc or 0, if none of the previous is true.
+    Checks if the path points to a directory, a pdf document or a single image. Returns 1 if the path points to a
+    directory, 2 if the path points to a pdf doc and 3 if path points to either a PNG, JPG or JPEG or 0 if none of the
+    previous is true.
 
     :param path: A path
-    :return: A value of 0,1,2
+    :return: A value of 0,1,2,3
     """
 
-    is_dir = os.path.isdir(path)
-    if is_dir:
+    if os.path.isdir(path):
         return 1
     file_name = os.path.split(path)[1]
-    is_pdf = is_file_extension(file_name, ".pdf")
-    if is_pdf:
+    if is_file_extension(file_name, ".pdf"):
         return 2
+    if is_file_extension(file_name, [".png", ".jpeg", ".jpg", ".tif"]):
+        return 3
     return 0
 
 

deepdoctection/utils/pdf_utils.py

@@ -26,7 +26,7 @@ from errno import ENOENT
 from io import BytesIO
 from pathlib import Path
 from shutil import copyfile
-from typing import Generator, Literal, Optional
+from typing import Generator, Literal, Optional, Union
 
 from lazy_imports import try_import
 from numpy import uint8
@@ -46,6 +46,7 @@ with try_import() as pt_import_guard:
 
 __all__ = [
     "decrypt_pdf_document",
+    "decrypt_pdf_document_from_bytes",
     "get_pdf_file_reader",
     "get_pdf_file_writer",
     "PDFStreamer",
@@ -68,7 +69,6 @@ def decrypt_pdf_document(path: PathLikeOrStr) -> bool:
     :param path: A path to the pdf file
     :return: True if document has been successfully decrypted
     """
-
     if qpdf_available():
         path_base, file_name = os.path.split(path)
         file_name_tmp = os.path.splitext(file_name)[0] + "tmp.pdf"
@@ -86,41 +86,69 @@ def decrypt_pdf_document(path: PathLikeOrStr) -> bool:
     return False
 
 
-def get_pdf_file_reader(path: PathLikeOrStr) -> PdfReader:
+def decrypt_pdf_document_from_bytes(input_bytes: bytes) -> bytes:
+    """
+    Decrypting a pdf given as bytes. Under the hood, it saves the bytes to a temporary file and then calls
+
+    qpdf: <http://qpdf.sourceforge.net/>
+
+    :param input_bytes: A bytes object representing the pdf file
+    :return: The decrypted bytes object
+    """
+    with save_tmp_file(input_bytes, "pdf_") as (_, input_file_name):
+        is_decrypted = decrypt_pdf_document(input_file_name)
+        if is_decrypted:
+            with open(input_file_name, "rb") as file:
+                return file.read()
+        else:
+            logger.error(LoggingRecord("pdf bytes cannot be decrypted and therefore cannot be processed further."))
+            sys.exit()
+
+
+def get_pdf_file_reader(path_or_bytes: Union[PathLikeOrStr, bytes]) -> PdfReader:
     """
     Creates a file reader object from a pdf document. Will try to decrypt the document if it is
     encrypted. (See `decrypt_pdf_document` to understand what is meant with "decrypt").
 
-    :param path: A path to a pdf document
+    :param path_or_bytes: A path to a pdf document
     :return: A file reader object from which you can iterate through the document.
     """
 
-    if not os.path.isfile(path):
-        raise FileNotFoundError(str(path))
-    file_name = os.path.split(path)[1]
+    if isinstance(path_or_bytes, bytes):
+        try:
+            reader = PdfReader(BytesIO(path_or_bytes))
+        except (errors.PdfReadError, AttributeError):
+            decrypted_bytes = decrypt_pdf_document_from_bytes(path_or_bytes)
+            reader = PdfReader(BytesIO(decrypted_bytes))
+        return reader
+
+    if not os.path.isfile(path_or_bytes):
+        raise FileNotFoundError(str(path_or_bytes))
+    file_name = os.path.split(path_or_bytes)[1]
     if not is_file_extension(file_name, ".pdf"):
         raise FileExtensionError(f"must be a pdf file: {file_name}")
 
-    with open(path, "rb") as file:
+    with open(path_or_bytes, "rb") as file:
         qpdf_called = False
         try:
-            input_pdf_as_bytes = PdfReader(file)
+            reader = PdfReader(file)
         except (errors.PdfReadError, AttributeError):
-            _ = decrypt_pdf_document(path)
+            _ = decrypt_pdf_document(path_or_bytes)
             qpdf_called = True
 
         if not qpdf_called:
-            if input_pdf_as_bytes.is_encrypted:
-                is_decrypted = decrypt_pdf_document(path)
+            if reader.is_encrypted:
+                is_decrypted = decrypt_pdf_document(path_or_bytes)
                 if not is_decrypted:
                     logger.error(
                         LoggingRecord(
-                            f"pdf document {path} cannot be decrypted and therefore cannot be " f"processed further."
+                            f"pdf document {path_or_bytes} cannot be decrypted and therefore cannot "
+                            f"be processed further."
                         )
                     )
                     sys.exit()
 
-    return PdfReader(os.fspath(path))
+    return PdfReader(os.fspath(path_or_bytes))
 
 
 def get_pdf_file_writer() -> PdfWriter:
@@ -157,11 +185,11 @@ class PDFStreamer:
 
     """
 
-    def __init__(self, path: PathLikeOrStr) -> None:
+    def __init__(self, path_or_bytes: Union[PathLikeOrStr, bytes]) -> None:
         """
-        :param path: to a pdf.
+        :param path_or_bytes: to a pdf.
         """
-        self.file_reader = get_pdf_file_reader(path)
+        self.file_reader = get_pdf_file_reader(path_or_bytes)
         self.file_writer = PdfWriter()
 
     def __len__(self) -> int:

deepdoctection/utils/viz.py

@@ -312,6 +312,7 @@ class VizPackageHandler:
             "interactive_imshow": "_cv2_interactive_imshow",
             "encode": "_cv2_encode",
             "rotate_image": "_cv2_rotate_image",
+            "convert_bytes_to_np": "_cv2_convert_bytes_to_np",
         },
         "pillow": {
             "read_image": "_pillow_read_image",
@@ -325,6 +326,7 @@ class VizPackageHandler:
             "interactive_imshow": "_pillow_interactive_imshow",
             "encode": "_pillow_encode",
             "rotate_image": "_pillow_rotate_image",
+            "convert_bytes_to_np": "_pillow_convert_bytes_to_np",
         },
     }
 
@@ -484,6 +486,37 @@ class VizPackageHandler:
         pil_image = Image.open(im_file)
         return np.array(pil_image)[:, :, ::-1]
 
+    def convert_bytes_to_np(self, image_bytes: bytes) -> PixelValues:
+        """Converting an image as bytes into np.array
+
+        :param image_bytes: Image as np.array
+        """
+        return getattr(self, self.pkg_func_dict["convert_bytes_to_np"])(image_bytes)
+
+    @staticmethod
+    def _cv2_convert_bytes_to_np(image_bytes: bytes) -> PixelValues:
+        """
+        Convert image bytes to a numpy array using OpenCV.
+
+        :param image_bytes: Image bytes
+        :return: Image as numpy array
+        """
+        np_array = np.frombuffer(image_bytes, np.uint8)
+        np_image = cv2.imdecode(np_array, cv2.IMREAD_COLOR)
+        return np_image
+
+    @staticmethod
+    def _pillow_convert_bytes_to_np(image_bytes: bytes) -> PixelValues:
+        """
+        Convert image bytes to a numpy array using Pillow.
+
+        :param image_bytes: Image bytes
+        :return: Image as numpy array
+        """
+        image = Image.open(BytesIO(image_bytes))
+        np_image = np.array(image)
+        return np_image
+
     def resize(self, image: PixelValues, width: int, height: int, interpolation: str) -> PixelValues:
         """
         Resize a given image to new width, height. Specifying an interpolation method is required. Depending on the

{deepdoctection-0.36.dist-info → deepdoctection-0.37.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: deepdoctection
-Version: 0.36
+Version: 0.37.1
 Summary: Repository for Document AI
 Home-page: https://github.com/deepdoctection/deepdoctection
 Author: Dr. Janis Meyer

{deepdoctection-0.36.dist-info → deepdoctection-0.37.1.dist-info}/RECORD

@@ -1,9 +1,9 @@
-deepdoctection/__init__.py,sha256=fNUbaFAlK1JUXgPCmTu2UOLUMqW4HIgkaW4uOUYjYYg,12571
+deepdoctection/__init__.py,sha256=i23UZBqMlkcvUILJxvUQAdj-3d2yV9edzxFsC5RoMHA,12655
 deepdoctection/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 deepdoctection/analyzer/__init__.py,sha256=icClxrd20XutD6LxLgEPIWceSs4j_QfI3szCE-9BL2w,729
 deepdoctection/analyzer/_config.py,sha256=NZl_REM8Ge2xfxvHN-mZR5KURcHfZii3xfMlKQwckbA,4864
 deepdoctection/analyzer/dd.py,sha256=DUOhOtwipHw5nabYqn3WGR9aZcgP0ma_bi_tjf9xscw,5973
-deepdoctection/analyzer/factory.py,sha256=xmo5F9X7I6lp0ZWJv8QavpMyG8UWYLvMi4qogsZV1_s,31507
+deepdoctection/analyzer/factory.py,sha256=dEUOtdBS3yQGLqMqLR_kq5EYCR3IE30DjHNzE0spoQE,31519
 deepdoctection/configs/__init__.py,sha256=TX_P6tqDOF1LK1mi9ruAl7x0mtv1Asm8cYWCz3Pe2dk,646
 deepdoctection/configs/conf_dd_one.yaml,sha256=td7XsyVhdXkhh5Pie7sT_WNjGTaxBOWgpxhkobHd1H0,2325
 deepdoctection/configs/conf_tesseract.yaml,sha256=oF6szDyoi15FHvq7yFUNIEjfA_jNLhGxoowiRsz_zY4,35
@@ -11,15 +11,15 @@ deepdoctection/dataflow/__init__.py,sha256=CWRHMpmJaPk4xY_oIIFubCt-z11SguWrMWxHZ
 deepdoctection/dataflow/base.py,sha256=z4DCComSj5wStEPjtk0093cNNGfUMiDqx8dqz36nS_o,6221
 deepdoctection/dataflow/common.py,sha256=MyGA2VWlNMjQdIN_Jd-o0Ec3bDJmjQit4Nv0v43OCSQ,10119
 deepdoctection/dataflow/custom.py,sha256=3CK_1oL9p6nbOq8WtH5_vQUo70_8Z8pXY7kG0OFqzug,6803
-deepdoctection/dataflow/custom_serialize.py,sha256=CKeyw2Ayq_qAl0O5BoKkIOFJgteCt78h9QFTI23XhmQ,22818
+deepdoctection/dataflow/custom_serialize.py,sha256=WocuiYo2gkih5Z9lWAoIIfUewwYSDOhHzG7ZZjKlUic,22827
 deepdoctection/dataflow/parallel_map.py,sha256=8FhxJBWV-kjJrJ27jQtP3yYF6Ev6rz98worO60oi96c,15837
 deepdoctection/dataflow/serialize.py,sha256=4pYC7m9h53JCu99waVeKpHDpsCDDdYCrSZpP2QYSsgs,4555
 deepdoctection/dataflow/stats.py,sha256=Bsr6v7lcesKXUYtO9wjqlzx_Yq_uyIF3Lel-tQ0i4wI,9619
 deepdoctection/datapoint/__init__.py,sha256=3K406GbOPhoEp8koVaSbMocmSsmWifnSZ1SPb7C1lOY,1643
 deepdoctection/datapoint/annotation.py,sha256=FEgz4COxVDfjic0gG7kS6iHnWLBIgFnquQ63Cbj2a4Y,22531
 deepdoctection/datapoint/box.py,sha256=tkFuVM6xfx2jL7W4UED4qHXV572LSRdIsVJbrEiyIxI,23524
-deepdoctection/datapoint/convert.py,sha256=Gw2IjNiEotPu1yuMZqrIYB0mCAwafKt-VgMnrHj6S7U,6808
-deepdoctection/datapoint/image.py,sha256=EvZlVwJjMAcL1z8RNPBvZ8fwdJvkGuGpcFxCP1y26Go,33045
+deepdoctection/datapoint/convert.py,sha256=O7920pIomyEkzXwxpFsrzfhn7Pl6UzVGhNzv90VcuKU,7099
+deepdoctection/datapoint/image.py,sha256=AM34br9eM1syTIUXcJIrAaP7pEnejbUl-w-CK5pr9z8,33233
 deepdoctection/datapoint/view.py,sha256=1rVMuqucCrI5zlwyXMADJQBV38V_zSNFqFyBi3cMA1E,44914
 deepdoctection/datasets/__init__.py,sha256=-A3aR90aDsHPmVM35JavfnQ2itYSCn3ujl4krRni1QU,1076
 deepdoctection/datasets/adapter.py,sha256=Ly_vbOAgVI73V41FUccnSX1ECTOyesW_qsuvQuvOZbw,7796
@@ -27,7 +27,7 @@ deepdoctection/datasets/base.py,sha256=DT4i-d74sIEiUNC6UspIHNJuHSK0t1dBv7qwadg4r
 deepdoctection/datasets/dataflow_builder.py,sha256=cYU2zV3gZW2bFvMHimlO9VIl3BAUaCwML08cCIQ8Em4,4107
 deepdoctection/datasets/info.py,sha256=6y5TfiUhQppynbMFP5JmUPk95ggsVCtGIw4dYh2lVus,20501
 deepdoctection/datasets/registry.py,sha256=ZjzVzjsCgNXJuZZZtR98_yKocADmh4EBGV5JqJbGjWk,2543
-deepdoctection/datasets/save.py,sha256=khYQ4t94FOu9RWMimP9E4kASq25f61SIow78NHaX1pg,3349
+deepdoctection/datasets/save.py,sha256=Y9508Qqp8gIGN7pbGgVBBnkiC6NdCb9L2YR4wVvEUxM,3350
 deepdoctection/datasets/instances/__init__.py,sha256=XEc_4vT5lDn6bbZID9ujDEumWu8Ec2W-QS4pI_bfWWE,1388
 deepdoctection/datasets/instances/doclaynet.py,sha256=wRZT7wMTilZBLZ1gKY2cWReD1EGT735vOOTy0pD0N6M,12038
 deepdoctection/datasets/instances/fintabnet.py,sha256=qYzFK1dWF6MEPkHamP255DvAzlQT_GnkvDe1aM7CgjA,12006
@@ -50,13 +50,13 @@ deepdoctection/eval/registry.py,sha256=v4mp-s67vBVRu1nQzuGlYPViQnMSeIXEcF_WmvfUC
 deepdoctection/eval/tedsmetric.py,sha256=rKw-734Y9CpBtIfkBSPQF2vAZxnIdWrI9Zc723P7RxI,9529
 deepdoctection/eval/tp_eval_callback.py,sha256=SXsXumoyxq-MIH9Cep5eUOwnNshMbKmC6mYOGwCg0pM,5283
 deepdoctection/extern/__init__.py,sha256=9Iks9b4Q_LynjcV167TVCoK8YsQRUcA2jjmAmDNA_X8,1056
-deepdoctection/extern/base.py,sha256=ajzFzD9BrFwnly4SziN8PadI-PBOzzVRlIGPm_sNllE,24142
+deepdoctection/extern/base.py,sha256=ONPgappl_P5HSwQr42FatuRnwMTvUPecPsCztDTN0Hw,24108
 deepdoctection/extern/d2detect.py,sha256=zrKv1yurApnjD7QZIZk_8LYCahjmN82MQUjHjv8zvkQ,22127
 deepdoctection/extern/deskew.py,sha256=sPoixu8S9he-0wbs-jgxtPE2V9BiP4-3uZlb6F5Y1SA,3077
 deepdoctection/extern/doctrocr.py,sha256=T3_tvlih22_dVCBZypS1Y8tjQQB1fkAxIbGdUGHIapQ,24473
 deepdoctection/extern/fastlang.py,sha256=F4gK-SEwcCujjxH327ZDzMGWToJ49xS_dCKcePQ9IlY,4780
 deepdoctection/extern/hfdetr.py,sha256=1NPW_u5eH2tP3ixZ91l4WR-O-wLVcrFsLWA7BqID0oM,12055
-deepdoctection/extern/hflayoutlm.py,sha256=KfoWx9_Rpa1Y2L51HLrYvenfWaTB4SVTmVJH00Cqb-s,56510
+deepdoctection/extern/hflayoutlm.py,sha256=_OUeQsbNgfjbV7TPYBjkqc4HoTBQqkOINnwpewPJpl8,56494
 deepdoctection/extern/hflm.py,sha256=kwS6kcSlY_2m9u0RzBLTRq-UMM7c1PhyUaDTvSdejus,9217
 deepdoctection/extern/model.py,sha256=ViHHKPvbGmLCPw7ZESv_rmjlkA90UiBU6oZiHOMqNSw,59869
 deepdoctection/extern/pdftext.py,sha256=KS_t27SUiYn_IOS_J2lF9lSSo22vLagxmxvYCY3CqXA,7228
@@ -95,7 +95,7 @@ deepdoctection/mapper/hfstruct.py,sha256=2PjGKsYturVJBimLT1CahYh09KSRAFEHz_QNtC1
 deepdoctection/mapper/laylmstruct.py,sha256=abMZkYU2W0e_VcCm_c0ZXNFuv-lfMFWcTedcZS5EYvE,42935
 deepdoctection/mapper/maputils.py,sha256=eI6ZcDg9W5uB6xQNBZpMIdEd86HlCxTtkJuyROdTqiw,8146
 deepdoctection/mapper/match.py,sha256=pCWZpz2R8JahiKXCw7dxKRTLiPgJXeVDgkddDPLy_c0,9643
-deepdoctection/mapper/misc.py,sha256=rCqHOcsCfVPXs36AWK0rZ2kk0CUM3yXV370_zyIGBJ4,6518
+deepdoctection/mapper/misc.py,sha256=NLSSgk066Tkrrdi075HkqV7cP-iqT9fv_MtyAJ-8gOg,6743
 deepdoctection/mapper/pascalstruct.py,sha256=TzVU1p0oiw0nOuxTFFbEB9vXJxH1v6VUvTJ7MD0manU,3828
 deepdoctection/mapper/prodigystruct.py,sha256=Re4Sd_zAp6qOvbXZLmMJeG0IGEfMQxebuyDeZgMcTa8,6827
 deepdoctection/mapper/pubstruct.py,sha256=YxsrZ-E0pD45Mm_VCPQB9yEgHsTPkw4htt-3DwCRX1k,23361
@@ -103,10 +103,10 @@ deepdoctection/mapper/tpstruct.py,sha256=YNABRibvcISD5Lavg3jouoE4FMdqXEJoM-hNoB_
 deepdoctection/mapper/xfundstruct.py,sha256=_3r3c0K82fnF2h1HxA85h-9ETYrHwcERa6MNc6Ko6Z8,8807
 deepdoctection/pipe/__init__.py,sha256=ywTVoetftdL6plXg2YlBzMfmqBZupq7yXblSVyvvkcQ,1127
 deepdoctection/pipe/anngen.py,sha256=3319l4aaXzcY4w6ItVBNPX8LGS5fHFDVtyVY9KMefac,16393
-deepdoctection/pipe/base.py,sha256=Davjkf3D837y9AIITcx7yXdebmVaz6Moyw_5Wi3nfmg,13561
+deepdoctection/pipe/base.py,sha256=ynNg5SSRuUVxN69VWOO3Oi7WSeGrYwn3A56NQMBJDvw,14222
 deepdoctection/pipe/common.py,sha256=haOb4v0jLX3r41BSC8cVseX2E320_HkSrGlZsQiKE2g,17728
 deepdoctection/pipe/concurrency.py,sha256=AAKRsVgaBEYNluntbDa46SBF1JZ_XqnWLDSWrNvAzEo,9657
-deepdoctection/pipe/doctectionpipe.py,sha256=I6B6HT_BG2ByQ3Rjsui3-Ct31yLmodx-iuZnujXaiSc,8953
+deepdoctection/pipe/doctectionpipe.py,sha256=uhsrSuwaHcOMj8b8i6wCpPaZlSxCTaeHVhMokJ8vRSI,11835
 deepdoctection/pipe/language.py,sha256=5zI0UQC6Fh12_r2pfVL42HoCGz2hpHrOhpXAn5m-rYw,5451
 deepdoctection/pipe/layout.py,sha256=xIhnJpyUSbvLbhTXyAKXY1hmG9352jihGYFSclTH_1g,5567
 deepdoctection/pipe/lm.py,sha256=Sp-b7smeslNDyioEfNjuNBUxAuFKn3-OKpCZkGXri_c,16643
@@ -129,20 +129,20 @@ deepdoctection/utils/develop.py,sha256=4HyTarkFbJwctL-Hgu1TU_LSJppHvaroDbcyHsxhI
 deepdoctection/utils/env_info.py,sha256=TnCA-LOTj4WIHd9yvn1AaoPWsLmPgc42l-BJmGV6zmM,19147
 deepdoctection/utils/error.py,sha256=_3q9VepKfEhsM3H033_Fu0hwBzMSjsWALsjyJbGAZr8,2367
 deepdoctection/utils/file_utils.py,sha256=IRElrcND0YEiU1QELw5hfXeNA39uE2_nyzh9-X7YcxI,19477
-deepdoctection/utils/fs.py,sha256=C4ktrzjoVtX9kgycv5YrEigDI9byi65b6_D0aKsGM4Y,10161
+deepdoctection/utils/fs.py,sha256=x842BxUP5bbjJ2cofw-g4dKJv4QAaGzda4qnAazabO4,10281
 deepdoctection/utils/identifier.py,sha256=QkNaGGqPynHwDPnd3_m8iur4Cv64rcQa7qolCE7Qphk,2159
 deepdoctection/utils/logger.py,sha256=J0OVKiXP_2A82MWbbJoOeMEJ-75aZu5npgaS_yI6mVA,10003
 deepdoctection/utils/metacfg.py,sha256=hD76KQ_RnD_5B02qLI2Zxf3WfnsnXhEI_KUTKpw91RI,5711
 deepdoctection/utils/mocks.py,sha256=IkN3-IzAl4eX0ibgKIHg8IY7ykVw6BnpF6XnxKnKaZI,2389
-deepdoctection/utils/pdf_utils.py,sha256=OAQjE9xHVNcDsFqAvX47Lu-mgmoMpVXqIf5pOK8AwxY,11595
+deepdoctection/utils/pdf_utils.py,sha256=G0m8kUn2HwwyZWH_BcrDkm-m3MP9GN9SWHj5VhB7swY,12845
 deepdoctection/utils/settings.py,sha256=k6OyuWbj-IPeaO9zT9RZ-5Yad1wNhWGYqGLZdtgXAZY,12464
 deepdoctection/utils/tqdm.py,sha256=cBUtR0L1x0KMeYrLP2rrzyzCamCjpQAKroHXLv81_pk,1820
 deepdoctection/utils/transform.py,sha256=3kCgsEeRkG1efCdkfvj7tUFMs-e2jbjbflq826F2GPU,8502
 deepdoctection/utils/types.py,sha256=_3dmPdCIZNLbgU5QP5k_c5phDf18xLe1kYL6t2nM45s,2953
 deepdoctection/utils/utils.py,sha256=csVs_VvCq4QBETPoE2JdTTL4MFYnD4xh-Js5vRb612g,6492
-deepdoctection/utils/viz.py,sha256=Mok1d0V7NwlhAvO1S1Iq5YitKpVmOfH_XHTSlRelCB0,25902
-deepdoctection-0.36.dist-info/LICENSE,sha256=GQ0rUvuGdrMNEI3iHK5UQx6dIMU1QwAuyXsxUHn5MEQ,11351
-deepdoctection-0.36.dist-info/METADATA,sha256=E-zXgx0bTdSqbd88D_abscR_poEJaKJGIwlv2RFbQs8,19543
-deepdoctection-0.36.dist-info/WHEEL,sha256=PZUExdf71Ui_so67QXpySuHtCi3-J3wvF4ORK6k_S8U,91
-deepdoctection-0.36.dist-info/top_level.txt,sha256=hs2DdoOL9h4mnHhmO82BT4pz4QATIoOZ20PZmlnxFI8,15
-deepdoctection-0.36.dist-info/RECORD,,
+deepdoctection/utils/viz.py,sha256=Jf8ePNYWlpuyaS6SeTYQ4OyA3eNhtgjvAQZnGNdgHC0,27051
+deepdoctection-0.37.1.dist-info/LICENSE,sha256=GQ0rUvuGdrMNEI3iHK5UQx6dIMU1QwAuyXsxUHn5MEQ,11351
+deepdoctection-0.37.1.dist-info/METADATA,sha256=M-HjpJpxuM4tHN0ld8DscsZPgKRUoNmsbx9slFkj6tg,19545
+deepdoctection-0.37.1.dist-info/WHEEL,sha256=PZUExdf71Ui_so67QXpySuHtCi3-J3wvF4ORK6k_S8U,91
+deepdoctection-0.37.1.dist-info/top_level.txt,sha256=hs2DdoOL9h4mnHhmO82BT4pz4QATIoOZ20PZmlnxFI8,15
+deepdoctection-0.37.1.dist-info/RECORD,,