Skip to content

Commit 8114f05

Browse files
gcf-owl-bot[bot]gcf-owl-bot[bot]
authored
feat: added sharding_config field in DocumentOutputConfig.GcsOutputConfig in document_io.proto (#433)
* feat: added sharding_config field in DocumentOutputConfig.GcsOutputConfig in document_io.proto feat: added process_options field in ProcessRequest in document_processor_service.proto feat: added sample_document_uris field in ProcessorType in processor_type.proto PiperOrigin-RevId: 495360288 Source-Link: googleapis/googleapis@5f39f46 Source-Link: googleapis/googleapis-gen@8520d57 Copy-Tag: eyJwIjoiLmdpdGh1Yi8uT3dsQm90LnlhbWwiLCJoIjoiODUyMGQ1NzA5M2QzNjBhZmVjMTk0Y2QwMjliNzEzMjRlNTk3ZjYyNiJ9 * 🦉 Updates from OwlBot post-processor See https://github.com/googleapis/repo-automation-bots/blob/main/packages/owl-bot/README.md * feat: added sharding_config field in DocumentOutputConfig.GcsOutputConfig in document_io.proto feat: added process_options field in ProcessRequest in document_processor_service.proto feat: added sample_document_uris field in ProcessorType in processor_type.proto PiperOrigin-RevId: 495363748 Source-Link: googleapis/googleapis@3dce9ff Source-Link: googleapis/googleapis-gen@66f3518 Copy-Tag: eyJwIjoiLmdpdGh1Yi8uT3dsQm90LnlhbWwiLCJoIjoiNjZmMzUxODIxN2YwYzA0YTQ3ZmFlZmViMzMyMmZmMjI2MTA2Yjg1NyJ9 * 🦉 Updates from OwlBot post-processor See https://github.com/googleapis/repo-automation-bots/blob/main/packages/owl-bot/README.md Co-authored-by: Owl Bot <gcf-owl-bot[bot]@users.noreply.github.com>
1 parent f68dee4 commit 8114f05

14 files changed

Lines changed: 97 additions & 25 deletions

‎packages/google-cloud-documentai/google/cloud/documentai_v1/types/document.py‎

Lines changed: 10 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -72,7 +72,7 @@ class Document(proto.Message):
7272
Optional. UTF-8 encoded text in reading order
7373
from the document.
7474
text_styles (MutableSequence[google.cloud.documentai_v1.types.Document.Style]):
75-
Placeholder. Styles for the
75+
Styles for the
7676
[Document.text][google.cloud.documentai.v1.Document.text].
7777
pages (MutableSequence[google.cloud.documentai_v1.types.Document.Page]):
7878
Visual page layout for the
@@ -775,9 +775,9 @@ class FormField(proto.Message):
775775
If the value is non-textual, this field represents the type.
776776
Current valid values are:
777777
778-
- blank (this indicates the field_value is normal text)
779-
- "unfilled_checkbox"
780-
- "filled_checkbox".
778+
- blank (this indicates the ``field_value`` is normal text)
779+
- ``unfilled_checkbox``
780+
- ``filled_checkbox``
781781
corrected_key_text (str):
782782
Created for Labeling UI to export key text. If corrections
783783
were made to the text identified by the
@@ -1117,8 +1117,8 @@ class NormalizedValue(proto.Message):
11171117
For some entity types, one of respective
11181118
``structured_value`` fields may also be populated. Also not
11191119
all the types of ``structured_value`` will be normalized.
1120-
For example, some processors may not generate float or int
1121-
normalized text by default.
1120+
For example, some processors may not generate ``float`` or
1121+
``integer`` normalized text by default.
11221122
11231123
Below are sample formats mapped to structured values.
11241124
@@ -1323,9 +1323,10 @@ class PageRef(proto.Message):
13231323
page (int):
13241324
Required. Index into the
13251325
[Document.pages][google.cloud.documentai.v1.Document.pages]
1326-
element, for example using [Document.pages][page_refs.page]
1327-
to locate the related page element. This field is skipped
1328-
when its value is the default 0. See
1326+
element, for example using
1327+
``[Document.pages][page_refs.page]`` to locate the related
1328+
page element. This field is skipped when its value is the
1329+
default ``0``. See
13291330
https://developers.google.com/protocol-buffers/docs/proto3#json.
13301331
layout_type (google.cloud.documentai_v1.types.Document.PageAnchor.PageRef.LayoutType):
13311332
Optional. The type of the layout element that

‎packages/google-cloud-documentai/google/cloud/documentai_v1/types/document_schema.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -66,7 +66,7 @@ class EntityType(proto.Message):
6666
following naming conventions:
6767
6868
- *use ``snake_casing``*
69-
- name matching is case-insensitive
69+
- name matching is case-sensitive
7070
- Maximum 64 characters.
7171
- Must start with a letter.
7272
- Allowed characters: ASCII letters ``[a-z0-9_-]``. (For

‎packages/google-cloud-documentai/google/cloud/documentai_v1/types/processor_type.py‎

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -50,6 +50,9 @@ class ProcessorType(proto.Message):
5050
access.
5151
launch_stage (google.api.launch_stage_pb2.LaunchStage):
5252
Launch stage of the processor type
53+
sample_document_uris (MutableSequence[str]):
54+
A set of Cloud Storage URIs of sample
55+
documents for this processor.
5356
"""
5457

5558
class LocationInfo(proto.Message):
@@ -92,6 +95,10 @@ class LocationInfo(proto.Message):
9295
number=8,
9396
enum=launch_stage_pb2.LaunchStage,
9497
)
98+
sample_document_uris: MutableSequence[str] = proto.RepeatedField(
99+
proto.STRING,
100+
number=9,
101+
)
95102

96103

97104
__all__ = tuple(sorted(__protobuf__.manifest))

‎packages/google-cloud-documentai/google/cloud/documentai_v1beta3/__init__.py‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -30,6 +30,7 @@
3030
GcsDocument,
3131
GcsDocuments,
3232
GcsPrefix,
33+
OcrConfig,
3334
RawDocument,
3435
)
3536
from .types.document_processor_service import (
@@ -67,6 +68,7 @@
6768
ListProcessorTypesResponse,
6869
ListProcessorVersionsRequest,
6970
ListProcessorVersionsResponse,
71+
ProcessOptions,
7072
ProcessRequest,
7173
ProcessResponse,
7274
ReviewDocumentOperationMetadata,
@@ -138,6 +140,8 @@
138140
"ListProcessorsRequest",
139141
"ListProcessorsResponse",
140142
"NormalizedVertex",
143+
"OcrConfig",
144+
"ProcessOptions",
141145
"ProcessRequest",
142146
"ProcessResponse",
143147
"Processor",

‎packages/google-cloud-documentai/google/cloud/documentai_v1beta3/types/__init__.py‎

Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -21,6 +21,7 @@
2121
GcsDocument,
2222
GcsDocuments,
2323
GcsPrefix,
24+
OcrConfig,
2425
RawDocument,
2526
)
2627
from .document_processor_service import (
@@ -58,6 +59,7 @@
5859
ListProcessorTypesResponse,
5960
ListProcessorVersionsRequest,
6061
ListProcessorVersionsResponse,
62+
ProcessOptions,
6163
ProcessRequest,
6264
ProcessResponse,
6365
ReviewDocumentOperationMetadata,
@@ -88,6 +90,7 @@
8890
"GcsDocument",
8991
"GcsDocuments",
9092
"GcsPrefix",
93+
"OcrConfig",
9194
"RawDocument",
9295
"BatchProcessMetadata",
9396
"BatchProcessRequest",
@@ -123,6 +126,7 @@
123126
"ListProcessorTypesResponse",
124127
"ListProcessorVersionsRequest",
125128
"ListProcessorVersionsResponse",
129+
"ProcessOptions",
126130
"ProcessRequest",
127131
"ProcessResponse",
128132
"ReviewDocumentOperationMetadata",

‎packages/google-cloud-documentai/google/cloud/documentai_v1beta3/types/document.py‎

Lines changed: 10 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -72,7 +72,7 @@ class Document(proto.Message):
7272
Optional. UTF-8 encoded text in reading order
7373
from the document.
7474
text_styles (MutableSequence[google.cloud.documentai_v1beta3.types.Document.Style]):
75-
Placeholder. Styles for the
75+
Styles for the
7676
[Document.text][google.cloud.documentai.v1beta3.Document.text].
7777
pages (MutableSequence[google.cloud.documentai_v1beta3.types.Document.Page]):
7878
Visual page layout for the
@@ -779,9 +779,9 @@ class FormField(proto.Message):
779779
If the value is non-textual, this field represents the type.
780780
Current valid values are:
781781
782-
- blank (this indicates the field_value is normal text)
783-
- "unfilled_checkbox"
784-
- "filled_checkbox".
782+
- blank (this indicates the ``field_value`` is normal text)
783+
- ``unfilled_checkbox``
784+
- ``filled_checkbox``
785785
corrected_key_text (str):
786786
Created for Labeling UI to export key text. If corrections
787787
were made to the text identified by the
@@ -1121,8 +1121,8 @@ class NormalizedValue(proto.Message):
11211121
For some entity types, one of respective
11221122
``structured_value`` fields may also be populated. Also not
11231123
all the types of ``structured_value`` will be normalized.
1124-
For example, some processors may not generate float or int
1125-
normalized text by default.
1124+
For example, some processors may not generate ``float`` or
1125+
``integer`` normalized text by default.
11261126
11271127
Below are sample formats mapped to structured values.
11281128
@@ -1327,9 +1327,10 @@ class PageRef(proto.Message):
13271327
page (int):
13281328
Required. Index into the
13291329
[Document.pages][google.cloud.documentai.v1beta3.Document.pages]
1330-
element, for example using [Document.pages][page_refs.page]
1331-
to locate the related page element. This field is skipped
1332-
when its value is the default 0. See
1330+
element, for example using
1331+
``[Document.pages][page_refs.page]`` to locate the related
1332+
page element. This field is skipped when its value is the
1333+
default ``0``. See
13331334
https://developers.google.com/protocol-buffers/docs/proto3#json.
13341335
layout_type (google.cloud.documentai_v1beta3.types.Document.PageAnchor.PageRef.LayoutType):
13351336
Optional. The type of the layout element that

‎packages/google-cloud-documentai/google/cloud/documentai_v1beta3/types/document_io.py‎

Lines changed: 17 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -27,6 +27,7 @@
2727
"GcsPrefix",
2828
"BatchDocumentsInputConfig",
2929
"DocumentOutputConfig",
30+
"OcrConfig",
3031
},
3132
)
3233

@@ -218,4 +219,20 @@ class ShardingConfig(proto.Message):
218219
)
219220

220221

222+
class OcrConfig(proto.Message):
223+
r"""Config for Document OCR.
224+
225+
Attributes:
226+
enable_native_pdf_parsing (bool):
227+
Enables special handling for PDFs with
228+
existing text information. Results in better
229+
text extraction quality in such PDF inputs.
230+
"""
231+
232+
enable_native_pdf_parsing: bool = proto.Field(
233+
proto.BOOL,
234+
number=3,
235+
)
236+
237+
221238
__all__ = tuple(sorted(__protobuf__.manifest))

‎packages/google-cloud-documentai/google/cloud/documentai_v1beta3/types/document_processor_service.py‎

Lines changed: 31 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -31,6 +31,7 @@
3131
__protobuf__ = proto.module(
3232
package="google.cloud.documentai.v1beta3",
3333
manifest={
34+
"ProcessOptions",
3435
"ProcessRequest",
3536
"HumanReviewStatus",
3637
"ProcessResponse",
@@ -83,6 +84,22 @@
8384
)
8485

8586

87+
class ProcessOptions(proto.Message):
88+
r"""Options for Process API
89+
90+
Attributes:
91+
ocr_config (google.cloud.documentai_v1beta3.types.OcrConfig):
92+
Only applicable to "Document OCR Processor".
93+
Returns error if set on other processor types.
94+
"""
95+
96+
ocr_config: document_io.OcrConfig = proto.Field(
97+
proto.MESSAGE,
98+
number=1,
99+
message=document_io.OcrConfig,
100+
)
101+
102+
86103
class ProcessRequest(proto.Message):
87104
r"""Request message for the process document method.
88105
@@ -125,6 +142,8 @@ class ProcessRequest(proto.Message):
125142
document. Only supports top level document and pages field
126143
so it must be in the form of ``{document_field_name}`` or
127144
``pages.{page_field_name}``.
145+
process_options (google.cloud.documentai_v1beta3.types.ProcessOptions):
146+
Inference-time options for the process API
128147
"""
129148

130149
inline_document: gcd_document.Document = proto.Field(
@@ -157,6 +176,11 @@ class ProcessRequest(proto.Message):
157176
number=6,
158177
message=field_mask_pb2.FieldMask,
159178
)
179+
process_options: "ProcessOptions" = proto.Field(
180+
proto.MESSAGE,
181+
number=7,
182+
message="ProcessOptions",
183+
)
160184

161185

162186
class HumanReviewStatus(proto.Message):
@@ -259,6 +283,8 @@ class BatchProcessRequest(proto.Message):
259283
skip_human_review (bool):
260284
Whether Human Review feature should be
261285
skipped for this request. Default to false.
286+
process_options (google.cloud.documentai_v1beta3.types.ProcessOptions):
287+
Inference-time options for the process API
262288
"""
263289

264290
class BatchInputConfig(proto.Message):
@@ -326,6 +352,11 @@ class BatchOutputConfig(proto.Message):
326352
proto.BOOL,
327353
number=4,
328354
)
355+
process_options: "ProcessOptions" = proto.Field(
356+
proto.MESSAGE,
357+
number=7,
358+
message="ProcessOptions",
359+
)
329360

330361

331362
class BatchProcessResponse(proto.Message):

‎packages/google-cloud-documentai/google/cloud/documentai_v1beta3/types/document_schema.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -66,7 +66,7 @@ class EntityType(proto.Message):
6666
following naming conventions:
6767
6868
- *use ``snake_casing``*
69-
- name matching is case-insensitive
69+
- name matching is case-sensitive
7070
- Maximum 64 characters.
7171
- Must start with a letter.
7272
- Allowed characters: ASCII letters ``[a-z0-9_-]``. (For

‎packages/google-cloud-documentai/google/cloud/documentai_v1beta3/types/processor_type.py‎

Lines changed: 7 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -50,6 +50,9 @@ class ProcessorType(proto.Message):
5050
access.
5151
launch_stage (google.api.launch_stage_pb2.LaunchStage):
5252
Launch stage of the processor type
53+
sample_document_uris (MutableSequence[str]):
54+
A set of Cloud Storage URIs of sample
55+
documents for this processor.
5356
"""
5457

5558
class LocationInfo(proto.Message):
@@ -92,6 +95,10 @@ class LocationInfo(proto.Message):
9295
number=8,
9396
enum=launch_stage_pb2.LaunchStage,
9497
)
98+
sample_document_uris: MutableSequence[str] = proto.RepeatedField(
99+
proto.STRING,
100+
number=9,
101+
)
95102

96103

97104
__all__ = tuple(sorted(__protobuf__.manifest))

0 commit comments

Comments
 (0)