deepdoctection 0.33__py3-none-any.whl → 0.35__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Potentially problematic release.


This version of deepdoctection might be problematic. Click here for more details.

Files changed (40) hide show
  1. deepdoctection/__init__.py +11 -12
  2. deepdoctection/analyzer/__init__.py +1 -0
  3. deepdoctection/analyzer/_config.py +150 -0
  4. deepdoctection/analyzer/dd.py +42 -358
  5. deepdoctection/analyzer/factory.py +522 -0
  6. deepdoctection/configs/conf_dd_one.yaml +1 -0
  7. deepdoctection/datapoint/annotation.py +41 -3
  8. deepdoctection/datapoint/convert.py +6 -4
  9. deepdoctection/datapoint/image.py +132 -46
  10. deepdoctection/datapoint/view.py +2 -1
  11. deepdoctection/datasets/base.py +1 -1
  12. deepdoctection/datasets/instances/fintabnet.py +1 -1
  13. deepdoctection/datasets/instances/xfund.py +29 -7
  14. deepdoctection/eval/eval.py +7 -1
  15. deepdoctection/extern/model.py +2 -1
  16. deepdoctection/extern/pdftext.py +96 -5
  17. deepdoctection/extern/tessocr.py +1 -0
  18. deepdoctection/mapper/cats.py +11 -13
  19. deepdoctection/mapper/cocostruct.py +6 -2
  20. deepdoctection/mapper/d2struct.py +2 -1
  21. deepdoctection/mapper/laylmstruct.py +1 -1
  22. deepdoctection/mapper/match.py +31 -0
  23. deepdoctection/mapper/misc.py +1 -1
  24. deepdoctection/mapper/prodigystruct.py +1 -1
  25. deepdoctection/pipe/anngen.py +27 -0
  26. deepdoctection/pipe/base.py +23 -0
  27. deepdoctection/pipe/common.py +123 -38
  28. deepdoctection/pipe/segment.py +1 -1
  29. deepdoctection/pipe/sub_layout.py +1 -1
  30. deepdoctection/utils/env_info.py +31 -2
  31. deepdoctection/utils/file_utils.py +19 -0
  32. deepdoctection/utils/fs.py +27 -4
  33. deepdoctection/utils/metacfg.py +12 -0
  34. deepdoctection/utils/pdf_utils.py +114 -6
  35. deepdoctection/utils/settings.py +3 -0
  36. {deepdoctection-0.33.dist-info → deepdoctection-0.35.dist-info}/METADATA +20 -11
  37. {deepdoctection-0.33.dist-info → deepdoctection-0.35.dist-info}/RECORD +40 -38
  38. {deepdoctection-0.33.dist-info → deepdoctection-0.35.dist-info}/WHEEL +1 -1
  39. {deepdoctection-0.33.dist-info → deepdoctection-0.35.dist-info}/LICENSE +0 -0
  40. {deepdoctection-0.33.dist-info → deepdoctection-0.35.dist-info}/top_level.txt +0 -0
@@ -15,30 +15,22 @@ if importlib.util.find_spec("dotenv") is not None:
15
15
 
16
16
 
17
17
  # pylint: disable=wrong-import-position
18
- import os
19
18
  import sys
20
19
  from typing import TYPE_CHECKING
21
20
 
22
- from .utils.env_info import collect_env_info
21
+ from .utils.env_info import auto_select_pdf_render_framework, collect_env_info
23
22
  from .utils.file_utils import _LazyModule, get_tf_version, pytorch_available, tf_available
24
23
  from .utils.logger import LoggingRecord, logger
25
24
 
26
25
  # pylint: enable=wrong-import-position
27
26
 
28
- __version__ = 0.33
27
+ __version__ = 0.35
29
28
 
30
29
  _IMPORT_STRUCTURE = {
31
30
  "analyzer": [
32
- "maybe_copy_config_to_cache",
33
31
  "config_sanity_checks",
34
- "build_detector",
35
- "build_padder",
36
- "build_service",
37
- "build_sub_image_service",
38
- "build_ocr",
39
- "build_doctr_word",
40
32
  "get_dd_analyzer",
41
- "build_analyzer",
33
+ "ServiceFactory"
42
34
  ],
43
35
  "configs": [],
44
36
  "dataflow": [
@@ -76,6 +68,7 @@ _IMPORT_STRUCTURE = {
76
68
  ],
77
69
  "datapoint": [
78
70
  "ann_from_dict",
71
+ "AnnotationMap",
79
72
  "Annotation",
80
73
  "CategoryAnnotation",
81
74
  "ImageAnnotation",
@@ -198,6 +191,7 @@ _IMPORT_STRUCTURE = {
198
191
  "print_model_infos",
199
192
  "ModelDownloadManager",
200
193
  "PdfPlumberTextDetector",
194
+ "Pdfmium2TextDetector",
201
195
  "TesseractOcrDetector",
202
196
  "TesseractRotationTransformer",
203
197
  "TextractOcrDetector",
@@ -237,6 +231,7 @@ _IMPORT_STRUCTURE = {
237
231
  "LabelSummarizer",
238
232
  "curry",
239
233
  "match_anns_by_intersection",
234
+ "match_anns_by_distance",
240
235
  "to_image",
241
236
  "maybe_load_image",
242
237
  "maybe_remove_image",
@@ -265,6 +260,8 @@ _IMPORT_STRUCTURE = {
265
260
  "DetectResultGenerator",
266
261
  "SubImageLayoutService",
267
262
  "ImageCroppingService",
263
+ "IntersectionMatcher",
264
+ "NeighbourMatcher",
268
265
  "MatchingService",
269
266
  "PageParsingService",
270
267
  "AnnotationNmsService",
@@ -302,6 +299,7 @@ _IMPORT_STRUCTURE = {
302
299
  "timed_operation",
303
300
  "collect_env_info",
304
301
  "auto_select_viz_library",
302
+ "auto_select_pdf_render_framework",
305
303
  "get_tensorflow_requirement",
306
304
  "tf_addons_available",
307
305
  "get_tf_addons_requirements",
@@ -364,6 +362,7 @@ _IMPORT_STRUCTURE = {
364
362
  "get_configs_dir_path",
365
363
  "get_weights_dir_path",
366
364
  "get_dataset_dir_path",
365
+ "maybe_copy_config_to_cache",
367
366
  "is_uuid_like",
368
367
  "get_uuid_from_str",
369
368
  "get_uuid",
@@ -424,7 +423,7 @@ _IMPORT_STRUCTURE = {
424
423
  # Setting some environment variables so that standard functions can be invoked with available hardware
425
424
  env_info = collect_env_info()
426
425
  logger.debug(LoggingRecord(msg=env_info))
427
-
426
+ auto_select_pdf_render_framework()
428
427
 
429
428
  # Direct imports for type-checking
430
429
  if TYPE_CHECKING:
@@ -20,3 +20,4 @@ Package for pre-built pipelines
20
20
  """
21
21
 
22
22
  from .dd import *
23
+ from .factory import *
@@ -0,0 +1,150 @@
1
+ # -*- coding: utf-8 -*-
2
+ # File: config.py
3
+
4
+ # Copyright 2024 Dr. Janis Meyer. All rights reserved.
5
+ #
6
+ # Licensed under the Apache License, Version 2.0 (the "License");
7
+ # you may not use this file except in compliance with the License.
8
+ # You may obtain a copy of the License at
9
+ #
10
+ # http://www.apache.org/licenses/LICENSE-2.0
11
+ #
12
+ # Unless required by applicable law or agreed to in writing, software
13
+ # distributed under the License is distributed on an "AS IS" BASIS,
14
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
15
+ # See the License for the specific language governing permissions and
16
+ # limitations under the License.
17
+
18
+ """Pipeline configuration for deepdoctection analyzer. Do not change the defaults in this file. """
19
+
20
+ from ..utils.metacfg import AttrDict
21
+ from ..utils.settings import CellType, LayoutType
22
+
23
+ cfg = AttrDict()
24
+
25
+ cfg.LANGUAGE = None
26
+ cfg.LIB = None
27
+ cfg.DEVICE = None
28
+ cfg.USE_ROTATOR = False
29
+ cfg.USE_LAYOUT = True
30
+ cfg.USE_TABLE_SEGMENTATION = True
31
+
32
+ cfg.TF.LAYOUT.WEIGHTS = "layout/model-800000_inf_only.data-00000-of-00001"
33
+ cfg.TF.LAYOUT.FILTER = None
34
+
35
+
36
+ cfg.TF.CELL.WEIGHTS = "cell/model-1800000_inf_only.data-00000-of-00001"
37
+ cfg.TF.CELL.FILTER = None
38
+
39
+
40
+ cfg.TF.ITEM.WEIGHTS = "item/model-1620000_inf_only.data-00000-of-00001"
41
+ cfg.TF.ITEM.FILTER = None
42
+
43
+ cfg.PT.LAYOUT.WEIGHTS = "layout/d2_model_0829999_layout_inf_only.pt"
44
+ cfg.PT.LAYOUT.WEIGHTS_TS = "layout/d2_model_0829999_layout_inf_only.ts"
45
+ cfg.PT.LAYOUT.FILTER = None
46
+ cfg.PT.LAYOUT.PAD.TOP = 60
47
+ cfg.PT.LAYOUT.PAD.RIGHT = 60
48
+ cfg.PT.LAYOUT.PAD.BOTTOM = 60
49
+ cfg.PT.LAYOUT.PAD.LEFT = 60
50
+
51
+ cfg.PT.ITEM.WEIGHTS = "item/d2_model_1639999_item_inf_only.pt"
52
+ cfg.PT.ITEM.WEIGHTS_TS = "item/d2_model_1639999_item_inf_only.ts"
53
+ cfg.PT.ITEM.FILTER = None
54
+ cfg.PT.ITEM.PAD.TOP = 60
55
+ cfg.PT.ITEM.PAD.RIGHT = 60
56
+ cfg.PT.ITEM.PAD.BOTTOM = 60
57
+ cfg.PT.ITEM.PAD.LEFT = 60
58
+
59
+ cfg.PT.CELL.WEIGHTS = "cell/d2_model_1849999_cell_inf_only.pt"
60
+ cfg.PT.CELL.WEIGHTS_TS = "cell/d2_model_1849999_cell_inf_only.ts"
61
+ cfg.PT.CELL.FILTER = None
62
+
63
+ cfg.USE_LAYOUT_NMS = False
64
+ cfg.LAYOUT_NMS_PAIRS.COMBINATIONS = None
65
+ cfg.LAYOUT_NMS_PAIRS.THRESHOLDS = None
66
+ cfg.LAYOUT_NMS_PAIRS.PRIORITY = None
67
+
68
+ cfg.SEGMENTATION.ASSIGNMENT_RULE = "ioa"
69
+ cfg.SEGMENTATION.THRESHOLD_ROWS = 0.4
70
+ cfg.SEGMENTATION.THRESHOLD_COLS = 0.4
71
+ cfg.SEGMENTATION.FULL_TABLE_TILING = True
72
+ cfg.SEGMENTATION.REMOVE_IOU_THRESHOLD_ROWS = 0.001
73
+ cfg.SEGMENTATION.REMOVE_IOU_THRESHOLD_COLS = 0.001
74
+ cfg.SEGMENTATION.CELL_CATEGORY_ID = 12
75
+ cfg.SEGMENTATION.TABLE_NAME = LayoutType.TABLE
76
+ cfg.SEGMENTATION.PUBTABLES_CELL_NAMES = [
77
+ CellType.SPANNING,
78
+ CellType.ROW_HEADER,
79
+ CellType.COLUMN_HEADER,
80
+ CellType.PROJECTED_ROW_HEADER,
81
+ LayoutType.CELL,
82
+ ]
83
+ cfg.SEGMENTATION.PUBTABLES_SPANNING_CELL_NAMES = [
84
+ CellType.SPANNING,
85
+ CellType.ROW_HEADER,
86
+ CellType.COLUMN_HEADER,
87
+ CellType.PROJECTED_ROW_HEADER,
88
+ ]
89
+ cfg.SEGMENTATION.PUBTABLES_ITEM_NAMES = [LayoutType.ROW, LayoutType.COLUMN]
90
+ cfg.SEGMENTATION.PUBTABLES_SUB_ITEM_NAMES = [CellType.ROW_NUMBER, CellType.COLUMN_NUMBER]
91
+ cfg.SEGMENTATION.CELL_NAMES = [CellType.HEADER, CellType.BODY, LayoutType.CELL]
92
+ cfg.SEGMENTATION.ITEM_NAMES = [LayoutType.ROW, LayoutType.COLUMN]
93
+ cfg.SEGMENTATION.SUB_ITEM_NAMES = [CellType.ROW_NUMBER, CellType.COLUMN_NUMBER]
94
+
95
+ cfg.SEGMENTATION.STRETCH_RULE = "equal"
96
+
97
+ cfg.USE_TABLE_REFINEMENT = True
98
+ cfg.USE_PDF_MINER = False
99
+
100
+ cfg.PDF_MINER.X_TOLERANCE = 3
101
+ cfg.PDF_MINER.Y_TOLERANCE = 3
102
+
103
+ cfg.USE_OCR = True
104
+
105
+ cfg.OCR.USE_TESSERACT = True
106
+ cfg.OCR.USE_DOCTR = False
107
+ cfg.OCR.USE_TEXTRACT = False
108
+ cfg.OCR.CONFIG.TESSERACT = "dd/conf_tesseract.yaml"
109
+
110
+ cfg.OCR.WEIGHTS.DOCTR_WORD.TF = "doctr/db_resnet50/tf/db_resnet50-adcafc63.zip"
111
+ cfg.OCR.WEIGHTS.DOCTR_WORD.PT = "doctr/db_resnet50/pt/db_resnet50-ac60cadc.pt"
112
+ cfg.OCR.WEIGHTS.DOCTR_RECOGNITION.TF = "doctr/crnn_vgg16_bn/tf/crnn_vgg16_bn-76b7f2c6.zip"
113
+ cfg.OCR.WEIGHTS.DOCTR_RECOGNITION.PT = "doctr/crnn_vgg16_bn/pt/crnn_vgg16_bn-9762b0b0.pt"
114
+
115
+ cfg.TEXT_CONTAINER = LayoutType.WORD
116
+ cfg.WORD_MATCHING.PARENTAL_CATEGORIES = [
117
+ LayoutType.TEXT,
118
+ LayoutType.TITLE,
119
+ LayoutType.LIST,
120
+ LayoutType.CELL,
121
+ CellType.COLUMN_HEADER,
122
+ CellType.PROJECTED_ROW_HEADER,
123
+ CellType.SPANNING,
124
+ CellType.ROW_HEADER,
125
+ ]
126
+ cfg.WORD_MATCHING.RULE = "ioa"
127
+ cfg.WORD_MATCHING.THRESHOLD = 0.6
128
+ cfg.WORD_MATCHING.MAX_PARENT_ONLY = True
129
+
130
+ cfg.TEXT_ORDERING.TEXT_BLOCK_CATEGORIES = [
131
+ LayoutType.TEXT,
132
+ LayoutType.TITLE,
133
+ LayoutType.LIST,
134
+ LayoutType.CELL,
135
+ CellType.COLUMN_HEADER,
136
+ CellType.PROJECTED_ROW_HEADER,
137
+ CellType.SPANNING,
138
+ CellType.ROW_HEADER,
139
+ ]
140
+ cfg.TEXT_ORDERING.FLOATING_TEXT_BLOCK_CATEGORIES = [
141
+ LayoutType.TEXT,
142
+ LayoutType.TITLE,
143
+ LayoutType.LIST,
144
+ ]
145
+ cfg.TEXT_ORDERING.INCLUDE_RESIDUAL_TEXT_CONTAINER = False
146
+ cfg.TEXT_ORDERING.STARTING_POINT_TOLERANCE = 0.005
147
+ cfg.TEXT_ORDERING.BROKEN_LINE_TOLERANCE = 0.003
148
+ cfg.TEXT_ORDERING.HEIGHT_TOLERANCE = 2.0
149
+ cfg.TEXT_ORDERING.PARAGRAPH_BREAK = 0.035
150
+ cfg.freeze()