google-cloud-document_ai-v1 0.4.0 → 0.6.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/AUTHENTICATION.md +1 -1
- data/lib/google/cloud/document_ai/v1/document_processor_service/client.rb +1492 -58
- data/lib/google/cloud/document_ai/v1/document_processor_service/paths.rb +38 -0
- data/lib/google/cloud/document_ai/v1/version.rb +1 -1
- data/lib/google/cloud/documentai/v1/barcode_pb.rb +24 -0
- data/lib/google/cloud/documentai/v1/document_io_pb.rb +3 -0
- data/lib/google/cloud/documentai/v1/document_pb.rb +25 -0
- data/lib/google/cloud/documentai/v1/document_processor_service_pb.rb +140 -0
- data/lib/google/cloud/documentai/v1/document_processor_service_services_pb.rb +34 -0
- data/lib/google/cloud/documentai/v1/document_schema_pb.rb +60 -0
- data/lib/google/cloud/documentai/v1/operation_metadata_pb.rb +1 -0
- data/lib/google/cloud/documentai/v1/processor_pb.rb +73 -0
- data/lib/google/cloud/documentai/v1/processor_type_pb.rb +34 -0
- data/proto_docs/google/api/launch_stage.rb +71 -0
- data/proto_docs/google/cloud/documentai/v1/barcode.rb +73 -0
- data/proto_docs/google/cloud/documentai/v1/document.rb +158 -135
- data/proto_docs/google/cloud/documentai/v1/document_io.rb +7 -2
- data/proto_docs/google/cloud/documentai/v1/document_processor_service.rb +354 -6
- data/proto_docs/google/cloud/documentai/v1/document_schema.rb +155 -0
- data/proto_docs/google/cloud/documentai/v1/operation_metadata.rb +3 -0
- data/proto_docs/google/cloud/documentai/v1/processor.rb +174 -0
- data/proto_docs/google/cloud/documentai/v1/processor_type.rb +62 -0
- data/proto_docs/google/protobuf/empty.rb +0 -2
- data/proto_docs/google/protobuf/field_mask.rb +229 -0
- metadata +34 -4
@@ -21,11 +21,10 @@ module Google
|
|
21
21
|
module Cloud
|
22
22
|
module DocumentAI
|
23
23
|
module V1
|
24
|
-
# Document represents the canonical document resource in Document
|
25
|
-
#
|
26
|
-
#
|
27
|
-
#
|
28
|
-
# optimize for quality.
|
24
|
+
# Document represents the canonical document resource in Document AI. It is an
|
25
|
+
# interchange format that provides insights into documents and allows for
|
26
|
+
# collaboration between users and Document AI to iterate and optimize for
|
27
|
+
# quality.
|
29
28
|
# @!attribute [rw] uri
|
30
29
|
# @return [::String]
|
31
30
|
# Optional. Currently supports Google Cloud Storage URI of the form
|
@@ -48,24 +47,22 @@ module Google
|
|
48
47
|
# Optional. UTF-8 encoded text in reading order from the document.
|
49
48
|
# @!attribute [rw] text_styles
|
50
49
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Style>]
|
51
|
-
# Styles for the {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
50
|
+
# Placeholder. Styles for the {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
52
51
|
# @!attribute [rw] pages
|
53
52
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page>]
|
54
53
|
# Visual page layout for the {::Google::Cloud::DocumentAI::V1::Document Document}.
|
55
54
|
# @!attribute [rw] entities
|
56
55
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Entity>]
|
57
|
-
# A list of entities detected on
|
58
|
-
#
|
59
|
-
# shards, entities in this list may cross shard boundaries.
|
56
|
+
# A list of entities detected on {::Google::Cloud::DocumentAI::V1::Document#text Document.text}. For document shards,
|
57
|
+
# entities in this list may cross shard boundaries.
|
60
58
|
# @!attribute [rw] entity_relations
|
61
59
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::EntityRelation>]
|
62
|
-
# Relationship among
|
63
|
-
# {::Google::Cloud::DocumentAI::V1::Document#entities Document.entities}.
|
60
|
+
# Placeholder. Relationship among {::Google::Cloud::DocumentAI::V1::Document#entities Document.entities}.
|
64
61
|
# @!attribute [rw] text_changes
|
65
62
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::TextChange>]
|
66
|
-
# A list of text corrections made to
|
67
|
-
# used for annotating corrections to OCR mistakes. Text changes
|
68
|
-
# revision may not overlap with each other.
|
63
|
+
# Placeholder. A list of text corrections made to {::Google::Cloud::DocumentAI::V1::Document#text Document.text}. This
|
64
|
+
# is usually used for annotating corrections to OCR mistakes. Text changes
|
65
|
+
# for a given revision may not overlap with each other.
|
69
66
|
# @!attribute [rw] shard_info
|
70
67
|
# @return [::Google::Cloud::DocumentAI::V1::Document::ShardInfo]
|
71
68
|
# Information about the sharding if this document is sharded part of a larger
|
@@ -75,7 +72,7 @@ module Google
|
|
75
72
|
# Any error that occurred while processing this document.
|
76
73
|
# @!attribute [rw] revisions
|
77
74
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Revision>]
|
78
|
-
# Revision history of this document.
|
75
|
+
# Placeholder. Revision history of this document.
|
79
76
|
class Document
|
80
77
|
include ::Google::Protobuf::MessageExts
|
81
78
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
@@ -91,8 +88,7 @@ module Google
|
|
91
88
|
# Total number of shards.
|
92
89
|
# @!attribute [rw] text_offset
|
93
90
|
# @return [::Integer]
|
94
|
-
# The index of the first character in
|
95
|
-
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text} in the overall
|
91
|
+
# The index of the first character in {::Google::Cloud::DocumentAI::V1::Document#text Document.text} in the overall
|
96
92
|
# document global text.
|
97
93
|
class ShardInfo
|
98
94
|
include ::Google::Protobuf::MessageExts
|
@@ -103,8 +99,7 @@ module Google
|
|
103
99
|
# conventions as much as possible.
|
104
100
|
# @!attribute [rw] text_anchor
|
105
101
|
# @return [::Google::Cloud::DocumentAI::V1::Document::TextAnchor]
|
106
|
-
# Text anchor indexing into the
|
107
|
-
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
102
|
+
# Text anchor indexing into the {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
108
103
|
# @!attribute [rw] color
|
109
104
|
# @return [::Google::Type::Color]
|
110
105
|
# Text color.
|
@@ -127,6 +122,10 @@ module Google
|
|
127
122
|
# @!attribute [rw] font_size
|
128
123
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Style::FontSize]
|
129
124
|
# Font size.
|
125
|
+
# @!attribute [rw] font_family
|
126
|
+
# @return [::String]
|
127
|
+
# Font family such as `Arial`, `Times New Roman`.
|
128
|
+
# https://www.w3schools.com/cssref/pr_font_font-family.asp
|
130
129
|
class Style
|
131
130
|
include ::Google::Protobuf::MessageExts
|
132
131
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
@@ -147,11 +146,9 @@ module Google
|
|
147
146
|
# A page in a {::Google::Cloud::DocumentAI::V1::Document Document}.
|
148
147
|
# @!attribute [rw] page_number
|
149
148
|
# @return [::Integer]
|
150
|
-
# 1-based index for current
|
151
|
-
# {::Google::Cloud::DocumentAI::V1::Document
|
152
|
-
#
|
153
|
-
# taken out of a {::Google::Cloud::DocumentAI::V1::Document Document} for
|
154
|
-
# individual processing.
|
149
|
+
# 1-based index for current {::Google::Cloud::DocumentAI::V1::Document::Page Page} in a parent {::Google::Cloud::DocumentAI::V1::Document Document}.
|
150
|
+
# Useful when a page is taken out of a {::Google::Cloud::DocumentAI::V1::Document Document} for individual
|
151
|
+
# processing.
|
155
152
|
# @!attribute [rw] image
|
156
153
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Image]
|
157
154
|
# Rendered image for this page. This image is preprocessed to remove any
|
@@ -199,6 +196,12 @@ module Google
|
|
199
196
|
# @!attribute [rw] symbols
|
200
197
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::Symbol>]
|
201
198
|
# A list of visually detected symbols on the page.
|
199
|
+
# @!attribute [rw] detected_barcodes
|
200
|
+
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedBarcode>]
|
201
|
+
# A list of detected barcodes.
|
202
|
+
# @!attribute [rw] image_quality_scores
|
203
|
+
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::ImageQualityScores]
|
204
|
+
# Image Quality Scores.
|
202
205
|
# @!attribute [rw] provenance
|
203
206
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Provenance]
|
204
207
|
# The history of this page.
|
@@ -264,23 +267,18 @@ module Google
|
|
264
267
|
# Visual element describing a layout unit on a page.
|
265
268
|
# @!attribute [rw] text_anchor
|
266
269
|
# @return [::Google::Cloud::DocumentAI::V1::Document::TextAnchor]
|
267
|
-
# Text anchor indexing into the
|
268
|
-
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
270
|
+
# Text anchor indexing into the {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
269
271
|
# @!attribute [rw] confidence
|
270
272
|
# @return [::Float]
|
271
|
-
# Confidence of the current
|
272
|
-
#
|
273
|
-
#
|
274
|
-
# single token, a table, a visual element, etc. depending on context.
|
275
|
-
# Range [0, 1].
|
273
|
+
# Confidence of the current {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} within context of the object this
|
274
|
+
# layout is for. e.g. confidence can be for a single token, a table,
|
275
|
+
# a visual element, etc. depending on context. Range `[0, 1]`.
|
276
276
|
# @!attribute [rw] bounding_poly
|
277
277
|
# @return [::Google::Cloud::DocumentAI::V1::BoundingPoly]
|
278
|
-
# The bounding polygon for the
|
279
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout}.
|
278
|
+
# The bounding polygon for the {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout}.
|
280
279
|
# @!attribute [rw] orientation
|
281
280
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout::Orientation]
|
282
|
-
# Detected orientation for the
|
283
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout}.
|
281
|
+
# Detected orientation for the {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout}.
|
284
282
|
class Layout
|
285
283
|
include ::Google::Protobuf::MessageExts
|
286
284
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
@@ -311,8 +309,7 @@ module Google
|
|
311
309
|
# common line-spacing and orientation.
|
312
310
|
# @!attribute [rw] layout
|
313
311
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
314
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
315
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Block Block}.
|
312
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for {::Google::Cloud::DocumentAI::V1::Document::Page::Block Block}.
|
316
313
|
# @!attribute [rw] detected_languages
|
317
314
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
318
315
|
# A list of detected languages together with confidence.
|
@@ -327,8 +324,7 @@ module Google
|
|
327
324
|
# A collection of lines that a human would perceive as a paragraph.
|
328
325
|
# @!attribute [rw] layout
|
329
326
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
330
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
331
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Paragraph Paragraph}.
|
327
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for {::Google::Cloud::DocumentAI::V1::Document::Page::Paragraph Paragraph}.
|
332
328
|
# @!attribute [rw] detected_languages
|
333
329
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
334
330
|
# A list of detected languages together with confidence.
|
@@ -344,8 +340,7 @@ module Google
|
|
344
340
|
# Does not cross column boundaries, can be horizontal, vertical, etc.
|
345
341
|
# @!attribute [rw] layout
|
346
342
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
347
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
348
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Line Line}.
|
343
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for {::Google::Cloud::DocumentAI::V1::Document::Page::Line Line}.
|
349
344
|
# @!attribute [rw] detected_languages
|
350
345
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
351
346
|
# A list of detected languages together with confidence.
|
@@ -360,24 +355,21 @@ module Google
|
|
360
355
|
# A detected token.
|
361
356
|
# @!attribute [rw] layout
|
362
357
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
363
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
364
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Token Token}.
|
358
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for {::Google::Cloud::DocumentAI::V1::Document::Page::Token Token}.
|
365
359
|
# @!attribute [rw] detected_break
|
366
360
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Token::DetectedBreak]
|
367
|
-
# Detected break at the end of a
|
368
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Token Token}.
|
361
|
+
# Detected break at the end of a {::Google::Cloud::DocumentAI::V1::Document::Page::Token Token}.
|
369
362
|
# @!attribute [rw] detected_languages
|
370
363
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
371
364
|
# A list of detected languages together with confidence.
|
372
365
|
# @!attribute [rw] provenance
|
373
366
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Provenance]
|
374
|
-
# The
|
367
|
+
# The history of this annotation.
|
375
368
|
class Token
|
376
369
|
include ::Google::Protobuf::MessageExts
|
377
370
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
378
371
|
|
379
|
-
# Detected break at the end of a
|
380
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Token Token}.
|
372
|
+
# Detected break at the end of a {::Google::Cloud::DocumentAI::V1::Document::Page::Token Token}.
|
381
373
|
# @!attribute [rw] type
|
382
374
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Token::DetectedBreak::Type]
|
383
375
|
# Detected break type.
|
@@ -405,8 +397,7 @@ module Google
|
|
405
397
|
# A detected symbol.
|
406
398
|
# @!attribute [rw] layout
|
407
399
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
408
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
409
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Symbol Symbol}.
|
400
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for {::Google::Cloud::DocumentAI::V1::Document::Page::Symbol Symbol}.
|
410
401
|
# @!attribute [rw] detected_languages
|
411
402
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
412
403
|
# A list of detected languages together with confidence.
|
@@ -419,12 +410,10 @@ module Google
|
|
419
410
|
# page.
|
420
411
|
# @!attribute [rw] layout
|
421
412
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
422
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
423
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::VisualElement VisualElement}.
|
413
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for {::Google::Cloud::DocumentAI::V1::Document::Page::VisualElement VisualElement}.
|
424
414
|
# @!attribute [rw] type
|
425
415
|
# @return [::String]
|
426
|
-
# Type of the
|
427
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::VisualElement VisualElement}.
|
416
|
+
# Type of the {::Google::Cloud::DocumentAI::V1::Document::Page::VisualElement VisualElement}.
|
428
417
|
# @!attribute [rw] detected_languages
|
429
418
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
430
419
|
# A list of detected languages together with confidence.
|
@@ -436,8 +425,7 @@ module Google
|
|
436
425
|
# A table representation similar to HTML table structure.
|
437
426
|
# @!attribute [rw] layout
|
438
427
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
439
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
440
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Table Table}.
|
428
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for {::Google::Cloud::DocumentAI::V1::Document::Page::Table Table}.
|
441
429
|
# @!attribute [rw] header_rows
|
442
430
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::Table::TableRow>]
|
443
431
|
# Header rows of the table.
|
@@ -447,6 +435,9 @@ module Google
|
|
447
435
|
# @!attribute [rw] detected_languages
|
448
436
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
449
437
|
# A list of detected languages together with confidence.
|
438
|
+
# @!attribute [rw] provenance
|
439
|
+
# @return [::Google::Cloud::DocumentAI::V1::Document::Provenance]
|
440
|
+
# The history of this table.
|
450
441
|
class Table
|
451
442
|
include ::Google::Protobuf::MessageExts
|
452
443
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
@@ -463,8 +454,7 @@ module Google
|
|
463
454
|
# A cell representation inside the table.
|
464
455
|
# @!attribute [rw] layout
|
465
456
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
466
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
467
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Table::TableCell TableCell}.
|
457
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for {::Google::Cloud::DocumentAI::V1::Document::Page::Table::TableCell TableCell}.
|
468
458
|
# @!attribute [rw] row_span
|
469
459
|
# @return [::Integer]
|
470
460
|
# How many rows this cell spans.
|
@@ -483,13 +473,11 @@ module Google
|
|
483
473
|
# A form field detected on the page.
|
484
474
|
# @!attribute [rw] field_name
|
485
475
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
486
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for the
|
487
|
-
#
|
488
|
-
# e.g. `Address`, `Email`, `Grand total`, `Phone number`, etc.
|
476
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for the {::Google::Cloud::DocumentAI::V1::Document::Page::FormField FormField} name. e.g. `Address`, `Email`,
|
477
|
+
# `Grand total`, `Phone number`, etc.
|
489
478
|
# @!attribute [rw] field_value
|
490
479
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
491
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for the
|
492
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::FormField FormField} value.
|
480
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for the {::Google::Cloud::DocumentAI::V1::Document::Page::FormField FormField} value.
|
493
481
|
# @!attribute [rw] name_detected_languages
|
494
482
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
495
483
|
# A list of detected languages for name together with confidence.
|
@@ -503,6 +491,16 @@ module Google
|
|
503
491
|
# - blank (this indicates the field_value is normal text)
|
504
492
|
# - "unfilled_checkbox"
|
505
493
|
# - "filled_checkbox"
|
494
|
+
# @!attribute [rw] corrected_key_text
|
495
|
+
# @return [::String]
|
496
|
+
# Created for Labeling UI to export key text.
|
497
|
+
# If corrections were made to the text identified by the
|
498
|
+
# `field_name.text_anchor`, this field will contain the correction.
|
499
|
+
# @!attribute [rw] corrected_value_text
|
500
|
+
# @return [::String]
|
501
|
+
# Created for Labeling UI to export value text.
|
502
|
+
# If corrections were made to the text identified by the
|
503
|
+
# `field_value.text_anchor`, this field will contain the correction.
|
506
504
|
# @!attribute [rw] provenance
|
507
505
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Provenance]
|
508
506
|
# The history of this annotation.
|
@@ -511,19 +509,65 @@ module Google
|
|
511
509
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
512
510
|
end
|
513
511
|
|
512
|
+
# A detected barcode.
|
513
|
+
# @!attribute [rw] layout
|
514
|
+
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
515
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for {::Google::Cloud::DocumentAI::V1::Document::Page::DetectedBarcode DetectedBarcode}.
|
516
|
+
# @!attribute [rw] barcode
|
517
|
+
# @return [::Google::Cloud::DocumentAI::V1::Barcode]
|
518
|
+
# Detailed barcode information of the {::Google::Cloud::DocumentAI::V1::Document::Page::DetectedBarcode DetectedBarcode}.
|
519
|
+
class DetectedBarcode
|
520
|
+
include ::Google::Protobuf::MessageExts
|
521
|
+
extend ::Google::Protobuf::MessageExts::ClassMethods
|
522
|
+
end
|
523
|
+
|
514
524
|
# Detected language for a structural component.
|
515
525
|
# @!attribute [rw] language_code
|
516
526
|
# @return [::String]
|
517
|
-
# The BCP-47 language code, such as
|
527
|
+
# The BCP-47 language code, such as `en-US` or `sr-Latn`. For more
|
518
528
|
# information, see
|
519
529
|
# https://www.unicode.org/reports/tr35/#Unicode_locale_identifier.
|
520
530
|
# @!attribute [rw] confidence
|
521
531
|
# @return [::Float]
|
522
|
-
# Confidence of detected language. Range [0, 1]
|
532
|
+
# Confidence of detected language. Range `[0, 1]`.
|
523
533
|
class DetectedLanguage
|
524
534
|
include ::Google::Protobuf::MessageExts
|
525
535
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
526
536
|
end
|
537
|
+
|
538
|
+
# Image Quality Scores for the page image
|
539
|
+
# @!attribute [rw] quality_score
|
540
|
+
# @return [::Float]
|
541
|
+
# The overall quality score. Range `[0, 1]` where 1 is perfect quality.
|
542
|
+
# @!attribute [rw] detected_defects
|
543
|
+
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::ImageQualityScores::DetectedDefect>]
|
544
|
+
# A list of detected defects.
|
545
|
+
class ImageQualityScores
|
546
|
+
include ::Google::Protobuf::MessageExts
|
547
|
+
extend ::Google::Protobuf::MessageExts::ClassMethods
|
548
|
+
|
549
|
+
# Image Quality Defects
|
550
|
+
# @!attribute [rw] type
|
551
|
+
# @return [::String]
|
552
|
+
# Name of the defect type. Supported values are:
|
553
|
+
#
|
554
|
+
# - `quality/defect_blurry`
|
555
|
+
# - `quality/defect_noisy`
|
556
|
+
# - `quality/defect_dark`
|
557
|
+
# - `quality/defect_faint`
|
558
|
+
# - `quality/defect_text_too_small`
|
559
|
+
# - `quality/defect_document_cutoff`
|
560
|
+
# - `quality/defect_text_cutoff`
|
561
|
+
# - `quality/defect_glare`
|
562
|
+
# @!attribute [rw] confidence
|
563
|
+
# @return [::Float]
|
564
|
+
# Confidence of detected defect. Range `[0, 1]` where 1 indicates
|
565
|
+
# strong confidence of that the defect exists.
|
566
|
+
class DetectedDefect
|
567
|
+
include ::Google::Protobuf::MessageExts
|
568
|
+
extend ::Google::Protobuf::MessageExts::ClassMethods
|
569
|
+
end
|
570
|
+
end
|
527
571
|
end
|
528
572
|
|
529
573
|
# An entity that could be a phrase in the text or a property that belongs to
|
@@ -532,46 +576,43 @@ module Google
|
|
532
576
|
# @!attribute [rw] text_anchor
|
533
577
|
# @return [::Google::Cloud::DocumentAI::V1::Document::TextAnchor]
|
534
578
|
# Optional. Provenance of the entity.
|
535
|
-
# Text anchor indexing into the
|
536
|
-
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
579
|
+
# Text anchor indexing into the {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
537
580
|
# @!attribute [rw] type
|
538
581
|
# @return [::String]
|
539
582
|
# Required. Entity type from a schema e.g. `Address`.
|
540
583
|
# @!attribute [rw] mention_text
|
541
584
|
# @return [::String]
|
542
|
-
# Optional. Text value
|
543
|
-
# the entity is not present in the document, this field will be empty.
|
585
|
+
# Optional. Text value of the entity e.g. `1600 Amphitheatre Pkwy`.
|
544
586
|
# @!attribute [rw] mention_id
|
545
587
|
# @return [::String]
|
546
588
|
# Optional. Deprecated. Use `id` field instead.
|
547
589
|
# @!attribute [rw] confidence
|
548
590
|
# @return [::Float]
|
549
|
-
# Optional. Confidence of detected Schema entity. Range [0, 1]
|
591
|
+
# Optional. Confidence of detected Schema entity. Range `[0, 1]`.
|
550
592
|
# @!attribute [rw] page_anchor
|
551
593
|
# @return [::Google::Cloud::DocumentAI::V1::Document::PageAnchor]
|
552
|
-
# Optional. Represents the provenance of this entity wrt. the location on
|
553
|
-
#
|
594
|
+
# Optional. Represents the provenance of this entity wrt. the location on the
|
595
|
+
# page where it was found.
|
554
596
|
# @!attribute [rw] id
|
555
597
|
# @return [::String]
|
556
598
|
# Optional. Canonical id. This will be a unique value in the entity list
|
557
599
|
# for this document.
|
558
600
|
# @!attribute [rw] normalized_value
|
559
601
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Entity::NormalizedValue]
|
560
|
-
# Optional. Normalized entity value. Absent if the extracted value could
|
561
|
-
#
|
602
|
+
# Optional. Normalized entity value. Absent if the extracted value could not be
|
603
|
+
# converted or the type (e.g. address) is not supported for certain
|
562
604
|
# parsers. This field is also only populated for certain supported document
|
563
605
|
# types.
|
564
606
|
# @!attribute [rw] properties
|
565
607
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Entity>]
|
566
|
-
# Optional. Entities can be nested to form a hierarchical data structure
|
567
|
-
#
|
608
|
+
# Optional. Entities can be nested to form a hierarchical data structure representing
|
609
|
+
# the content in the document.
|
568
610
|
# @!attribute [rw] provenance
|
569
611
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Provenance]
|
570
612
|
# Optional. The history of this annotation.
|
571
613
|
# @!attribute [rw] redacted
|
572
614
|
# @return [::Boolean]
|
573
|
-
# Optional. Whether the entity will be redacted for de-identification
|
574
|
-
# purposes.
|
615
|
+
# Optional. Whether the entity will be redacted for de-identification purposes.
|
575
616
|
class Entity
|
576
617
|
include ::Google::Protobuf::MessageExts
|
577
618
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
@@ -597,6 +638,12 @@ module Google
|
|
597
638
|
# @return [::Boolean]
|
598
639
|
# Boolean value. Can be used for entities with binary values, or for
|
599
640
|
# checkboxes.
|
641
|
+
# @!attribute [rw] integer_value
|
642
|
+
# @return [::Integer]
|
643
|
+
# Integer value.
|
644
|
+
# @!attribute [rw] float_value
|
645
|
+
# @return [::Float]
|
646
|
+
# Float value.
|
600
647
|
# @!attribute [rw] text
|
601
648
|
# @return [::String]
|
602
649
|
# Optional. An optional field to store a normalized string.
|
@@ -606,6 +653,7 @@ module Google
|
|
606
653
|
# or int normalized text by default.
|
607
654
|
#
|
608
655
|
# Below are sample formats mapped to structured values.
|
656
|
+
#
|
609
657
|
# - Money/Currency type (`money_value`) is in the ISO 4217 text format.
|
610
658
|
# - Date type (`date_value`) is in the ISO 8601 text format.
|
611
659
|
# - Datetime type (`datetime_value`) is in the ISO 8601 text format.
|
@@ -615,8 +663,7 @@ module Google
|
|
615
663
|
end
|
616
664
|
end
|
617
665
|
|
618
|
-
# Relationship between
|
619
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Entity Entities}.
|
666
|
+
# Relationship between {::Google::Cloud::DocumentAI::V1::Document::Entity Entities}.
|
620
667
|
# @!attribute [rw] subject_id
|
621
668
|
# @return [::String]
|
622
669
|
# Subject entity id.
|
@@ -631,12 +678,10 @@ module Google
|
|
631
678
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
632
679
|
end
|
633
680
|
|
634
|
-
# Text reference indexing into the
|
635
|
-
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
681
|
+
# Text reference indexing into the {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
636
682
|
# @!attribute [rw] text_segments
|
637
683
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::TextAnchor::TextSegment>]
|
638
|
-
# The text segments from the
|
639
|
-
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
684
|
+
# The text segments from the {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
640
685
|
# @!attribute [rw] content
|
641
686
|
# @return [::String]
|
642
687
|
# Contains the content of the text span so that users do
|
@@ -646,20 +691,15 @@ module Google
|
|
646
691
|
include ::Google::Protobuf::MessageExts
|
647
692
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
648
693
|
|
649
|
-
# A text segment in the
|
650
|
-
#
|
651
|
-
#
|
652
|
-
# document shard for large sharded documents. See
|
653
|
-
# {::Google::Cloud::DocumentAI::V1::Document::ShardInfo#text_offset ShardInfo.text_offset}
|
694
|
+
# A text segment in the {::Google::Cloud::DocumentAI::V1::Document#text Document.text}. The indices may be out of bounds
|
695
|
+
# which indicate that the text extends into another document shard for
|
696
|
+
# large sharded documents. See {::Google::Cloud::DocumentAI::V1::Document::ShardInfo#text_offset ShardInfo.text_offset}
|
654
697
|
# @!attribute [rw] start_index
|
655
698
|
# @return [::Integer]
|
656
|
-
# {::Google::Cloud::DocumentAI::V1::Document::TextAnchor::TextSegment TextSegment}
|
657
|
-
# start UTF-8 char index in the
|
658
|
-
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
699
|
+
# {::Google::Cloud::DocumentAI::V1::Document::TextAnchor::TextSegment TextSegment} start UTF-8 char index in the {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
659
700
|
# @!attribute [rw] end_index
|
660
701
|
# @return [::Integer]
|
661
|
-
# {::Google::Cloud::DocumentAI::V1::Document::TextAnchor::TextSegment TextSegment}
|
662
|
-
# half open end UTF-8 char index in the
|
702
|
+
# {::Google::Cloud::DocumentAI::V1::Document::TextAnchor::TextSegment TextSegment} half open end UTF-8 char index in the
|
663
703
|
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
664
704
|
class TextSegment
|
665
705
|
include ::Google::Protobuf::MessageExts
|
@@ -667,10 +707,9 @@ module Google
|
|
667
707
|
end
|
668
708
|
end
|
669
709
|
|
670
|
-
# Referencing the visual context of the entity in the
|
671
|
-
#
|
672
|
-
#
|
673
|
-
# reference specific layout element types.
|
710
|
+
# Referencing the visual context of the entity in the {::Google::Cloud::DocumentAI::V1::Document#pages Document.pages}.
|
711
|
+
# Page anchors can be cross-page, consist of multiple bounding polygons and
|
712
|
+
# optionally reference specific layout element types.
|
674
713
|
# @!attribute [rw] page_refs
|
675
714
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::PageAnchor::PageRef>]
|
676
715
|
# One or more references to visual page elements
|
@@ -681,29 +720,22 @@ module Google
|
|
681
720
|
# Represents a weak reference to a page element within a document.
|
682
721
|
# @!attribute [rw] page
|
683
722
|
# @return [::Integer]
|
684
|
-
# Required. Index into the
|
685
|
-
#
|
686
|
-
#
|
687
|
-
# related page element. This field is skipped when its value is the
|
688
|
-
# default 0. See
|
723
|
+
# Required. Index into the {::Google::Cloud::DocumentAI::V1::Document#pages Document.pages} element, for example using
|
724
|
+
# [Document.pages][page_refs.page] to locate the related page element.
|
725
|
+
# This field is skipped when its value is the default 0. See
|
689
726
|
# https://developers.google.com/protocol-buffers/docs/proto3#json.
|
690
727
|
# @!attribute [rw] layout_type
|
691
728
|
# @return [::Google::Cloud::DocumentAI::V1::Document::PageAnchor::PageRef::LayoutType]
|
692
|
-
# Optional. The type of the layout element that is being referenced if
|
693
|
-
# any.
|
729
|
+
# Optional. The type of the layout element that is being referenced if any.
|
694
730
|
# @!attribute [rw] layout_id
|
695
731
|
# @return [::String]
|
696
|
-
# Optional. Deprecated. Use
|
697
|
-
# {::Google::Cloud::DocumentAI::V1::Document::PageAnchor::PageRef#bounding_poly PageRef.bounding_poly}
|
698
|
-
# instead.
|
732
|
+
# Optional. Deprecated. Use {::Google::Cloud::DocumentAI::V1::Document::PageAnchor::PageRef#bounding_poly PageRef.bounding_poly} instead.
|
699
733
|
# @!attribute [rw] bounding_poly
|
700
734
|
# @return [::Google::Cloud::DocumentAI::V1::BoundingPoly]
|
701
|
-
# Optional. Identifies the bounding polygon of a layout element on the
|
702
|
-
# page.
|
735
|
+
# Optional. Identifies the bounding polygon of a layout element on the page.
|
703
736
|
# @!attribute [rw] confidence
|
704
737
|
# @return [::Float]
|
705
|
-
# Optional. Confidence of detected page element, if applicable. Range [0,
|
706
|
-
# 1].
|
738
|
+
# Optional. Confidence of detected page element, if applicable. Range `[0, 1]`.
|
707
739
|
class PageRef
|
708
740
|
include ::Google::Protobuf::MessageExts
|
709
741
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
@@ -713,38 +745,25 @@ module Google
|
|
713
745
|
# Layout Unspecified.
|
714
746
|
LAYOUT_TYPE_UNSPECIFIED = 0
|
715
747
|
|
716
|
-
# References a
|
717
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page#blocks Page.blocks}
|
718
|
-
# element.
|
748
|
+
# References a {::Google::Cloud::DocumentAI::V1::Document::Page#blocks Page.blocks} element.
|
719
749
|
BLOCK = 1
|
720
750
|
|
721
|
-
# References a
|
722
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page#paragraphs Page.paragraphs}
|
723
|
-
# element.
|
751
|
+
# References a {::Google::Cloud::DocumentAI::V1::Document::Page#paragraphs Page.paragraphs} element.
|
724
752
|
PARAGRAPH = 2
|
725
753
|
|
726
|
-
# References a
|
727
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page#lines Page.lines} element.
|
754
|
+
# References a {::Google::Cloud::DocumentAI::V1::Document::Page#lines Page.lines} element.
|
728
755
|
LINE = 3
|
729
756
|
|
730
|
-
# References a
|
731
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page#tokens Page.tokens}
|
732
|
-
# element.
|
757
|
+
# References a {::Google::Cloud::DocumentAI::V1::Document::Page#tokens Page.tokens} element.
|
733
758
|
TOKEN = 4
|
734
759
|
|
735
|
-
# References a
|
736
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page#visual_elements Page.visual_elements}
|
737
|
-
# element.
|
760
|
+
# References a {::Google::Cloud::DocumentAI::V1::Document::Page#visual_elements Page.visual_elements} element.
|
738
761
|
VISUAL_ELEMENT = 5
|
739
762
|
|
740
|
-
# Refrrences a
|
741
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page#tables Page.tables}
|
742
|
-
# element.
|
763
|
+
# Refrrences a {::Google::Cloud::DocumentAI::V1::Document::Page#tables Page.tables} element.
|
743
764
|
TABLE = 6
|
744
765
|
|
745
|
-
# References a
|
746
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page#form_fields Page.form_fields}
|
747
|
-
# element.
|
766
|
+
# References a {::Google::Cloud::DocumentAI::V1::Document::Page#form_fields Page.form_fields} element.
|
748
767
|
FORM_FIELD = 7
|
749
768
|
end
|
750
769
|
end
|
@@ -830,6 +849,11 @@ module Google
|
|
830
849
|
# The revisions that this revision is based on. This can include one or
|
831
850
|
# more parent (when documents are merged.) This field represents the
|
832
851
|
# index into the `revisions` field.
|
852
|
+
# @!attribute [rw] parent_ids
|
853
|
+
# @return [::Array<::String>]
|
854
|
+
# The revisions that this revision is based on. Must include all the ids
|
855
|
+
# that have anything to do with this revision - eg. there are
|
856
|
+
# `provenance.parent.revision` fields that index into this field.
|
833
857
|
# @!attribute [rw] create_time
|
834
858
|
# @return [::Google::Protobuf::Timestamp]
|
835
859
|
# The time that the revision was created.
|
@@ -858,9 +882,8 @@ module Google
|
|
858
882
|
# @!attribute [rw] text_anchor
|
859
883
|
# @return [::Google::Cloud::DocumentAI::V1::Document::TextAnchor]
|
860
884
|
# Provenance of the correction.
|
861
|
-
# Text anchor indexing into the
|
862
|
-
#
|
863
|
-
# only be a single `TextAnchor.text_segments` element. If the start and
|
885
|
+
# Text anchor indexing into the {::Google::Cloud::DocumentAI::V1::Document#text Document.text}. There can only be a
|
886
|
+
# single `TextAnchor.text_segments` element. If the start and
|
864
887
|
# end index of the text segment are the same, the text change is inserted
|
865
888
|
# before that index.
|
866
889
|
# @!attribute [rw] changed_text
|
@@ -28,7 +28,7 @@ module Google
|
|
28
28
|
# @!attribute [rw] mime_type
|
29
29
|
# @return [::String]
|
30
30
|
# An IANA MIME type (RFC6838) indicating the nature and format of the
|
31
|
-
#
|
31
|
+
# {::Google::Cloud::DocumentAI::V1::RawDocument#content content}.
|
32
32
|
class RawDocument
|
33
33
|
include ::Google::Protobuf::MessageExts
|
34
34
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
@@ -67,7 +67,7 @@ module Google
|
|
67
67
|
# The common config to specify a set of documents used as input.
|
68
68
|
# @!attribute [rw] gcs_prefix
|
69
69
|
# @return [::Google::Cloud::DocumentAI::V1::GcsPrefix]
|
70
|
-
# The set of documents that match the specified Cloud Storage
|
70
|
+
# The set of documents that match the specified Cloud Storage `gcs_prefix`.
|
71
71
|
# @!attribute [rw] gcs_documents
|
72
72
|
# @return [::Google::Cloud::DocumentAI::V1::GcsDocuments]
|
73
73
|
# The set of documents individually specified on Cloud Storage.
|
@@ -89,6 +89,11 @@ module Google
|
|
89
89
|
# @!attribute [rw] gcs_uri
|
90
90
|
# @return [::String]
|
91
91
|
# The Cloud Storage uri (a directory) of the output.
|
92
|
+
# @!attribute [rw] field_mask
|
93
|
+
# @return [::Google::Protobuf::FieldMask]
|
94
|
+
# Specifies which fields to include in the output documents.
|
95
|
+
# Only supports top level document and pages field so it must be in the
|
96
|
+
# form of `{document_field_name}` or `pages.{page_field_name}`.
|
92
97
|
class GcsOutputConfig
|
93
98
|
include ::Google::Protobuf::MessageExts
|
94
99
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|