google-cloud-document_ai-v1 0.2.4 → 0.3.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/README.md +1 -1
- data/lib/google/cloud/document_ai/v1/version.rb +1 -1
- data/lib/google/cloud/documentai/v1/document_pb.rb +7 -1
- data/lib/google/cloud/documentai/v1/document_processor_service_pb.rb +1 -0
- data/proto_docs/google/cloud/documentai/v1/document.rb +158 -81
- data/proto_docs/google/cloud/documentai/v1/document_processor_service.rb +3 -0
- metadata +3 -3
checksums.yaml
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
---
|
|
2
2
|
SHA256:
|
|
3
|
-
metadata.gz:
|
|
4
|
-
data.tar.gz:
|
|
3
|
+
metadata.gz: db1b8b3e4a2a0e11379da959689aab8db44a9e531a7dab68763e71fba5cb4a8b
|
|
4
|
+
data.tar.gz: 4328522a4c2e5400b29affed08535e9dd13271113920ee8d9552a20a7cec2b5f
|
|
5
5
|
SHA512:
|
|
6
|
-
metadata.gz:
|
|
7
|
-
data.tar.gz:
|
|
6
|
+
metadata.gz: 14ac2abace39ad77e14fa79cd2621b1f62e89a313aca9cdfe8511591277bd078dd4680de95dac79c36c5157fda31fadde180c5f3eac51c74e0324633a04edebc
|
|
7
|
+
data.tar.gz: 3a7bc1163f4a53739ca45a393e57c975a1eb59858399104d51f02d21b41238fa2c1055938aef0a639fbc774fdb7f653137b0ae748a6754d91bc389a49cff2a71
|
data/README.md
CHANGED
|
@@ -37,7 +37,7 @@ request = ::Google::Cloud::DocumentAI::V1::ProcessRequest.new # (request fields
|
|
|
37
37
|
response = client.process_document request
|
|
38
38
|
```
|
|
39
39
|
|
|
40
|
-
View the [Client Library Documentation](https://
|
|
40
|
+
View the [Client Library Documentation](https://cloud.google.com/ruby/docs/reference/google-cloud-document_ai-v1/latest)
|
|
41
41
|
for class and method documentation.
|
|
42
42
|
|
|
43
43
|
See also the [Product Documentation](https://cloud.google.com/document-ai/)
|
|
@@ -1,6 +1,7 @@
|
|
|
1
1
|
# Generated by the protocol buffer compiler. DO NOT EDIT!
|
|
2
2
|
# source: google/cloud/documentai/v1/document.proto
|
|
3
3
|
|
|
4
|
+
require 'google/api/annotations_pb'
|
|
4
5
|
require 'google/api/field_behavior_pb'
|
|
5
6
|
require 'google/cloud/documentai/v1/geometry_pb'
|
|
6
7
|
require 'google/protobuf/timestamp_pb'
|
|
@@ -10,7 +11,6 @@ require 'google/type/date_pb'
|
|
|
10
11
|
require 'google/type/datetime_pb'
|
|
11
12
|
require 'google/type/money_pb'
|
|
12
13
|
require 'google/type/postal_address_pb'
|
|
13
|
-
require 'google/api/annotations_pb'
|
|
14
14
|
require 'google/protobuf'
|
|
15
15
|
|
|
16
16
|
Google::Protobuf::DescriptorPool.generated_pool.build do
|
|
@@ -63,6 +63,7 @@ Google::Protobuf::DescriptorPool.generated_pool.build do
|
|
|
63
63
|
repeated :visual_elements, :message, 9, "google.cloud.documentai.v1.Document.Page.VisualElement"
|
|
64
64
|
repeated :tables, :message, 10, "google.cloud.documentai.v1.Document.Page.Table"
|
|
65
65
|
repeated :form_fields, :message, 11, "google.cloud.documentai.v1.Document.Page.FormField"
|
|
66
|
+
repeated :symbols, :message, 12, "google.cloud.documentai.v1.Document.Page.Symbol"
|
|
66
67
|
optional :provenance, :message, 16, "google.cloud.documentai.v1.Document.Provenance"
|
|
67
68
|
end
|
|
68
69
|
add_message "google.cloud.documentai.v1.Document.Page.Dimension" do
|
|
@@ -125,6 +126,10 @@ Google::Protobuf::DescriptorPool.generated_pool.build do
|
|
|
125
126
|
value :WIDE_SPACE, 2
|
|
126
127
|
value :HYPHEN, 3
|
|
127
128
|
end
|
|
129
|
+
add_message "google.cloud.documentai.v1.Document.Page.Symbol" do
|
|
130
|
+
optional :layout, :message, 1, "google.cloud.documentai.v1.Document.Page.Layout"
|
|
131
|
+
repeated :detected_languages, :message, 2, "google.cloud.documentai.v1.Document.Page.DetectedLanguage"
|
|
132
|
+
end
|
|
128
133
|
add_message "google.cloud.documentai.v1.Document.Page.VisualElement" do
|
|
129
134
|
optional :layout, :message, 1, "google.cloud.documentai.v1.Document.Page.Layout"
|
|
130
135
|
optional :type, :string, 2
|
|
@@ -275,6 +280,7 @@ module Google
|
|
|
275
280
|
Document::Page::Token = ::Google::Protobuf::DescriptorPool.generated_pool.lookup("google.cloud.documentai.v1.Document.Page.Token").msgclass
|
|
276
281
|
Document::Page::Token::DetectedBreak = ::Google::Protobuf::DescriptorPool.generated_pool.lookup("google.cloud.documentai.v1.Document.Page.Token.DetectedBreak").msgclass
|
|
277
282
|
Document::Page::Token::DetectedBreak::Type = ::Google::Protobuf::DescriptorPool.generated_pool.lookup("google.cloud.documentai.v1.Document.Page.Token.DetectedBreak.Type").enummodule
|
|
283
|
+
Document::Page::Symbol = ::Google::Protobuf::DescriptorPool.generated_pool.lookup("google.cloud.documentai.v1.Document.Page.Symbol").msgclass
|
|
278
284
|
Document::Page::VisualElement = ::Google::Protobuf::DescriptorPool.generated_pool.lookup("google.cloud.documentai.v1.Document.Page.VisualElement").msgclass
|
|
279
285
|
Document::Page::Table = ::Google::Protobuf::DescriptorPool.generated_pool.lookup("google.cloud.documentai.v1.Document.Page.Table").msgclass
|
|
280
286
|
Document::Page::Table::TableRow = ::Google::Protobuf::DescriptorPool.generated_pool.lookup("google.cloud.documentai.v1.Document.Page.Table.TableRow").msgclass
|
|
@@ -87,6 +87,7 @@ Google::Protobuf::DescriptorPool.generated_pool.build do
|
|
|
87
87
|
end
|
|
88
88
|
add_message "google.cloud.documentai.v1.ReviewDocumentOperationMetadata" do
|
|
89
89
|
optional :common_metadata, :message, 5, "google.cloud.documentai.v1.CommonOperationMetadata"
|
|
90
|
+
optional :question_id, :string, 6
|
|
90
91
|
end
|
|
91
92
|
end
|
|
92
93
|
end
|
|
@@ -54,11 +54,13 @@ module Google
|
|
|
54
54
|
# Visual page layout for the {::Google::Cloud::DocumentAI::V1::Document Document}.
|
|
55
55
|
# @!attribute [rw] entities
|
|
56
56
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Entity>]
|
|
57
|
-
# A list of entities detected on
|
|
58
|
-
#
|
|
57
|
+
# A list of entities detected on
|
|
58
|
+
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}. For document
|
|
59
|
+
# shards, entities in this list may cross shard boundaries.
|
|
59
60
|
# @!attribute [rw] entity_relations
|
|
60
61
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::EntityRelation>]
|
|
61
|
-
# Relationship among
|
|
62
|
+
# Relationship among
|
|
63
|
+
# {::Google::Cloud::DocumentAI::V1::Document#entities Document.entities}.
|
|
62
64
|
# @!attribute [rw] text_changes
|
|
63
65
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::TextChange>]
|
|
64
66
|
# A list of text corrections made to [Document.text]. This is usually
|
|
@@ -89,7 +91,8 @@ module Google
|
|
|
89
91
|
# Total number of shards.
|
|
90
92
|
# @!attribute [rw] text_offset
|
|
91
93
|
# @return [::Integer]
|
|
92
|
-
# The index of the first character in
|
|
94
|
+
# The index of the first character in
|
|
95
|
+
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text} in the overall
|
|
93
96
|
# document global text.
|
|
94
97
|
class ShardInfo
|
|
95
98
|
include ::Google::Protobuf::MessageExts
|
|
@@ -100,7 +103,8 @@ module Google
|
|
|
100
103
|
# conventions as much as possible.
|
|
101
104
|
# @!attribute [rw] text_anchor
|
|
102
105
|
# @return [::Google::Cloud::DocumentAI::V1::Document::TextAnchor]
|
|
103
|
-
# Text anchor indexing into the
|
|
106
|
+
# Text anchor indexing into the
|
|
107
|
+
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
|
104
108
|
# @!attribute [rw] color
|
|
105
109
|
# @return [::Google::Type::Color]
|
|
106
110
|
# Text color.
|
|
@@ -143,9 +147,11 @@ module Google
|
|
|
143
147
|
# A page in a {::Google::Cloud::DocumentAI::V1::Document Document}.
|
|
144
148
|
# @!attribute [rw] page_number
|
|
145
149
|
# @return [::Integer]
|
|
146
|
-
# 1-based index for current
|
|
147
|
-
#
|
|
148
|
-
#
|
|
150
|
+
# 1-based index for current
|
|
151
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page Page} in a parent
|
|
152
|
+
# {::Google::Cloud::DocumentAI::V1::Document Document}. Useful when a page is
|
|
153
|
+
# taken out of a {::Google::Cloud::DocumentAI::V1::Document Document} for
|
|
154
|
+
# individual processing.
|
|
149
155
|
# @!attribute [rw] image
|
|
150
156
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Image]
|
|
151
157
|
# Rendered image for this page. This image is preprocessed to remove any
|
|
@@ -190,6 +196,9 @@ module Google
|
|
|
190
196
|
# @!attribute [rw] form_fields
|
|
191
197
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::FormField>]
|
|
192
198
|
# A list of visually detected form fields on the page.
|
|
199
|
+
# @!attribute [rw] symbols
|
|
200
|
+
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::Symbol>]
|
|
201
|
+
# A list of visually detected symbols on the page.
|
|
193
202
|
# @!attribute [rw] provenance
|
|
194
203
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Provenance]
|
|
195
204
|
# The history of this page.
|
|
@@ -255,18 +264,23 @@ module Google
|
|
|
255
264
|
# Visual element describing a layout unit on a page.
|
|
256
265
|
# @!attribute [rw] text_anchor
|
|
257
266
|
# @return [::Google::Cloud::DocumentAI::V1::Document::TextAnchor]
|
|
258
|
-
# Text anchor indexing into the
|
|
267
|
+
# Text anchor indexing into the
|
|
268
|
+
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
|
259
269
|
# @!attribute [rw] confidence
|
|
260
270
|
# @return [::Float]
|
|
261
|
-
# Confidence of the current
|
|
262
|
-
#
|
|
263
|
-
#
|
|
271
|
+
# Confidence of the current
|
|
272
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} within
|
|
273
|
+
# context of the object this layout is for. e.g. confidence can be for a
|
|
274
|
+
# single token, a table, a visual element, etc. depending on context.
|
|
275
|
+
# Range [0, 1].
|
|
264
276
|
# @!attribute [rw] bounding_poly
|
|
265
277
|
# @return [::Google::Cloud::DocumentAI::V1::BoundingPoly]
|
|
266
|
-
# The bounding polygon for the
|
|
278
|
+
# The bounding polygon for the
|
|
279
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout}.
|
|
267
280
|
# @!attribute [rw] orientation
|
|
268
281
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout::Orientation]
|
|
269
|
-
# Detected orientation for the
|
|
282
|
+
# Detected orientation for the
|
|
283
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout}.
|
|
270
284
|
class Layout
|
|
271
285
|
include ::Google::Protobuf::MessageExts
|
|
272
286
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
|
@@ -297,7 +311,8 @@ module Google
|
|
|
297
311
|
# common line-spacing and orientation.
|
|
298
312
|
# @!attribute [rw] layout
|
|
299
313
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
|
300
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
314
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
315
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Block Block}.
|
|
301
316
|
# @!attribute [rw] detected_languages
|
|
302
317
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
|
303
318
|
# A list of detected languages together with confidence.
|
|
@@ -312,7 +327,8 @@ module Google
|
|
|
312
327
|
# A collection of lines that a human would perceive as a paragraph.
|
|
313
328
|
# @!attribute [rw] layout
|
|
314
329
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
|
315
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
330
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
331
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Paragraph Paragraph}.
|
|
316
332
|
# @!attribute [rw] detected_languages
|
|
317
333
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
|
318
334
|
# A list of detected languages together with confidence.
|
|
@@ -328,7 +344,8 @@ module Google
|
|
|
328
344
|
# Does not cross column boundaries, can be horizontal, vertical, etc.
|
|
329
345
|
# @!attribute [rw] layout
|
|
330
346
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
|
331
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
347
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
348
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Line Line}.
|
|
332
349
|
# @!attribute [rw] detected_languages
|
|
333
350
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
|
334
351
|
# A list of detected languages together with confidence.
|
|
@@ -343,10 +360,12 @@ module Google
|
|
|
343
360
|
# A detected token.
|
|
344
361
|
# @!attribute [rw] layout
|
|
345
362
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
|
346
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
363
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
364
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Token Token}.
|
|
347
365
|
# @!attribute [rw] detected_break
|
|
348
366
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Token::DetectedBreak]
|
|
349
|
-
# Detected break at the end of a
|
|
367
|
+
# Detected break at the end of a
|
|
368
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Token Token}.
|
|
350
369
|
# @!attribute [rw] detected_languages
|
|
351
370
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
|
352
371
|
# A list of detected languages together with confidence.
|
|
@@ -357,7 +376,8 @@ module Google
|
|
|
357
376
|
include ::Google::Protobuf::MessageExts
|
|
358
377
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
|
359
378
|
|
|
360
|
-
# Detected break at the end of a
|
|
379
|
+
# Detected break at the end of a
|
|
380
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Token Token}.
|
|
361
381
|
# @!attribute [rw] type
|
|
362
382
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Token::DetectedBreak::Type]
|
|
363
383
|
# Detected break type.
|
|
@@ -382,14 +402,29 @@ module Google
|
|
|
382
402
|
end
|
|
383
403
|
end
|
|
384
404
|
|
|
405
|
+
# A detected symbol.
|
|
406
|
+
# @!attribute [rw] layout
|
|
407
|
+
# @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}.
|
|
410
|
+
# @!attribute [rw] detected_languages
|
|
411
|
+
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
|
412
|
+
# A list of detected languages together with confidence.
|
|
413
|
+
class Symbol
|
|
414
|
+
include ::Google::Protobuf::MessageExts
|
|
415
|
+
extend ::Google::Protobuf::MessageExts::ClassMethods
|
|
416
|
+
end
|
|
417
|
+
|
|
385
418
|
# Detected non-text visual elements e.g. checkbox, signature etc. on the
|
|
386
419
|
# page.
|
|
387
420
|
# @!attribute [rw] layout
|
|
388
421
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
|
389
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
422
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
423
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::VisualElement VisualElement}.
|
|
390
424
|
# @!attribute [rw] type
|
|
391
425
|
# @return [::String]
|
|
392
|
-
# Type of the
|
|
426
|
+
# Type of the
|
|
427
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::VisualElement VisualElement}.
|
|
393
428
|
# @!attribute [rw] detected_languages
|
|
394
429
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
|
395
430
|
# A list of detected languages together with confidence.
|
|
@@ -401,7 +436,8 @@ module Google
|
|
|
401
436
|
# A table representation similar to HTML table structure.
|
|
402
437
|
# @!attribute [rw] layout
|
|
403
438
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
|
404
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
439
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
440
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Table Table}.
|
|
405
441
|
# @!attribute [rw] header_rows
|
|
406
442
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::Table::TableRow>]
|
|
407
443
|
# Header rows of the table.
|
|
@@ -427,7 +463,8 @@ module Google
|
|
|
427
463
|
# A cell representation inside the table.
|
|
428
464
|
# @!attribute [rw] layout
|
|
429
465
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
|
430
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
466
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for
|
|
467
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Table::TableCell TableCell}.
|
|
431
468
|
# @!attribute [rw] row_span
|
|
432
469
|
# @return [::Integer]
|
|
433
470
|
# How many rows this cell spans.
|
|
@@ -446,11 +483,13 @@ module Google
|
|
|
446
483
|
# A form field detected on the page.
|
|
447
484
|
# @!attribute [rw] field_name
|
|
448
485
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
|
449
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for the
|
|
450
|
-
#
|
|
486
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for the
|
|
487
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::FormField FormField} name.
|
|
488
|
+
# e.g. `Address`, `Email`, `Grand total`, `Phone number`, etc.
|
|
451
489
|
# @!attribute [rw] field_value
|
|
452
490
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Page::Layout]
|
|
453
|
-
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for the
|
|
491
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::Layout Layout} for the
|
|
492
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page::FormField FormField} value.
|
|
454
493
|
# @!attribute [rw] name_detected_languages
|
|
455
494
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Page::DetectedLanguage>]
|
|
456
495
|
# A list of detected languages for name together with confidence.
|
|
@@ -477,7 +516,7 @@ module Google
|
|
|
477
516
|
# @return [::String]
|
|
478
517
|
# The BCP-47 language code, such as "en-US" or "sr-Latn". For more
|
|
479
518
|
# information, see
|
|
480
|
-
#
|
|
519
|
+
# https://www.unicode.org/reports/tr35/#Unicode_locale_identifier.
|
|
481
520
|
# @!attribute [rw] confidence
|
|
482
521
|
# @return [::Float]
|
|
483
522
|
# Confidence of detected language. Range [0, 1].
|
|
@@ -487,18 +526,21 @@ module Google
|
|
|
487
526
|
end
|
|
488
527
|
end
|
|
489
528
|
|
|
490
|
-
#
|
|
491
|
-
#
|
|
529
|
+
# An entity that could be a phrase in the text or a property that belongs to
|
|
530
|
+
# the document. It is a known entity type, such as a person, an organization,
|
|
531
|
+
# or location.
|
|
492
532
|
# @!attribute [rw] text_anchor
|
|
493
533
|
# @return [::Google::Cloud::DocumentAI::V1::Document::TextAnchor]
|
|
494
534
|
# Optional. Provenance of the entity.
|
|
495
|
-
# Text anchor indexing into the
|
|
535
|
+
# Text anchor indexing into the
|
|
536
|
+
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
|
496
537
|
# @!attribute [rw] type
|
|
497
538
|
# @return [::String]
|
|
498
539
|
# Entity type from a schema e.g. `Address`.
|
|
499
540
|
# @!attribute [rw] mention_text
|
|
500
541
|
# @return [::String]
|
|
501
|
-
# Optional. Text value in the document e.g. `1600 Amphitheatre Pkwy`.
|
|
542
|
+
# Optional. Text value in the document e.g. `1600 Amphitheatre Pkwy`. If
|
|
543
|
+
# the entity is not present in the document, this field will be empty.
|
|
502
544
|
# @!attribute [rw] mention_id
|
|
503
545
|
# @return [::String]
|
|
504
546
|
# Optional. Deprecated. Use `id` field instead.
|
|
@@ -507,28 +549,29 @@ module Google
|
|
|
507
549
|
# Optional. Confidence of detected Schema entity. Range [0, 1].
|
|
508
550
|
# @!attribute [rw] page_anchor
|
|
509
551
|
# @return [::Google::Cloud::DocumentAI::V1::Document::PageAnchor]
|
|
510
|
-
# Optional. Represents the provenance of this entity wrt. the location on
|
|
511
|
-
# page where it was found.
|
|
552
|
+
# Optional. Represents the provenance of this entity wrt. the location on
|
|
553
|
+
# the page where it was found.
|
|
512
554
|
# @!attribute [rw] id
|
|
513
555
|
# @return [::String]
|
|
514
556
|
# Optional. Canonical id. This will be a unique value in the entity list
|
|
515
557
|
# for this document.
|
|
516
558
|
# @!attribute [rw] normalized_value
|
|
517
559
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Entity::NormalizedValue]
|
|
518
|
-
# Optional. Normalized entity value. Absent if the extracted value could
|
|
519
|
-
# converted or the type (e.g. address) is not supported for certain
|
|
560
|
+
# Optional. Normalized entity value. Absent if the extracted value could
|
|
561
|
+
# not be converted or the type (e.g. address) is not supported for certain
|
|
520
562
|
# parsers. This field is also only populated for certain supported document
|
|
521
563
|
# types.
|
|
522
564
|
# @!attribute [rw] properties
|
|
523
565
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::Entity>]
|
|
524
|
-
# Optional. Entities can be nested to form a hierarchical data structure
|
|
525
|
-
# the content in the document.
|
|
566
|
+
# Optional. Entities can be nested to form a hierarchical data structure
|
|
567
|
+
# representing the content in the document.
|
|
526
568
|
# @!attribute [rw] provenance
|
|
527
569
|
# @return [::Google::Cloud::DocumentAI::V1::Document::Provenance]
|
|
528
570
|
# Optional. The history of this annotation.
|
|
529
571
|
# @!attribute [rw] redacted
|
|
530
572
|
# @return [::Boolean]
|
|
531
|
-
# Optional. Whether the entity will be redacted for de-identification
|
|
573
|
+
# Optional. Whether the entity will be redacted for de-identification
|
|
574
|
+
# purposes.
|
|
532
575
|
class Entity
|
|
533
576
|
include ::Google::Protobuf::MessageExts
|
|
534
577
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
|
@@ -556,10 +599,13 @@ module Google
|
|
|
556
599
|
# checkboxes.
|
|
557
600
|
# @!attribute [rw] text
|
|
558
601
|
# @return [::String]
|
|
559
|
-
#
|
|
560
|
-
#
|
|
561
|
-
#
|
|
602
|
+
# Optional. An optional field to store a normalized string.
|
|
603
|
+
# For some entity types, one of respective `structured_value` fields may
|
|
604
|
+
# also be populated. Also not all the types of `structured_value` will be
|
|
605
|
+
# normalized. For example, some processors may not generate float
|
|
606
|
+
# or int normalized text by default.
|
|
562
607
|
#
|
|
608
|
+
# Below are sample formats mapped to structured values.
|
|
563
609
|
# - Money/Currency type (`money_value`) is in the ISO 4217 text format.
|
|
564
610
|
# - Date type (`date_value`) is in the ISO 8601 text format.
|
|
565
611
|
# - Datetime type (`datetime_value`) is in the ISO 8601 text format.
|
|
@@ -569,7 +615,8 @@ module Google
|
|
|
569
615
|
end
|
|
570
616
|
end
|
|
571
617
|
|
|
572
|
-
# Relationship between
|
|
618
|
+
# Relationship between
|
|
619
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Entity Entities}.
|
|
573
620
|
# @!attribute [rw] subject_id
|
|
574
621
|
# @return [::String]
|
|
575
622
|
# Subject entity id.
|
|
@@ -584,27 +631,35 @@ module Google
|
|
|
584
631
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
|
585
632
|
end
|
|
586
633
|
|
|
587
|
-
# Text reference indexing into the
|
|
634
|
+
# Text reference indexing into the
|
|
635
|
+
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
|
588
636
|
# @!attribute [rw] text_segments
|
|
589
637
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::TextAnchor::TextSegment>]
|
|
590
|
-
# The text segments from the
|
|
638
|
+
# The text segments from the
|
|
639
|
+
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
|
591
640
|
# @!attribute [rw] content
|
|
592
641
|
# @return [::String]
|
|
593
642
|
# Contains the content of the text span so that users do
|
|
594
|
-
# not have to look it up in the text_segments.
|
|
643
|
+
# not have to look it up in the text_segments. It is always
|
|
644
|
+
# populated for formFields.
|
|
595
645
|
class TextAnchor
|
|
596
646
|
include ::Google::Protobuf::MessageExts
|
|
597
647
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
|
598
648
|
|
|
599
|
-
# A text segment in the
|
|
600
|
-
#
|
|
601
|
-
#
|
|
649
|
+
# A text segment in the
|
|
650
|
+
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}. The indices
|
|
651
|
+
# may be out of bounds which indicate that the text extends into another
|
|
652
|
+
# document shard for large sharded documents. See
|
|
653
|
+
# {::Google::Cloud::DocumentAI::V1::Document::ShardInfo#text_offset ShardInfo.text_offset}
|
|
602
654
|
# @!attribute [rw] start_index
|
|
603
655
|
# @return [::Integer]
|
|
604
|
-
# {::Google::Cloud::DocumentAI::V1::Document::TextAnchor::TextSegment TextSegment}
|
|
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}.
|
|
605
659
|
# @!attribute [rw] end_index
|
|
606
660
|
# @return [::Integer]
|
|
607
|
-
# {::Google::Cloud::DocumentAI::V1::Document::TextAnchor::TextSegment TextSegment}
|
|
661
|
+
# {::Google::Cloud::DocumentAI::V1::Document::TextAnchor::TextSegment TextSegment}
|
|
662
|
+
# half open end UTF-8 char index in the
|
|
608
663
|
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}.
|
|
609
664
|
class TextSegment
|
|
610
665
|
include ::Google::Protobuf::MessageExts
|
|
@@ -612,9 +667,10 @@ module Google
|
|
|
612
667
|
end
|
|
613
668
|
end
|
|
614
669
|
|
|
615
|
-
# Referencing the visual context of the entity in the
|
|
616
|
-
# Page anchors
|
|
617
|
-
#
|
|
670
|
+
# Referencing the visual context of the entity in the
|
|
671
|
+
# {::Google::Cloud::DocumentAI::V1::Document#pages Document.pages}. Page anchors
|
|
672
|
+
# can be cross-page, consist of multiple bounding polygons and optionally
|
|
673
|
+
# reference specific layout element types.
|
|
618
674
|
# @!attribute [rw] page_refs
|
|
619
675
|
# @return [::Array<::Google::Cloud::DocumentAI::V1::Document::PageAnchor::PageRef>]
|
|
620
676
|
# One or more references to visual page elements
|
|
@@ -625,22 +681,29 @@ module Google
|
|
|
625
681
|
# Represents a weak reference to a page element within a document.
|
|
626
682
|
# @!attribute [rw] page
|
|
627
683
|
# @return [::Integer]
|
|
628
|
-
# Required. Index into the
|
|
629
|
-
#
|
|
630
|
-
#
|
|
684
|
+
# Required. Index into the
|
|
685
|
+
# {::Google::Cloud::DocumentAI::V1::Document#pages Document.pages} element,
|
|
686
|
+
# for example using [Document.pages][page_refs.page] to locate the
|
|
687
|
+
# related page element. This field is skipped when its value is the
|
|
688
|
+
# default 0. See
|
|
631
689
|
# https://developers.google.com/protocol-buffers/docs/proto3#json.
|
|
632
690
|
# @!attribute [rw] layout_type
|
|
633
691
|
# @return [::Google::Cloud::DocumentAI::V1::Document::PageAnchor::PageRef::LayoutType]
|
|
634
|
-
# Optional. The type of the layout element that is being referenced if
|
|
692
|
+
# Optional. The type of the layout element that is being referenced if
|
|
693
|
+
# any.
|
|
635
694
|
# @!attribute [rw] layout_id
|
|
636
695
|
# @return [::String]
|
|
637
|
-
# Optional. Deprecated. Use
|
|
696
|
+
# Optional. Deprecated. Use
|
|
697
|
+
# {::Google::Cloud::DocumentAI::V1::Document::PageAnchor::PageRef#bounding_poly PageRef.bounding_poly}
|
|
698
|
+
# instead.
|
|
638
699
|
# @!attribute [rw] bounding_poly
|
|
639
700
|
# @return [::Google::Cloud::DocumentAI::V1::BoundingPoly]
|
|
640
|
-
# Optional. Identifies the bounding polygon of a layout element on the
|
|
701
|
+
# Optional. Identifies the bounding polygon of a layout element on the
|
|
702
|
+
# page.
|
|
641
703
|
# @!attribute [rw] confidence
|
|
642
704
|
# @return [::Float]
|
|
643
|
-
# Optional. Confidence of detected page element, if applicable. Range [0,
|
|
705
|
+
# Optional. Confidence of detected page element, if applicable. Range [0,
|
|
706
|
+
# 1].
|
|
644
707
|
class PageRef
|
|
645
708
|
include ::Google::Protobuf::MessageExts
|
|
646
709
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
|
@@ -650,25 +713,38 @@ module Google
|
|
|
650
713
|
# Layout Unspecified.
|
|
651
714
|
LAYOUT_TYPE_UNSPECIFIED = 0
|
|
652
715
|
|
|
653
|
-
# References a
|
|
716
|
+
# References a
|
|
717
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page#blocks Page.blocks}
|
|
718
|
+
# element.
|
|
654
719
|
BLOCK = 1
|
|
655
720
|
|
|
656
|
-
# References a
|
|
721
|
+
# References a
|
|
722
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page#paragraphs Page.paragraphs}
|
|
723
|
+
# element.
|
|
657
724
|
PARAGRAPH = 2
|
|
658
725
|
|
|
659
|
-
# References a
|
|
726
|
+
# References a
|
|
727
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page#lines Page.lines} element.
|
|
660
728
|
LINE = 3
|
|
661
729
|
|
|
662
|
-
# References a
|
|
730
|
+
# References a
|
|
731
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page#tokens Page.tokens}
|
|
732
|
+
# element.
|
|
663
733
|
TOKEN = 4
|
|
664
734
|
|
|
665
|
-
# References a
|
|
735
|
+
# References a
|
|
736
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page#visual_elements Page.visual_elements}
|
|
737
|
+
# element.
|
|
666
738
|
VISUAL_ELEMENT = 5
|
|
667
739
|
|
|
668
|
-
# Refrrences a
|
|
740
|
+
# Refrrences a
|
|
741
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page#tables Page.tables}
|
|
742
|
+
# element.
|
|
669
743
|
TABLE = 6
|
|
670
744
|
|
|
671
|
-
# References a
|
|
745
|
+
# References a
|
|
746
|
+
# {::Google::Cloud::DocumentAI::V1::Document::Page#form_fields Page.form_fields}
|
|
747
|
+
# element.
|
|
672
748
|
FORM_FIELD = 7
|
|
673
749
|
end
|
|
674
750
|
end
|
|
@@ -693,16 +769,15 @@ module Google
|
|
|
693
769
|
include ::Google::Protobuf::MessageExts
|
|
694
770
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
|
695
771
|
|
|
696
|
-
#
|
|
697
|
-
#
|
|
698
|
-
# are replaced.
|
|
772
|
+
# The parent element the current element is based on. Used for
|
|
773
|
+
# referencing/aligning, removal and replacement operations.
|
|
699
774
|
# @!attribute [rw] revision
|
|
700
775
|
# @return [::Integer]
|
|
701
|
-
# The index of the
|
|
776
|
+
# The index of the index into current revision's parent_ids list.
|
|
702
777
|
# @!attribute [rw] index
|
|
703
778
|
# @return [::Integer]
|
|
704
|
-
# The index of the parent
|
|
705
|
-
#
|
|
779
|
+
# The index of the parent item in the corresponding item list (eg. list
|
|
780
|
+
# of entities, properties within entities, etc.) in the parent revision.
|
|
706
781
|
# @!attribute [rw] id
|
|
707
782
|
# @return [::Integer]
|
|
708
783
|
# The id of the parent provenance.
|
|
@@ -713,19 +788,20 @@ module Google
|
|
|
713
788
|
|
|
714
789
|
# If a processor or agent does an explicit operation on existing elements.
|
|
715
790
|
module OperationType
|
|
716
|
-
# Operation type unspecified.
|
|
791
|
+
# Operation type unspecified. If no operation is specified a provenance
|
|
792
|
+
# entry is simply used to match against a `parent`.
|
|
717
793
|
OPERATION_TYPE_UNSPECIFIED = 0
|
|
718
794
|
|
|
719
|
-
# Add an element.
|
|
795
|
+
# Add an element.
|
|
720
796
|
ADD = 1
|
|
721
797
|
|
|
722
|
-
#
|
|
798
|
+
# Remove an element identified by `parent`.
|
|
723
799
|
REMOVE = 2
|
|
724
800
|
|
|
725
|
-
#
|
|
801
|
+
# Replace an element identified by `parent`.
|
|
726
802
|
REPLACE = 3
|
|
727
803
|
|
|
728
|
-
#
|
|
804
|
+
# Request human review for the element identified by `parent`.
|
|
729
805
|
EVAL_REQUESTED = 4
|
|
730
806
|
|
|
731
807
|
# Element is reviewed and approved at human review, confidence will be
|
|
@@ -782,8 +858,9 @@ module Google
|
|
|
782
858
|
# @!attribute [rw] text_anchor
|
|
783
859
|
# @return [::Google::Cloud::DocumentAI::V1::Document::TextAnchor]
|
|
784
860
|
# Provenance of the correction.
|
|
785
|
-
# Text anchor indexing into the
|
|
786
|
-
#
|
|
861
|
+
# Text anchor indexing into the
|
|
862
|
+
# {::Google::Cloud::DocumentAI::V1::Document#text Document.text}. There can
|
|
863
|
+
# only be a single `TextAnchor.text_segments` element. If the start and
|
|
787
864
|
# end index of the text segment are the same, the text change is inserted
|
|
788
865
|
# before that index.
|
|
789
866
|
# @!attribute [rw] changed_text
|
|
@@ -227,6 +227,9 @@ module Google
|
|
|
227
227
|
# @!attribute [rw] common_metadata
|
|
228
228
|
# @return [::Google::Cloud::DocumentAI::V1::CommonOperationMetadata]
|
|
229
229
|
# The basic metadata of the long running operation.
|
|
230
|
+
# @!attribute [rw] question_id
|
|
231
|
+
# @return [::String]
|
|
232
|
+
# The question ID.
|
|
230
233
|
class ReviewDocumentOperationMetadata
|
|
231
234
|
include ::Google::Protobuf::MessageExts
|
|
232
235
|
extend ::Google::Protobuf::MessageExts::ClassMethods
|
metadata
CHANGED
|
@@ -1,14 +1,14 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: google-cloud-document_ai-v1
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 0.
|
|
4
|
+
version: 0.3.0
|
|
5
5
|
platform: ruby
|
|
6
6
|
authors:
|
|
7
7
|
- Google LLC
|
|
8
8
|
autorequire:
|
|
9
9
|
bindir: bin
|
|
10
10
|
cert_chain: []
|
|
11
|
-
date: 2022-
|
|
11
|
+
date: 2022-02-16 00:00:00.000000000 Z
|
|
12
12
|
dependencies:
|
|
13
13
|
- !ruby/object:Gem::Dependency
|
|
14
14
|
name: gapic-common
|
|
@@ -224,7 +224,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
|
|
|
224
224
|
- !ruby/object:Gem::Version
|
|
225
225
|
version: '0'
|
|
226
226
|
requirements: []
|
|
227
|
-
rubygems_version: 3.3.
|
|
227
|
+
rubygems_version: 3.3.5
|
|
228
228
|
signing_key:
|
|
229
229
|
specification_version: 4
|
|
230
230
|
summary: API Client library for the Document AI V1 API
|