TLDR-473 add dedoc utils (#340)

* use dedoc utils BBox class * use AdaptiveBinarizer from dedoc-utils * use SkewCorrector from dedoc-utils * fix style * fix rotated angle error * delete BBox from docs * fix angles * delete print * fix dedocutils * dedocutils set ver. 0.3.5 * fix mistakes and names --------- Co-authored-by: Nikita Shevtsov <shevtsov@ispras.ru>
ispras · Oct 2, 2023 · 896a31a · 896a31a
1 parent 840bed2
commit 896a31a
Show file tree

Hide file tree

Showing 44 changed files with 68 additions and 367 deletions.
diff --git a/dedoc/data_structures/__init__.py b/dedoc/data_structures/__init__.py
@@ -2,7 +2,6 @@
 import dedoc.data_structures.concrete_annotations as annotations
 from .annotation import Annotation
 from .attached_file import AttachedFile
-from .bbox import BBox
 from .cell_with_meta import CellWithMeta
 from .concrete_annotations import *
 from .document_content import DocumentContent
@@ -17,5 +16,5 @@
 from .tree_node import TreeNode
 from .unstructured_document import UnstructuredDocument
 
-__all__ = ['Annotation', 'AttachedFile', 'BBox', 'DocumentContent', 'DocumentMetadata', 'HierarchyLevel', 'LineMetadata',
+__all__ = ['Annotation', 'AttachedFile', 'DocumentContent', 'DocumentMetadata', 'HierarchyLevel', 'LineMetadata',
            'LineWithMeta', 'ParsedDocument', 'Serializable', 'Table', 'TableMetadata', 'CellWithMeta', 'TreeNode', 'UnstructuredDocument', *annotations.__all__]
diff --git a/dedoc/data_structures/bbox.py b/dedoc/data_structures/bbox.py
diff --git a/dedoc/data_structures/concrete_annotations/bbox_annotation.py b/dedoc/data_structures/concrete_annotations/bbox_annotation.py
@@ -1,9 +1,9 @@
 import json
 
+from dedocutils.data_structures import BBox
 from flask_restx import Api, Model, fields
 
 from dedoc.data_structures.annotation import Annotation
-from dedoc.data_structures.bbox import BBox
 
 
 class BBoxAnnotation(Annotation):

diff --git a/dedoc/readers/pdf_reader/data_classes/tables/cell.py b/dedoc/readers/pdf_reader/data_classes/tables/cell.py
@@ -1,8 +1,9 @@
 import uuid
 from typing import List, Optional
 
+from dedocutils.data_structures import BBox
+
 from dedoc.data_structures.annotation import Annotation
-from dedoc.data_structures.bbox import BBox
 from dedoc.data_structures.line_with_meta import LineWithMeta
 
 

diff --git a/dedoc/readers/pdf_reader/data_classes/tables/location.py b/dedoc/readers/pdf_reader/data_classes/tables/location.py
@@ -2,7 +2,7 @@
 from functools import total_ordering
 from typing import Any, Dict
 
-from dedoc.data_structures.bbox import BBox
+from dedocutils.data_structures import BBox
 
 
 @total_ordering

diff --git a/dedoc/readers/pdf_reader/data_classes/tables/scantable.py b/dedoc/readers/pdf_reader/data_classes/tables/scantable.py
@@ -3,8 +3,8 @@
 from typing import Any, List
 
 import numpy as np
+from dedocutils.data_structures import BBox
 
-from dedoc.data_structures.bbox import BBox
 from dedoc.readers.pdf_reader.data_classes.tables.cell import Cell
 from dedoc.readers.pdf_reader.data_classes.tables.location import Location
 

diff --git a/dedoc/readers/pdf_reader/data_classes/tables/table_tree.py b/dedoc/readers/pdf_reader/data_classes/tables/table_tree.py
@@ -4,8 +4,9 @@
 
 import cv2
 import numpy as np
+from dedocutils.data_structures import BBox
 
-from dedoc.data_structures import BBox, LineWithMeta
+from dedoc.data_structures import LineWithMeta
 from dedoc.readers.pdf_reader.pdf_image_reader.ocr.ocr_cell_extractor import OCRCellExtractor
 from dedoc.utils.image_utils import crop_image_text
 

diff --git a/dedoc/readers/pdf_reader/data_classes/text_with_bbox.py b/dedoc/readers/pdf_reader/data_classes/text_with_bbox.py
@@ -2,8 +2,9 @@
 from typing import List, Optional
 from uuid import uuid1
 
+from dedocutils.data_structures import BBox
+
 from dedoc.data_structures.annotation import Annotation
-from dedoc.data_structures.bbox import BBox
 from dedoc.data_structures.serializable import Serializable
 
 

diff --git a/dedoc/readers/pdf_reader/pdf_image_reader/adaptive_binarizer.py b/dedoc/readers/pdf_reader/pdf_image_reader/adaptive_binarizer.py
diff --git a/...rs/pdf_reader/pdf_image_reader/line_metadata_extractor/bold_classifier/bold_classifier.py b/...rs/pdf_reader/pdf_image_reader/line_metadata_extractor/bold_classifier/bold_classifier.py
@@ -1,8 +1,8 @@
 from typing import List
 
 import numpy as np
+from dedocutils.data_structures import BBox
 
-from dedoc.data_structures import BBox
 from dedoc.readers.pdf_reader.pdf_image_reader.line_metadata_extractor.bold_classifier.agglomerative_clusterizer import BoldAgglomerativeClusterizer
 from dedoc.readers.pdf_reader.pdf_image_reader.line_metadata_extractor.bold_classifier.valley_emphasis_binarizer import ValleyEmphasisBinarizer
 

diff --git a/dedoc/readers/pdf_reader/pdf_image_reader/ocr/ocr_cell_extractor.py b/dedoc/readers/pdf_reader/pdf_image_reader/ocr/ocr_cell_extractor.py
@@ -4,9 +4,9 @@
 
 import cv2
 import numpy as np
+from dedocutils.data_structures import BBox
 
 from dedoc.data_structures import BBoxAnnotation, ConfidenceAnnotation, LineMetadata, LineWithMeta
-from dedoc.data_structures.bbox import BBox
 from dedoc.readers.pdf_reader.pdf_image_reader.ocr.ocr_line_extractor import OCRLineExtractor
 from dedoc.readers.pdf_reader.pdf_image_reader.ocr.ocr_page.ocr_page import OcrPage
 from dedoc.readers.pdf_reader.pdf_image_reader.ocr.ocr_utils import get_text_with_bbox_from_cells

diff --git a/dedoc/readers/pdf_reader/pdf_image_reader/ocr/ocr_page/ocr_block.py b/dedoc/readers/pdf_reader/pdf_image_reader/ocr/ocr_page/ocr_block.py
@@ -1,7 +1,8 @@
 from collections import defaultdict
 from typing import List
 
-from dedoc.data_structures.bbox import BBox
+from dedocutils.data_structures import BBox
+
 from dedoc.readers.pdf_reader.pdf_image_reader.ocr.ocr_page.ocr_paragraph import OcrParagraph
 from dedoc.readers.pdf_reader.pdf_image_reader.ocr.ocr_page.ocr_tuple import OcrElement
 

diff --git a/dedoc/readers/pdf_reader/pdf_image_reader/ocr/ocr_page/ocr_line.py b/dedoc/readers/pdf_reader/pdf_image_reader/ocr/ocr_page/ocr_line.py
@@ -1,7 +1,8 @@
 from typing import List
 
+from dedocutils.data_structures import BBox
+
 from dedoc.data_structures.annotation import Annotation
-from dedoc.data_structures.bbox import BBox
 from dedoc.data_structures.concrete_annotations.bbox_annotation import BBoxAnnotation
 from dedoc.data_structures.concrete_annotations.confidence_annotation import ConfidenceAnnotation
 from dedoc.readers.pdf_reader.pdf_image_reader.ocr.ocr_page.ocr_tuple import OcrElement

diff --git a/dedoc/readers/pdf_reader/pdf_image_reader/ocr/ocr_page/ocr_paragraph.py b/dedoc/readers/pdf_reader/pdf_image_reader/ocr/ocr_page/ocr_paragraph.py
@@ -1,7 +1,8 @@
 from collections import defaultdict
 from typing import List
 
-from dedoc.data_structures.bbox import BBox
+from dedocutils.data_structures import BBox
+
 from dedoc.readers.pdf_reader.pdf_image_reader.ocr.ocr_page.ocr_line import OcrLine
 from dedoc.readers.pdf_reader.pdf_image_reader.ocr.ocr_page.ocr_tuple import OcrElement
 

diff --git a/dedoc/readers/pdf_reader/pdf_image_reader/ocr/ocr_page/ocr_tuple.py b/dedoc/readers/pdf_reader/pdf_image_reader/ocr/ocr_page/ocr_tuple.py
@@ -1,6 +1,6 @@
 from typing import Dict, Iterable
 
-from dedoc.data_structures.bbox import BBox
+from dedocutils.data_structures import BBox
 
 
 class OcrElement: