From 925119966b5c2de7bb21e931a29928c61006ce19 Mon Sep 17 00:00:00 2001 From: yoshi-automation Date: Wed, 9 Oct 2019 05:39:49 -0700 Subject: [PATCH] [CHANGE ME] Re-generated videointelligence to pick up changes in the API or client library generator. --- .../video_intelligence_service_client.py | 6 +- ...ideo_intelligence_service_client_config.py | 2 +- .../proto/video_intelligence.proto | 245 +++++++------- .../proto/video_intelligence_pb2.py | 316 ++++++++++-------- videointelligence/synth.metadata | 6 +- 5 files changed, 306 insertions(+), 269 deletions(-) diff --git a/videointelligence/google/cloud/videointelligence_v1/gapic/video_intelligence_service_client.py b/videointelligence/google/cloud/videointelligence_v1/gapic/video_intelligence_service_client.py index a9058fb38c33..f872f349fde8 100644 --- a/videointelligence/google/cloud/videointelligence_v1/gapic/video_intelligence_service_client.py +++ b/videointelligence/google/cloud/videointelligence_v1/gapic/video_intelligence_service_client.py @@ -241,19 +241,19 @@ def annotate_video( request as ``input_content``. If set, ``input_content`` should be unset. input_content (bytes): The video data bytes. If unset, the input video(s) should be specified via ``input_uri``. If set, ``input_uri`` should be unset. - features (list[~google.cloud.videointelligence_v1.types.Feature]): Requested video annotation features. + features (list[~google.cloud.videointelligence_v1.types.Feature]): Required. Requested video annotation features. video_context (Union[dict, ~google.cloud.videointelligence_v1.types.VideoContext]): Additional video context and/or feature-specific parameters. If a dict is provided, it must be of the same form as the protobuf message :class:`~google.cloud.videointelligence_v1.types.VideoContext` - output_uri (str): Optional location where the output (in JSON format) should be stored. + output_uri (str): Optional. Location where the output (in JSON format) should be stored. Currently, only `Google Cloud Storage `__ URIs are supported, which must be specified in the following format: ``gs://bucket-id/object-id`` (other URI formats return ``google.rpc.Code.INVALID_ARGUMENT``). For more information, see `Request URIs `__. - location_id (str): Optional cloud region where annotation should take place. Supported + location_id (str): Optional. Cloud region where annotation should take place. Supported cloud regions: ``us-east1``, ``us-west1``, ``europe-west1``, ``asia-east1``. If no region is specified, a region will be determined based on video file location. diff --git a/videointelligence/google/cloud/videointelligence_v1/gapic/video_intelligence_service_client_config.py b/videointelligence/google/cloud/videointelligence_v1/gapic/video_intelligence_service_client_config.py index b9eef5e2a6a7..74dc2121caff 100644 --- a/videointelligence/google/cloud/videointelligence_v1/gapic/video_intelligence_service_client_config.py +++ b/videointelligence/google/cloud/videointelligence_v1/gapic/video_intelligence_service_client_config.py @@ -18,7 +18,7 @@ }, "methods": { "AnnotateVideo": { - "timeout_millis": 600000, + "timeout_millis": 60000, "retry_codes_name": "idempotent", "retry_params_name": "default", } diff --git a/videointelligence/google/cloud/videointelligence_v1/proto/video_intelligence.proto b/videointelligence/google/cloud/videointelligence_v1/proto/video_intelligence.proto index 4c7a0ad131a2..6611e9eb2c61 100644 --- a/videointelligence/google/cloud/videointelligence_v1/proto/video_intelligence.proto +++ b/videointelligence/google/cloud/videointelligence_v1/proto/video_intelligence.proto @@ -19,6 +19,7 @@ package google.cloud.videointelligence.v1; import "google/api/annotations.proto"; import "google/api/client.proto"; +import "google/api/field_behavior.proto"; import "google/longrunning/operations.proto"; import "google/protobuf/duration.proto"; import "google/protobuf/timestamp.proto"; @@ -35,19 +36,22 @@ option ruby_package = "Google::Cloud::VideoIntelligence::V1"; // Service that implements Google Cloud Video Intelligence API. service VideoIntelligenceService { option (google.api.default_host) = "videointelligence.googleapis.com"; - option (google.api.oauth_scopes) = - "https://www.googleapis.com/auth/cloud-platform"; + option (google.api.oauth_scopes) = "https://www.googleapis.com/auth/cloud-platform"; // Performs asynchronous video annotation. Progress and results can be // retrieved through the `google.longrunning.Operations` interface. // `Operation.metadata` contains `AnnotateVideoProgress` (progress). // `Operation.response` contains `AnnotateVideoResponse` (results). - rpc AnnotateVideo(AnnotateVideoRequest) - returns (google.longrunning.Operation) { + rpc AnnotateVideo(AnnotateVideoRequest) returns (google.longrunning.Operation) { option (google.api.http) = { post: "/v1/videos:annotate" body: "*" }; + option (google.api.method_signature) = "input_uri,features"; + option (google.longrunning.operation_info) = { + response_type: "AnnotateVideoResponse" + metadata_type: "AnnotateVideoProgress" + }; } } @@ -57,10 +61,10 @@ message AnnotateVideoRequest { // [Google Cloud Storage](https://cloud.google.com/storage/) URIs are // supported, which must be specified in the following format: // `gs://bucket-id/object-id` (other URI formats return - // [google.rpc.Code.INVALID_ARGUMENT][google.rpc.Code.INVALID_ARGUMENT]). For - // more information, see [Request URIs](/storage/docs/reference-uris). A video - // URI may include wildcards in `object-id`, and thus identify multiple - // videos. Supported wildcards: '*' to match 0 or more characters; + // [google.rpc.Code.INVALID_ARGUMENT][google.rpc.Code.INVALID_ARGUMENT]). For more information, see + // [Request URIs](/storage/docs/reference-uris). + // A video URI may include wildcards in `object-id`, and thus identify + // multiple videos. Supported wildcards: '*' to match 0 or more characters; // '?' to match 1 character. If unset, the input video should be embedded // in the request as `input_content`. If set, `input_content` should be unset. string input_uri = 1; @@ -70,24 +74,24 @@ message AnnotateVideoRequest { // If set, `input_uri` should be unset. bytes input_content = 6; - // Requested video annotation features. - repeated Feature features = 2; + // Required. Requested video annotation features. + repeated Feature features = 2 [(google.api.field_behavior) = REQUIRED]; // Additional video context and/or feature-specific parameters. VideoContext video_context = 3; - // Optional location where the output (in JSON format) should be stored. + // Optional. Location where the output (in JSON format) should be stored. // Currently, only [Google Cloud Storage](https://cloud.google.com/storage/) // URIs are supported, which must be specified in the following format: // `gs://bucket-id/object-id` (other URI formats return - // [google.rpc.Code.INVALID_ARGUMENT][google.rpc.Code.INVALID_ARGUMENT]). For - // more information, see [Request URIs](/storage/docs/reference-uris). - string output_uri = 4; + // [google.rpc.Code.INVALID_ARGUMENT][google.rpc.Code.INVALID_ARGUMENT]). For more information, see + // [Request URIs](/storage/docs/reference-uris). + string output_uri = 4 [(google.api.field_behavior) = OPTIONAL]; - // Optional cloud region where annotation should take place. Supported cloud + // Optional. Cloud region where annotation should take place. Supported cloud // regions: `us-east1`, `us-west1`, `europe-west1`, `asia-east1`. If no region // is specified, a region will be determined based on video file location. - string location_id = 5; + string location_id = 5 [(google.api.field_behavior) = OPTIONAL]; } // Video context and/or feature-specific parameters. @@ -119,6 +123,69 @@ message VideoContext { ObjectTrackingConfig object_tracking_config = 13; } +// Video annotation feature. +enum Feature { + // Unspecified. + FEATURE_UNSPECIFIED = 0; + + // Label detection. Detect objects, such as dog or flower. + LABEL_DETECTION = 1; + + // Shot change detection. + SHOT_CHANGE_DETECTION = 2; + + // Explicit content detection. + EXPLICIT_CONTENT_DETECTION = 3; + + // Human face detection and tracking. + FACE_DETECTION = 4; + + // Speech transcription. + SPEECH_TRANSCRIPTION = 6; + + // OCR text detection and tracking. + TEXT_DETECTION = 7; + + // Object detection and tracking. + OBJECT_TRACKING = 9; +} + +// Label detection mode. +enum LabelDetectionMode { + // Unspecified. + LABEL_DETECTION_MODE_UNSPECIFIED = 0; + + // Detect shot-level labels. + SHOT_MODE = 1; + + // Detect frame-level labels. + FRAME_MODE = 2; + + // Detect both shot-level and frame-level labels. + SHOT_AND_FRAME_MODE = 3; +} + +// Bucketized representation of likelihood. +enum Likelihood { + // Unspecified likelihood. + LIKELIHOOD_UNSPECIFIED = 0; + + // Very unlikely. + VERY_UNLIKELY = 1; + + // Unlikely. + UNLIKELY = 2; + + // Possible. + POSSIBLE = 3; + + // Likely. + LIKELY = 4; + + // Very likely. + VERY_LIKELY = 5; +} + // Config for LABEL_DETECTION. message LabelDetectionConfig { // What labels should be detected with LABEL_DETECTION, in addition to @@ -161,9 +228,9 @@ message ShotChangeDetectionConfig { string model = 1; } -// Config for EXPLICIT_CONTENT_DETECTION. -message ExplicitContentDetectionConfig { - // Model to use for explicit content detection. +// Config for OBJECT_TRACKING. +message ObjectTrackingConfig { + // Model to use for object tracking. // Supported values: "builtin/stable" (the default if unset) and // "builtin/latest". string model = 1; @@ -180,9 +247,9 @@ message FaceDetectionConfig { bool include_bounding_boxes = 2; } -// Config for OBJECT_TRACKING. -message ObjectTrackingConfig { - // Model to use for object tracking. +// Config for EXPLICIT_CONTENT_DETECTION. +message ExplicitContentDetectionConfig { + // Model to use for explicit content detection. // Supported values: "builtin/stable" (the default if unset) and // "builtin/latest". string model = 1; @@ -344,8 +411,11 @@ message VideoAnnotationResults { repeated LabelAnnotation segment_label_annotations = 2; // Presence label annotations on video level or user specified segment level. - // There is exactly one element for each unique label. This will eventually - // get publicly exposed and the restriction will be removed. + // There is exactly one element for each unique label. Compared to the + // existing topical `segment_label_annotations`, this field presents more + // fine-grained, segment-level labels detected in video content and is made + // available only when the client sets `LabelDetectionConfig.model` to + // "builtin/latest" in the request. repeated LabelAnnotation segment_presence_label_annotations = 23; // Topical label annotations on shot level. @@ -353,8 +423,10 @@ message VideoAnnotationResults { repeated LabelAnnotation shot_label_annotations = 3; // Presence label annotations on shot level. There is exactly one element for - // each unique label. This will eventually get publicly exposed and the - // restriction will be removed. + // each unique label. Compared to the existing topical + // `shot_label_annotations`, this field presents more fine-grained, shot-level + // labels detected in video content and is made available only when the client + // sets `LabelDetectionConfig.model` to "builtin/latest" in the request. repeated LabelAnnotation shot_presence_label_annotations = 24; // Label annotations on frame level. @@ -429,72 +501,71 @@ message AnnotateVideoProgress { // Config for SPEECH_TRANSCRIPTION. message SpeechTranscriptionConfig { - // *Required* The language of the supplied audio as a + // Required. *Required* The language of the supplied audio as a // [BCP-47](https://www.rfc-editor.org/rfc/bcp/bcp47.txt) language tag. // Example: "en-US". // See [Language Support](https://cloud.google.com/speech/docs/languages) // for a list of the currently supported language codes. - string language_code = 1; + string language_code = 1 [(google.api.field_behavior) = REQUIRED]; - // *Optional* Maximum number of recognition hypotheses to be returned. + // Optional. Maximum number of recognition hypotheses to be returned. // Specifically, the maximum number of `SpeechRecognitionAlternative` messages // within each `SpeechTranscription`. The server may return fewer than // `max_alternatives`. Valid values are `0`-`30`. A value of `0` or `1` will // return a maximum of one. If omitted, will return a maximum of one. - int32 max_alternatives = 2; + int32 max_alternatives = 2 [(google.api.field_behavior) = OPTIONAL]; - // *Optional* If set to `true`, the server will attempt to filter out + // Optional. If set to `true`, the server will attempt to filter out // profanities, replacing all but the initial character in each filtered word // with asterisks, e.g. "f***". If set to `false` or omitted, profanities // won't be filtered out. - bool filter_profanity = 3; + bool filter_profanity = 3 [(google.api.field_behavior) = OPTIONAL]; - // *Optional* A means to provide context to assist the speech recognition. - repeated SpeechContext speech_contexts = 4; + // Optional. A means to provide context to assist the speech recognition. + repeated SpeechContext speech_contexts = 4 [(google.api.field_behavior) = OPTIONAL]; - // *Optional* If 'true', adds punctuation to recognition result hypotheses. + // Optional. If 'true', adds punctuation to recognition result hypotheses. // This feature is only available in select languages. Setting this for // requests in other languages has no effect at all. The default 'false' value // does not add punctuation to result hypotheses. NOTE: "This is currently // offered as an experimental service, complimentary to all users. In the // future this may be exclusively available as a premium feature." - bool enable_automatic_punctuation = 5; + bool enable_automatic_punctuation = 5 [(google.api.field_behavior) = OPTIONAL]; - // *Optional* For file formats, such as MXF or MKV, supporting multiple audio + // Optional. For file formats, such as MXF or MKV, supporting multiple audio // tracks, specify up to two tracks. Default: track 0. - repeated int32 audio_tracks = 6; + repeated int32 audio_tracks = 6 [(google.api.field_behavior) = OPTIONAL]; - // *Optional* If 'true', enables speaker detection for each recognized word in + // Optional. If 'true', enables speaker detection for each recognized word in // the top alternative of the recognition result using a speaker_tag provided // in the WordInfo. // Note: When this is true, we send all the words from the beginning of the // audio for the top alternative in every consecutive responses. // This is done in order to improve our speaker tags as our models learn to // identify the speakers in the conversation over time. - bool enable_speaker_diarization = 7; + bool enable_speaker_diarization = 7 [(google.api.field_behavior) = OPTIONAL]; - // *Optional* - // If set, specifies the estimated number of speakers in the conversation. + // Optional. If set, specifies the estimated number of speakers in the conversation. // If not set, defaults to '2'. // Ignored unless enable_speaker_diarization is set to true. - int32 diarization_speaker_count = 8; + int32 diarization_speaker_count = 8 [(google.api.field_behavior) = OPTIONAL]; - // *Optional* If `true`, the top result includes a list of words and the + // Optional. If `true`, the top result includes a list of words and the // confidence for those words. If `false`, no word-level confidence // information is returned. The default is `false`. - bool enable_word_confidence = 9; + bool enable_word_confidence = 9 [(google.api.field_behavior) = OPTIONAL]; } // Provides "hints" to the speech recognizer to favor specific words and phrases // in the results. message SpeechContext { - // *Optional* A list of strings containing words and phrases "hints" so that + // Optional. A list of strings containing words and phrases "hints" so that // the speech recognition is more likely to recognize them. This can be used // to improve the accuracy for specific words and phrases, for example, if // specific commands are typically spoken by the user. This can also be used // to add additional words to the vocabulary of the recognizer. See // [usage limits](https://cloud.google.com/speech/limits#content). - repeated string phrases = 1; + repeated string phrases = 1 [(google.api.field_behavior) = OPTIONAL]; } // A speech recognition result corresponding to a portion of the audio. @@ -505,11 +576,10 @@ message SpeechTranscription { // ranked by the recognizer. repeated SpeechRecognitionAlternative alternatives = 1; - // Output only. The - // [BCP-47](https://www.rfc-editor.org/rfc/bcp/bcp47.txt) language tag of the - // language in this result. This language code was detected to have the most - // likelihood of being spoken in the audio. - string language_code = 2; + // Output only. The [BCP-47](https://www.rfc-editor.org/rfc/bcp/bcp47.txt) language tag of + // the language in this result. This language code was detected to have the + // most likelihood of being spoken in the audio. + string language_code = 2 [(google.api.field_behavior) = OUTPUT_ONLY]; } // Alternative hypotheses (a.k.a. n-best list). @@ -523,12 +593,12 @@ message SpeechRecognitionAlternative { // This field is not guaranteed to be accurate and users should not rely on it // to be always provided. // The default of 0.0 is a sentinel value indicating `confidence` was not set. - float confidence = 2; + float confidence = 2 [(google.api.field_behavior) = OUTPUT_ONLY]; // Output only. A list of word-specific information for each recognized word. // Note: When `enable_speaker_diarization` is true, you will see all the words // from the beginning of the audio. - repeated WordInfo words = 3; + repeated WordInfo words = 3 [(google.api.field_behavior) = OUTPUT_ONLY]; } // Word-specific information for recognized words. Word information is only @@ -556,13 +626,13 @@ message WordInfo { // This field is not guaranteed to be accurate and users should not rely on it // to be always provided. // The default of 0.0 is a sentinel value indicating `confidence` was not set. - float confidence = 4; + float confidence = 4 [(google.api.field_behavior) = OUTPUT_ONLY]; // Output only. A distinct integer value is assigned for every speaker within // the audio. This field specifies which one of those speakers was detected to // have spoken this word. Value ranges from 1 up to diarization_speaker_count, // and is only set if speaker diarization is enabled. - int32 speaker_tag = 5; + int32 speaker_tag = 5 [(google.api.field_behavior) = OUTPUT_ONLY]; } // A vertex represents a 2D point in the image. @@ -673,66 +743,3 @@ message ObjectTrackingAnnotation { // Streaming mode: it can only be one ObjectTrackingFrame message in frames. repeated ObjectTrackingFrame frames = 2; } - -// Video annotation feature. -enum Feature { - // Unspecified. - FEATURE_UNSPECIFIED = 0; - - // Label detection. Detect objects, such as dog or flower. - LABEL_DETECTION = 1; - - // Shot change detection. - SHOT_CHANGE_DETECTION = 2; - - // Explicit content detection. - EXPLICIT_CONTENT_DETECTION = 3; - - // Human face detection and tracking. - FACE_DETECTION = 4; - - // Speech transcription. - SPEECH_TRANSCRIPTION = 6; - - // OCR text detection and tracking. - TEXT_DETECTION = 7; - - // Object detection and tracking. - OBJECT_TRACKING = 9; -} - -// Label detection mode. -enum LabelDetectionMode { - // Unspecified. - LABEL_DETECTION_MODE_UNSPECIFIED = 0; - - // Detect shot-level labels. - SHOT_MODE = 1; - - // Detect frame-level labels. - FRAME_MODE = 2; - - // Detect both shot-level and frame-level labels. - SHOT_AND_FRAME_MODE = 3; -} - -// Bucketized representation of likelihood. -enum Likelihood { - // Unspecified likelihood. - LIKELIHOOD_UNSPECIFIED = 0; - - // Very unlikely. - VERY_UNLIKELY = 1; - - // Unlikely. - UNLIKELY = 2; - - // Possible. - POSSIBLE = 3; - - // Likely. - LIKELY = 4; - - // Very likely. - VERY_LIKELY = 5; -} diff --git a/videointelligence/google/cloud/videointelligence_v1/proto/video_intelligence_pb2.py b/videointelligence/google/cloud/videointelligence_v1/proto/video_intelligence_pb2.py index 755f13437776..869eeae436ce 100644 --- a/videointelligence/google/cloud/videointelligence_v1/proto/video_intelligence_pb2.py +++ b/videointelligence/google/cloud/videointelligence_v1/proto/video_intelligence_pb2.py @@ -18,6 +18,7 @@ from google.api import annotations_pb2 as google_dot_api_dot_annotations__pb2 from google.api import client_pb2 as google_dot_api_dot_client__pb2 +from google.api import field_behavior_pb2 as google_dot_api_dot_field__behavior__pb2 from google.longrunning import ( operations_pb2 as google_dot_longrunning_dot_operations__pb2, ) @@ -34,11 +35,12 @@ "\n%com.google.cloud.videointelligence.v1B\035VideoIntelligenceServiceProtoP\001ZRgoogle.golang.org/genproto/googleapis/cloud/videointelligence/v1;videointelligence\252\002!Google.Cloud.VideoIntelligence.V1\312\002!Google\\Cloud\\VideoIntelligence\\V1\352\002$Google::Cloud::VideoIntelligence::V1" ), serialized_pb=_b( - '\n@google/cloud/videointelligence_v1/proto/video_intelligence.proto\x12!google.cloud.videointelligence.v1\x1a\x1cgoogle/api/annotations.proto\x1a\x17google/api/client.proto\x1a#google/longrunning/operations.proto\x1a\x1egoogle/protobuf/duration.proto\x1a\x1fgoogle/protobuf/timestamp.proto\x1a\x17google/rpc/status.proto"\xef\x01\n\x14\x41nnotateVideoRequest\x12\x11\n\tinput_uri\x18\x01 \x01(\t\x12\x15\n\rinput_content\x18\x06 \x01(\x0c\x12<\n\x08\x66\x65\x61tures\x18\x02 \x03(\x0e\x32*.google.cloud.videointelligence.v1.Feature\x12\x46\n\rvideo_context\x18\x03 \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoContext\x12\x12\n\noutput_uri\x18\x04 \x01(\t\x12\x13\n\x0blocation_id\x18\x05 \x01(\t"\xe6\x05\n\x0cVideoContext\x12\x41\n\x08segments\x18\x01 \x03(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment\x12W\n\x16label_detection_config\x18\x02 \x01(\x0b\x32\x37.google.cloud.videointelligence.v1.LabelDetectionConfig\x12\x62\n\x1cshot_change_detection_config\x18\x03 \x01(\x0b\x32<.google.cloud.videointelligence.v1.ShotChangeDetectionConfig\x12l\n!explicit_content_detection_config\x18\x04 \x01(\x0b\x32\x41.google.cloud.videointelligence.v1.ExplicitContentDetectionConfig\x12U\n\x15\x66\x61\x63\x65_detection_config\x18\x05 \x01(\x0b\x32\x36.google.cloud.videointelligence.v1.FaceDetectionConfig\x12\x61\n\x1bspeech_transcription_config\x18\x06 \x01(\x0b\x32<.google.cloud.videointelligence.v1.SpeechTranscriptionConfig\x12U\n\x15text_detection_config\x18\x08 \x01(\x0b\x32\x36.google.cloud.videointelligence.v1.TextDetectionConfig\x12W\n\x16object_tracking_config\x18\r \x01(\x0b\x32\x37.google.cloud.videointelligence.v1.ObjectTrackingConfig"\xdd\x01\n\x14LabelDetectionConfig\x12S\n\x14label_detection_mode\x18\x01 \x01(\x0e\x32\x35.google.cloud.videointelligence.v1.LabelDetectionMode\x12\x19\n\x11stationary_camera\x18\x02 \x01(\x08\x12\r\n\x05model\x18\x03 \x01(\t\x12"\n\x1a\x66rame_confidence_threshold\x18\x04 \x01(\x02\x12"\n\x1avideo_confidence_threshold\x18\x05 \x01(\x02"*\n\x19ShotChangeDetectionConfig\x12\r\n\x05model\x18\x01 \x01(\t"/\n\x1e\x45xplicitContentDetectionConfig\x12\r\n\x05model\x18\x01 \x01(\t"D\n\x13\x46\x61\x63\x65\x44\x65tectionConfig\x12\r\n\x05model\x18\x01 \x01(\t\x12\x1e\n\x16include_bounding_boxes\x18\x02 \x01(\x08"%\n\x14ObjectTrackingConfig\x12\r\n\x05model\x18\x01 \x01(\t"<\n\x13TextDetectionConfig\x12\x16\n\x0elanguage_hints\x18\x01 \x03(\t\x12\r\n\x05model\x18\x02 \x01(\t"x\n\x0cVideoSegment\x12\x34\n\x11start_time_offset\x18\x01 \x01(\x0b\x32\x19.google.protobuf.Duration\x12\x32\n\x0f\x65nd_time_offset\x18\x02 \x01(\x0b\x32\x19.google.protobuf.Duration"d\n\x0cLabelSegment\x12@\n\x07segment\x18\x01 \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment\x12\x12\n\nconfidence\x18\x02 \x01(\x02"P\n\nLabelFrame\x12.\n\x0btime_offset\x18\x01 \x01(\x0b\x32\x19.google.protobuf.Duration\x12\x12\n\nconfidence\x18\x02 \x01(\x02"G\n\x06\x45ntity\x12\x11\n\tentity_id\x18\x01 \x01(\t\x12\x13\n\x0b\x64\x65scription\x18\x02 \x01(\t\x12\x15\n\rlanguage_code\x18\x03 \x01(\t"\x94\x02\n\x0fLabelAnnotation\x12\x39\n\x06\x65ntity\x18\x01 \x01(\x0b\x32).google.cloud.videointelligence.v1.Entity\x12\x44\n\x11\x63\x61tegory_entities\x18\x02 \x03(\x0b\x32).google.cloud.videointelligence.v1.Entity\x12\x41\n\x08segments\x18\x03 \x03(\x0b\x32/.google.cloud.videointelligence.v1.LabelSegment\x12=\n\x06\x66rames\x18\x04 \x03(\x0b\x32-.google.cloud.videointelligence.v1.LabelFrame"\x95\x01\n\x14\x45xplicitContentFrame\x12.\n\x0btime_offset\x18\x01 \x01(\x0b\x32\x19.google.protobuf.Duration\x12M\n\x16pornography_likelihood\x18\x02 \x01(\x0e\x32-.google.cloud.videointelligence.v1.Likelihood"d\n\x19\x45xplicitContentAnnotation\x12G\n\x06\x66rames\x18\x01 \x03(\x0b\x32\x37.google.cloud.videointelligence.v1.ExplicitContentFrame"Q\n\x15NormalizedBoundingBox\x12\x0c\n\x04left\x18\x01 \x01(\x02\x12\x0b\n\x03top\x18\x02 \x01(\x02\x12\r\n\x05right\x18\x03 \x01(\x02\x12\x0e\n\x06\x62ottom\x18\x04 \x01(\x02"O\n\x0b\x46\x61\x63\x65Segment\x12@\n\x07segment\x18\x01 \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment"\x98\x01\n\tFaceFrame\x12[\n\x19normalized_bounding_boxes\x18\x01 \x03(\x0b\x32\x38.google.cloud.videointelligence.v1.NormalizedBoundingBox\x12.\n\x0btime_offset\x18\x02 \x01(\x0b\x32\x19.google.protobuf.Duration"\xa3\x01\n\x0e\x46\x61\x63\x65\x41nnotation\x12\x11\n\tthumbnail\x18\x01 \x01(\x0c\x12@\n\x08segments\x18\x02 \x03(\x0b\x32..google.cloud.videointelligence.v1.FaceSegment\x12<\n\x06\x66rames\x18\x03 \x03(\x0b\x32,.google.cloud.videointelligence.v1.FaceFrame"\xbd\x08\n\x16VideoAnnotationResults\x12\x11\n\tinput_uri\x18\x01 \x01(\t\x12@\n\x07segment\x18\n \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment\x12U\n\x19segment_label_annotations\x18\x02 \x03(\x0b\x32\x32.google.cloud.videointelligence.v1.LabelAnnotation\x12^\n"segment_presence_label_annotations\x18\x17 \x03(\x0b\x32\x32.google.cloud.videointelligence.v1.LabelAnnotation\x12R\n\x16shot_label_annotations\x18\x03 \x03(\x0b\x32\x32.google.cloud.videointelligence.v1.LabelAnnotation\x12[\n\x1fshot_presence_label_annotations\x18\x18 \x03(\x0b\x32\x32.google.cloud.videointelligence.v1.LabelAnnotation\x12S\n\x17\x66rame_label_annotations\x18\x04 \x03(\x0b\x32\x32.google.cloud.videointelligence.v1.LabelAnnotation\x12K\n\x10\x66\x61\x63\x65_annotations\x18\x05 \x03(\x0b\x32\x31.google.cloud.videointelligence.v1.FaceAnnotation\x12I\n\x10shot_annotations\x18\x06 \x03(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment\x12Y\n\x13\x65xplicit_annotation\x18\x07 \x01(\x0b\x32<.google.cloud.videointelligence.v1.ExplicitContentAnnotation\x12U\n\x15speech_transcriptions\x18\x0b \x03(\x0b\x32\x36.google.cloud.videointelligence.v1.SpeechTranscription\x12K\n\x10text_annotations\x18\x0c \x03(\x0b\x32\x31.google.cloud.videointelligence.v1.TextAnnotation\x12W\n\x12object_annotations\x18\x0e \x03(\x0b\x32;.google.cloud.videointelligence.v1.ObjectTrackingAnnotation\x12!\n\x05\x65rror\x18\t \x01(\x0b\x32\x12.google.rpc.Status"n\n\x15\x41nnotateVideoResponse\x12U\n\x12\x61nnotation_results\x18\x01 \x03(\x0b\x32\x39.google.cloud.videointelligence.v1.VideoAnnotationResults"\xa6\x02\n\x17VideoAnnotationProgress\x12\x11\n\tinput_uri\x18\x01 \x01(\t\x12\x18\n\x10progress_percent\x18\x02 \x01(\x05\x12.\n\nstart_time\x18\x03 \x01(\x0b\x32\x1a.google.protobuf.Timestamp\x12/\n\x0bupdate_time\x18\x04 \x01(\x0b\x32\x1a.google.protobuf.Timestamp\x12;\n\x07\x66\x65\x61ture\x18\x05 \x01(\x0e\x32*.google.cloud.videointelligence.v1.Feature\x12@\n\x07segment\x18\x06 \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment"p\n\x15\x41nnotateVideoProgress\x12W\n\x13\x61nnotation_progress\x18\x01 \x03(\x0b\x32:.google.cloud.videointelligence.v1.VideoAnnotationProgress"\xd4\x02\n\x19SpeechTranscriptionConfig\x12\x15\n\rlanguage_code\x18\x01 \x01(\t\x12\x18\n\x10max_alternatives\x18\x02 \x01(\x05\x12\x18\n\x10\x66ilter_profanity\x18\x03 \x01(\x08\x12I\n\x0fspeech_contexts\x18\x04 \x03(\x0b\x32\x30.google.cloud.videointelligence.v1.SpeechContext\x12$\n\x1c\x65nable_automatic_punctuation\x18\x05 \x01(\x08\x12\x14\n\x0c\x61udio_tracks\x18\x06 \x03(\x05\x12"\n\x1a\x65nable_speaker_diarization\x18\x07 \x01(\x08\x12!\n\x19\x64iarization_speaker_count\x18\x08 \x01(\x05\x12\x1e\n\x16\x65nable_word_confidence\x18\t \x01(\x08" \n\rSpeechContext\x12\x0f\n\x07phrases\x18\x01 \x03(\t"\x83\x01\n\x13SpeechTranscription\x12U\n\x0c\x61lternatives\x18\x01 \x03(\x0b\x32?.google.cloud.videointelligence.v1.SpeechRecognitionAlternative\x12\x15\n\rlanguage_code\x18\x02 \x01(\t"\x82\x01\n\x1cSpeechRecognitionAlternative\x12\x12\n\ntranscript\x18\x01 \x01(\t\x12\x12\n\nconfidence\x18\x02 \x01(\x02\x12:\n\x05words\x18\x03 \x03(\x0b\x32+.google.cloud.videointelligence.v1.WordInfo"\x9d\x01\n\x08WordInfo\x12-\n\nstart_time\x18\x01 \x01(\x0b\x32\x19.google.protobuf.Duration\x12+\n\x08\x65nd_time\x18\x02 \x01(\x0b\x32\x19.google.protobuf.Duration\x12\x0c\n\x04word\x18\x03 \x01(\t\x12\x12\n\nconfidence\x18\x04 \x01(\x02\x12\x13\n\x0bspeaker_tag\x18\x05 \x01(\x05"(\n\x10NormalizedVertex\x12\t\n\x01x\x18\x01 \x01(\x02\x12\t\n\x01y\x18\x02 \x01(\x02"_\n\x16NormalizedBoundingPoly\x12\x45\n\x08vertices\x18\x01 \x03(\x0b\x32\x33.google.cloud.videointelligence.v1.NormalizedVertex"\xa1\x01\n\x0bTextSegment\x12@\n\x07segment\x18\x01 \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment\x12\x12\n\nconfidence\x18\x02 \x01(\x02\x12<\n\x06\x66rames\x18\x03 \x03(\x0b\x32,.google.cloud.videointelligence.v1.TextFrame"\x94\x01\n\tTextFrame\x12W\n\x14rotated_bounding_box\x18\x01 \x01(\x0b\x32\x39.google.cloud.videointelligence.v1.NormalizedBoundingPoly\x12.\n\x0btime_offset\x18\x02 \x01(\x0b\x32\x19.google.protobuf.Duration"`\n\x0eTextAnnotation\x12\x0c\n\x04text\x18\x01 \x01(\t\x12@\n\x08segments\x18\x02 \x03(\x0b\x32..google.cloud.videointelligence.v1.TextSegment"\xa0\x01\n\x13ObjectTrackingFrame\x12Y\n\x17normalized_bounding_box\x18\x01 \x01(\x0b\x32\x38.google.cloud.videointelligence.v1.NormalizedBoundingBox\x12.\n\x0btime_offset\x18\x02 \x01(\x0b\x32\x19.google.protobuf.Duration"\x97\x02\n\x18ObjectTrackingAnnotation\x12\x42\n\x07segment\x18\x03 \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegmentH\x00\x12\x12\n\x08track_id\x18\x05 \x01(\x03H\x00\x12\x39\n\x06\x65ntity\x18\x01 \x01(\x0b\x32).google.cloud.videointelligence.v1.Entity\x12\x12\n\nconfidence\x18\x04 \x01(\x02\x12\x46\n\x06\x66rames\x18\x02 \x03(\x0b\x32\x36.google.cloud.videointelligence.v1.ObjectTrackingFrameB\x0c\n\ntrack_info*\xc9\x01\n\x07\x46\x65\x61ture\x12\x17\n\x13\x46\x45\x41TURE_UNSPECIFIED\x10\x00\x12\x13\n\x0fLABEL_DETECTION\x10\x01\x12\x19\n\x15SHOT_CHANGE_DETECTION\x10\x02\x12\x1e\n\x1a\x45XPLICIT_CONTENT_DETECTION\x10\x03\x12\x12\n\x0e\x46\x41\x43\x45_DETECTION\x10\x04\x12\x18\n\x14SPEECH_TRANSCRIPTION\x10\x06\x12\x12\n\x0eTEXT_DETECTION\x10\x07\x12\x13\n\x0fOBJECT_TRACKING\x10\t*r\n\x12LabelDetectionMode\x12$\n LABEL_DETECTION_MODE_UNSPECIFIED\x10\x00\x12\r\n\tSHOT_MODE\x10\x01\x12\x0e\n\nFRAME_MODE\x10\x02\x12\x17\n\x13SHOT_AND_FRAME_MODE\x10\x03*t\n\nLikelihood\x12\x1a\n\x16LIKELIHOOD_UNSPECIFIED\x10\x00\x12\x11\n\rVERY_UNLIKELY\x10\x01\x12\x0c\n\x08UNLIKELY\x10\x02\x12\x0c\n\x08POSSIBLE\x10\x03\x12\n\n\x06LIKELY\x10\x04\x12\x0f\n\x0bVERY_LIKELY\x10\x05\x32\xfa\x01\n\x18VideoIntelligenceService\x12\x87\x01\n\rAnnotateVideo\x12\x37.google.cloud.videointelligence.v1.AnnotateVideoRequest\x1a\x1d.google.longrunning.Operation"\x1e\x82\xd3\xe4\x93\x02\x18"\x13/v1/videos:annotate:\x01*\x1aT\xca\x41 videointelligence.googleapis.com\xd2\x41.https://www.googleapis.com/auth/cloud-platformB\x8b\x02\n%com.google.cloud.videointelligence.v1B\x1dVideoIntelligenceServiceProtoP\x01ZRgoogle.golang.org/genproto/googleapis/cloud/videointelligence/v1;videointelligence\xaa\x02!Google.Cloud.VideoIntelligence.V1\xca\x02!Google\\Cloud\\VideoIntelligence\\V1\xea\x02$Google::Cloud::VideoIntelligence::V1b\x06proto3' + '\n@google/cloud/videointelligence_v1/proto/video_intelligence.proto\x12!google.cloud.videointelligence.v1\x1a\x1cgoogle/api/annotations.proto\x1a\x17google/api/client.proto\x1a\x1fgoogle/api/field_behavior.proto\x1a#google/longrunning/operations.proto\x1a\x1egoogle/protobuf/duration.proto\x1a\x1fgoogle/protobuf/timestamp.proto\x1a\x17google/rpc/status.proto"\xfe\x01\n\x14\x41nnotateVideoRequest\x12\x11\n\tinput_uri\x18\x01 \x01(\t\x12\x15\n\rinput_content\x18\x06 \x01(\x0c\x12\x41\n\x08\x66\x65\x61tures\x18\x02 \x03(\x0e\x32*.google.cloud.videointelligence.v1.FeatureB\x03\xe0\x41\x02\x12\x46\n\rvideo_context\x18\x03 \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoContext\x12\x17\n\noutput_uri\x18\x04 \x01(\tB\x03\xe0\x41\x01\x12\x18\n\x0blocation_id\x18\x05 \x01(\tB\x03\xe0\x41\x01"\xe6\x05\n\x0cVideoContext\x12\x41\n\x08segments\x18\x01 \x03(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment\x12W\n\x16label_detection_config\x18\x02 \x01(\x0b\x32\x37.google.cloud.videointelligence.v1.LabelDetectionConfig\x12\x62\n\x1cshot_change_detection_config\x18\x03 \x01(\x0b\x32<.google.cloud.videointelligence.v1.ShotChangeDetectionConfig\x12l\n!explicit_content_detection_config\x18\x04 \x01(\x0b\x32\x41.google.cloud.videointelligence.v1.ExplicitContentDetectionConfig\x12U\n\x15\x66\x61\x63\x65_detection_config\x18\x05 \x01(\x0b\x32\x36.google.cloud.videointelligence.v1.FaceDetectionConfig\x12\x61\n\x1bspeech_transcription_config\x18\x06 \x01(\x0b\x32<.google.cloud.videointelligence.v1.SpeechTranscriptionConfig\x12U\n\x15text_detection_config\x18\x08 \x01(\x0b\x32\x36.google.cloud.videointelligence.v1.TextDetectionConfig\x12W\n\x16object_tracking_config\x18\r \x01(\x0b\x32\x37.google.cloud.videointelligence.v1.ObjectTrackingConfig"\xdd\x01\n\x14LabelDetectionConfig\x12S\n\x14label_detection_mode\x18\x01 \x01(\x0e\x32\x35.google.cloud.videointelligence.v1.LabelDetectionMode\x12\x19\n\x11stationary_camera\x18\x02 \x01(\x08\x12\r\n\x05model\x18\x03 \x01(\t\x12"\n\x1a\x66rame_confidence_threshold\x18\x04 \x01(\x02\x12"\n\x1avideo_confidence_threshold\x18\x05 \x01(\x02"*\n\x19ShotChangeDetectionConfig\x12\r\n\x05model\x18\x01 \x01(\t"%\n\x14ObjectTrackingConfig\x12\r\n\x05model\x18\x01 \x01(\t"D\n\x13\x46\x61\x63\x65\x44\x65tectionConfig\x12\r\n\x05model\x18\x01 \x01(\t\x12\x1e\n\x16include_bounding_boxes\x18\x02 \x01(\x08"/\n\x1e\x45xplicitContentDetectionConfig\x12\r\n\x05model\x18\x01 \x01(\t"<\n\x13TextDetectionConfig\x12\x16\n\x0elanguage_hints\x18\x01 \x03(\t\x12\r\n\x05model\x18\x02 \x01(\t"x\n\x0cVideoSegment\x12\x34\n\x11start_time_offset\x18\x01 \x01(\x0b\x32\x19.google.protobuf.Duration\x12\x32\n\x0f\x65nd_time_offset\x18\x02 \x01(\x0b\x32\x19.google.protobuf.Duration"d\n\x0cLabelSegment\x12@\n\x07segment\x18\x01 \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment\x12\x12\n\nconfidence\x18\x02 \x01(\x02"P\n\nLabelFrame\x12.\n\x0btime_offset\x18\x01 \x01(\x0b\x32\x19.google.protobuf.Duration\x12\x12\n\nconfidence\x18\x02 \x01(\x02"G\n\x06\x45ntity\x12\x11\n\tentity_id\x18\x01 \x01(\t\x12\x13\n\x0b\x64\x65scription\x18\x02 \x01(\t\x12\x15\n\rlanguage_code\x18\x03 \x01(\t"\x94\x02\n\x0fLabelAnnotation\x12\x39\n\x06\x65ntity\x18\x01 \x01(\x0b\x32).google.cloud.videointelligence.v1.Entity\x12\x44\n\x11\x63\x61tegory_entities\x18\x02 \x03(\x0b\x32).google.cloud.videointelligence.v1.Entity\x12\x41\n\x08segments\x18\x03 \x03(\x0b\x32/.google.cloud.videointelligence.v1.LabelSegment\x12=\n\x06\x66rames\x18\x04 \x03(\x0b\x32-.google.cloud.videointelligence.v1.LabelFrame"\x95\x01\n\x14\x45xplicitContentFrame\x12.\n\x0btime_offset\x18\x01 \x01(\x0b\x32\x19.google.protobuf.Duration\x12M\n\x16pornography_likelihood\x18\x02 \x01(\x0e\x32-.google.cloud.videointelligence.v1.Likelihood"d\n\x19\x45xplicitContentAnnotation\x12G\n\x06\x66rames\x18\x01 \x03(\x0b\x32\x37.google.cloud.videointelligence.v1.ExplicitContentFrame"Q\n\x15NormalizedBoundingBox\x12\x0c\n\x04left\x18\x01 \x01(\x02\x12\x0b\n\x03top\x18\x02 \x01(\x02\x12\r\n\x05right\x18\x03 \x01(\x02\x12\x0e\n\x06\x62ottom\x18\x04 \x01(\x02"O\n\x0b\x46\x61\x63\x65Segment\x12@\n\x07segment\x18\x01 \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment"\x98\x01\n\tFaceFrame\x12[\n\x19normalized_bounding_boxes\x18\x01 \x03(\x0b\x32\x38.google.cloud.videointelligence.v1.NormalizedBoundingBox\x12.\n\x0btime_offset\x18\x02 \x01(\x0b\x32\x19.google.protobuf.Duration"\xa3\x01\n\x0e\x46\x61\x63\x65\x41nnotation\x12\x11\n\tthumbnail\x18\x01 \x01(\x0c\x12@\n\x08segments\x18\x02 \x03(\x0b\x32..google.cloud.videointelligence.v1.FaceSegment\x12<\n\x06\x66rames\x18\x03 \x03(\x0b\x32,.google.cloud.videointelligence.v1.FaceFrame"\xbd\x08\n\x16VideoAnnotationResults\x12\x11\n\tinput_uri\x18\x01 \x01(\t\x12@\n\x07segment\x18\n \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment\x12U\n\x19segment_label_annotations\x18\x02 \x03(\x0b\x32\x32.google.cloud.videointelligence.v1.LabelAnnotation\x12^\n"segment_presence_label_annotations\x18\x17 \x03(\x0b\x32\x32.google.cloud.videointelligence.v1.LabelAnnotation\x12R\n\x16shot_label_annotations\x18\x03 \x03(\x0b\x32\x32.google.cloud.videointelligence.v1.LabelAnnotation\x12[\n\x1fshot_presence_label_annotations\x18\x18 \x03(\x0b\x32\x32.google.cloud.videointelligence.v1.LabelAnnotation\x12S\n\x17\x66rame_label_annotations\x18\x04 \x03(\x0b\x32\x32.google.cloud.videointelligence.v1.LabelAnnotation\x12K\n\x10\x66\x61\x63\x65_annotations\x18\x05 \x03(\x0b\x32\x31.google.cloud.videointelligence.v1.FaceAnnotation\x12I\n\x10shot_annotations\x18\x06 \x03(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment\x12Y\n\x13\x65xplicit_annotation\x18\x07 \x01(\x0b\x32<.google.cloud.videointelligence.v1.ExplicitContentAnnotation\x12U\n\x15speech_transcriptions\x18\x0b \x03(\x0b\x32\x36.google.cloud.videointelligence.v1.SpeechTranscription\x12K\n\x10text_annotations\x18\x0c \x03(\x0b\x32\x31.google.cloud.videointelligence.v1.TextAnnotation\x12W\n\x12object_annotations\x18\x0e \x03(\x0b\x32;.google.cloud.videointelligence.v1.ObjectTrackingAnnotation\x12!\n\x05\x65rror\x18\t \x01(\x0b\x32\x12.google.rpc.Status"n\n\x15\x41nnotateVideoResponse\x12U\n\x12\x61nnotation_results\x18\x01 \x03(\x0b\x32\x39.google.cloud.videointelligence.v1.VideoAnnotationResults"\xa6\x02\n\x17VideoAnnotationProgress\x12\x11\n\tinput_uri\x18\x01 \x01(\t\x12\x18\n\x10progress_percent\x18\x02 \x01(\x05\x12.\n\nstart_time\x18\x03 \x01(\x0b\x32\x1a.google.protobuf.Timestamp\x12/\n\x0bupdate_time\x18\x04 \x01(\x0b\x32\x1a.google.protobuf.Timestamp\x12;\n\x07\x66\x65\x61ture\x18\x05 \x01(\x0e\x32*.google.cloud.videointelligence.v1.Feature\x12@\n\x07segment\x18\x06 \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment"p\n\x15\x41nnotateVideoProgress\x12W\n\x13\x61nnotation_progress\x18\x01 \x03(\x0b\x32:.google.cloud.videointelligence.v1.VideoAnnotationProgress"\x81\x03\n\x19SpeechTranscriptionConfig\x12\x1a\n\rlanguage_code\x18\x01 \x01(\tB\x03\xe0\x41\x02\x12\x1d\n\x10max_alternatives\x18\x02 \x01(\x05\x42\x03\xe0\x41\x01\x12\x1d\n\x10\x66ilter_profanity\x18\x03 \x01(\x08\x42\x03\xe0\x41\x01\x12N\n\x0fspeech_contexts\x18\x04 \x03(\x0b\x32\x30.google.cloud.videointelligence.v1.SpeechContextB\x03\xe0\x41\x01\x12)\n\x1c\x65nable_automatic_punctuation\x18\x05 \x01(\x08\x42\x03\xe0\x41\x01\x12\x19\n\x0c\x61udio_tracks\x18\x06 \x03(\x05\x42\x03\xe0\x41\x01\x12\'\n\x1a\x65nable_speaker_diarization\x18\x07 \x01(\x08\x42\x03\xe0\x41\x01\x12&\n\x19\x64iarization_speaker_count\x18\x08 \x01(\x05\x42\x03\xe0\x41\x01\x12#\n\x16\x65nable_word_confidence\x18\t \x01(\x08\x42\x03\xe0\x41\x01"%\n\rSpeechContext\x12\x14\n\x07phrases\x18\x01 \x03(\tB\x03\xe0\x41\x01"\x88\x01\n\x13SpeechTranscription\x12U\n\x0c\x61lternatives\x18\x01 \x03(\x0b\x32?.google.cloud.videointelligence.v1.SpeechRecognitionAlternative\x12\x1a\n\rlanguage_code\x18\x02 \x01(\tB\x03\xe0\x41\x03"\x8c\x01\n\x1cSpeechRecognitionAlternative\x12\x12\n\ntranscript\x18\x01 \x01(\t\x12\x17\n\nconfidence\x18\x02 \x01(\x02\x42\x03\xe0\x41\x03\x12?\n\x05words\x18\x03 \x03(\x0b\x32+.google.cloud.videointelligence.v1.WordInfoB\x03\xe0\x41\x03"\xa7\x01\n\x08WordInfo\x12-\n\nstart_time\x18\x01 \x01(\x0b\x32\x19.google.protobuf.Duration\x12+\n\x08\x65nd_time\x18\x02 \x01(\x0b\x32\x19.google.protobuf.Duration\x12\x0c\n\x04word\x18\x03 \x01(\t\x12\x17\n\nconfidence\x18\x04 \x01(\x02\x42\x03\xe0\x41\x03\x12\x18\n\x0bspeaker_tag\x18\x05 \x01(\x05\x42\x03\xe0\x41\x03"(\n\x10NormalizedVertex\x12\t\n\x01x\x18\x01 \x01(\x02\x12\t\n\x01y\x18\x02 \x01(\x02"_\n\x16NormalizedBoundingPoly\x12\x45\n\x08vertices\x18\x01 \x03(\x0b\x32\x33.google.cloud.videointelligence.v1.NormalizedVertex"\xa1\x01\n\x0bTextSegment\x12@\n\x07segment\x18\x01 \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegment\x12\x12\n\nconfidence\x18\x02 \x01(\x02\x12<\n\x06\x66rames\x18\x03 \x03(\x0b\x32,.google.cloud.videointelligence.v1.TextFrame"\x94\x01\n\tTextFrame\x12W\n\x14rotated_bounding_box\x18\x01 \x01(\x0b\x32\x39.google.cloud.videointelligence.v1.NormalizedBoundingPoly\x12.\n\x0btime_offset\x18\x02 \x01(\x0b\x32\x19.google.protobuf.Duration"`\n\x0eTextAnnotation\x12\x0c\n\x04text\x18\x01 \x01(\t\x12@\n\x08segments\x18\x02 \x03(\x0b\x32..google.cloud.videointelligence.v1.TextSegment"\xa0\x01\n\x13ObjectTrackingFrame\x12Y\n\x17normalized_bounding_box\x18\x01 \x01(\x0b\x32\x38.google.cloud.videointelligence.v1.NormalizedBoundingBox\x12.\n\x0btime_offset\x18\x02 \x01(\x0b\x32\x19.google.protobuf.Duration"\x97\x02\n\x18ObjectTrackingAnnotation\x12\x42\n\x07segment\x18\x03 \x01(\x0b\x32/.google.cloud.videointelligence.v1.VideoSegmentH\x00\x12\x12\n\x08track_id\x18\x05 \x01(\x03H\x00\x12\x39\n\x06\x65ntity\x18\x01 \x01(\x0b\x32).google.cloud.videointelligence.v1.Entity\x12\x12\n\nconfidence\x18\x04 \x01(\x02\x12\x46\n\x06\x66rames\x18\x02 \x03(\x0b\x32\x36.google.cloud.videointelligence.v1.ObjectTrackingFrameB\x0c\n\ntrack_info*\xc9\x01\n\x07\x46\x65\x61ture\x12\x17\n\x13\x46\x45\x41TURE_UNSPECIFIED\x10\x00\x12\x13\n\x0fLABEL_DETECTION\x10\x01\x12\x19\n\x15SHOT_CHANGE_DETECTION\x10\x02\x12\x1e\n\x1a\x45XPLICIT_CONTENT_DETECTION\x10\x03\x12\x12\n\x0e\x46\x41\x43\x45_DETECTION\x10\x04\x12\x18\n\x14SPEECH_TRANSCRIPTION\x10\x06\x12\x12\n\x0eTEXT_DETECTION\x10\x07\x12\x13\n\x0fOBJECT_TRACKING\x10\t*r\n\x12LabelDetectionMode\x12$\n LABEL_DETECTION_MODE_UNSPECIFIED\x10\x00\x12\r\n\tSHOT_MODE\x10\x01\x12\x0e\n\nFRAME_MODE\x10\x02\x12\x17\n\x13SHOT_AND_FRAME_MODE\x10\x03*t\n\nLikelihood\x12\x1a\n\x16LIKELIHOOD_UNSPECIFIED\x10\x00\x12\x11\n\rVERY_UNLIKELY\x10\x01\x12\x0c\n\x08UNLIKELY\x10\x02\x12\x0c\n\x08POSSIBLE\x10\x03\x12\n\n\x06LIKELY\x10\x04\x12\x0f\n\x0bVERY_LIKELY\x10\x05\x32\xc0\x02\n\x18VideoIntelligenceService\x12\xcd\x01\n\rAnnotateVideo\x12\x37.google.cloud.videointelligence.v1.AnnotateVideoRequest\x1a\x1d.google.longrunning.Operation"d\x82\xd3\xe4\x93\x02\x18"\x13/v1/videos:annotate:\x01*\xda\x41\x12input_uri,features\xca\x41.\n\x15\x41nnotateVideoResponse\x12\x15\x41nnotateVideoProgress\x1aT\xca\x41 videointelligence.googleapis.com\xd2\x41.https://www.googleapis.com/auth/cloud-platformB\x8b\x02\n%com.google.cloud.videointelligence.v1B\x1dVideoIntelligenceServiceProtoP\x01ZRgoogle.golang.org/genproto/googleapis/cloud/videointelligence/v1;videointelligence\xaa\x02!Google.Cloud.VideoIntelligence.V1\xca\x02!Google\\Cloud\\VideoIntelligence\\V1\xea\x02$Google::Cloud::VideoIntelligence::V1b\x06proto3' ), dependencies=[ google_dot_api_dot_annotations__pb2.DESCRIPTOR, google_dot_api_dot_client__pb2.DESCRIPTOR, + google_dot_api_dot_field__behavior__pb2.DESCRIPTOR, google_dot_longrunning_dot_operations__pb2.DESCRIPTOR, google_dot_protobuf_dot_duration__pb2.DESCRIPTOR, google_dot_protobuf_dot_timestamp__pb2.DESCRIPTOR, @@ -103,8 +105,8 @@ ], containing_type=None, serialized_options=None, - serialized_start=6570, - serialized_end=6771, + serialized_start=6693, + serialized_end=6894, ) _sym_db.RegisterEnumDescriptor(_FEATURE) @@ -138,8 +140,8 @@ ], containing_type=None, serialized_options=None, - serialized_start=6773, - serialized_end=6887, + serialized_start=6896, + serialized_end=7010, ) _sym_db.RegisterEnumDescriptor(_LABELDETECTIONMODE) @@ -175,8 +177,8 @@ ], containing_type=None, serialized_options=None, - serialized_start=6889, - serialized_end=7005, + serialized_start=7012, + serialized_end=7128, ) _sym_db.RegisterEnumDescriptor(_LIKELIHOOD) @@ -259,7 +261,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\002"), file=DESCRIPTOR, ), _descriptor.FieldDescriptor( @@ -295,7 +297,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\001"), file=DESCRIPTOR, ), _descriptor.FieldDescriptor( @@ -313,7 +315,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\001"), file=DESCRIPTOR, ), ], @@ -325,8 +327,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=286, - serialized_end=525, + serialized_start=319, + serialized_end=573, ) @@ -490,8 +492,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=528, - serialized_end=1270, + serialized_start=576, + serialized_end=1318, ) @@ -601,8 +603,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=1273, - serialized_end=1494, + serialized_start=1321, + serialized_end=1542, ) @@ -640,21 +642,21 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=1496, - serialized_end=1538, + serialized_start=1544, + serialized_end=1586, ) -_EXPLICITCONTENTDETECTIONCONFIG = _descriptor.Descriptor( - name="ExplicitContentDetectionConfig", - full_name="google.cloud.videointelligence.v1.ExplicitContentDetectionConfig", +_OBJECTTRACKINGCONFIG = _descriptor.Descriptor( + name="ObjectTrackingConfig", + full_name="google.cloud.videointelligence.v1.ObjectTrackingConfig", filename=None, file=DESCRIPTOR, containing_type=None, fields=[ _descriptor.FieldDescriptor( name="model", - full_name="google.cloud.videointelligence.v1.ExplicitContentDetectionConfig.model", + full_name="google.cloud.videointelligence.v1.ObjectTrackingConfig.model", index=0, number=1, type=9, @@ -679,8 +681,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=1540, - serialized_end=1587, + serialized_start=1588, + serialized_end=1625, ) @@ -736,21 +738,21 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=1589, - serialized_end=1657, + serialized_start=1627, + serialized_end=1695, ) -_OBJECTTRACKINGCONFIG = _descriptor.Descriptor( - name="ObjectTrackingConfig", - full_name="google.cloud.videointelligence.v1.ObjectTrackingConfig", +_EXPLICITCONTENTDETECTIONCONFIG = _descriptor.Descriptor( + name="ExplicitContentDetectionConfig", + full_name="google.cloud.videointelligence.v1.ExplicitContentDetectionConfig", filename=None, file=DESCRIPTOR, containing_type=None, fields=[ _descriptor.FieldDescriptor( name="model", - full_name="google.cloud.videointelligence.v1.ObjectTrackingConfig.model", + full_name="google.cloud.videointelligence.v1.ExplicitContentDetectionConfig.model", index=0, number=1, type=9, @@ -775,8 +777,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=1659, - serialized_end=1696, + serialized_start=1697, + serialized_end=1744, ) @@ -832,8 +834,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=1698, - serialized_end=1758, + serialized_start=1746, + serialized_end=1806, ) @@ -889,8 +891,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=1760, - serialized_end=1880, + serialized_start=1808, + serialized_end=1928, ) @@ -946,8 +948,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=1882, - serialized_end=1982, + serialized_start=1930, + serialized_end=2030, ) @@ -1003,8 +1005,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=1984, - serialized_end=2064, + serialized_start=2032, + serialized_end=2112, ) @@ -1078,8 +1080,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=2066, - serialized_end=2137, + serialized_start=2114, + serialized_end=2185, ) @@ -1171,8 +1173,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=2140, - serialized_end=2416, + serialized_start=2188, + serialized_end=2464, ) @@ -1228,8 +1230,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=2419, - serialized_end=2568, + serialized_start=2467, + serialized_end=2616, ) @@ -1267,8 +1269,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=2570, - serialized_end=2670, + serialized_start=2618, + serialized_end=2718, ) @@ -1360,8 +1362,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=2672, - serialized_end=2753, + serialized_start=2720, + serialized_end=2801, ) @@ -1399,8 +1401,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=2755, - serialized_end=2834, + serialized_start=2803, + serialized_end=2882, ) @@ -1456,8 +1458,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=2837, - serialized_end=2989, + serialized_start=2885, + serialized_end=3037, ) @@ -1531,8 +1533,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=2992, - serialized_end=3155, + serialized_start=3040, + serialized_end=3203, ) @@ -1804,8 +1806,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=3158, - serialized_end=4243, + serialized_start=3206, + serialized_end=4291, ) @@ -1843,8 +1845,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=4245, - serialized_end=4355, + serialized_start=4293, + serialized_end=4403, ) @@ -1972,8 +1974,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=4358, - serialized_end=4652, + serialized_start=4406, + serialized_end=4700, ) @@ -2011,8 +2013,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=4654, - serialized_end=4766, + serialized_start=4702, + serialized_end=4814, ) @@ -2038,7 +2040,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\002"), file=DESCRIPTOR, ), _descriptor.FieldDescriptor( @@ -2056,7 +2058,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\001"), file=DESCRIPTOR, ), _descriptor.FieldDescriptor( @@ -2074,7 +2076,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\001"), file=DESCRIPTOR, ), _descriptor.FieldDescriptor( @@ -2092,7 +2094,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\001"), file=DESCRIPTOR, ), _descriptor.FieldDescriptor( @@ -2110,7 +2112,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\001"), file=DESCRIPTOR, ), _descriptor.FieldDescriptor( @@ -2128,7 +2130,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\001"), file=DESCRIPTOR, ), _descriptor.FieldDescriptor( @@ -2146,7 +2148,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\001"), file=DESCRIPTOR, ), _descriptor.FieldDescriptor( @@ -2164,7 +2166,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\001"), file=DESCRIPTOR, ), _descriptor.FieldDescriptor( @@ -2182,7 +2184,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\001"), file=DESCRIPTOR, ), ], @@ -2194,8 +2196,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=4769, - serialized_end=5109, + serialized_start=4817, + serialized_end=5202, ) @@ -2221,7 +2223,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\001"), file=DESCRIPTOR, ) ], @@ -2233,8 +2235,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=5111, - serialized_end=5143, + serialized_start=5204, + serialized_end=5241, ) @@ -2278,7 +2280,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\003"), file=DESCRIPTOR, ), ], @@ -2290,8 +2292,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=5146, - serialized_end=5277, + serialized_start=5244, + serialized_end=5380, ) @@ -2335,7 +2337,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\003"), file=DESCRIPTOR, ), _descriptor.FieldDescriptor( @@ -2353,7 +2355,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\003"), file=DESCRIPTOR, ), ], @@ -2365,8 +2367,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=5280, - serialized_end=5410, + serialized_start=5383, + serialized_end=5523, ) @@ -2446,7 +2448,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\003"), file=DESCRIPTOR, ), _descriptor.FieldDescriptor( @@ -2464,7 +2466,7 @@ containing_type=None, is_extension=False, extension_scope=None, - serialized_options=None, + serialized_options=_b("\340A\003"), file=DESCRIPTOR, ), ], @@ -2476,8 +2478,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=5413, - serialized_end=5570, + serialized_start=5526, + serialized_end=5693, ) @@ -2533,8 +2535,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=5572, - serialized_end=5612, + serialized_start=5695, + serialized_end=5735, ) @@ -2572,8 +2574,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=5614, - serialized_end=5709, + serialized_start=5737, + serialized_end=5832, ) @@ -2647,8 +2649,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=5712, - serialized_end=5873, + serialized_start=5835, + serialized_end=5996, ) @@ -2704,8 +2706,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=5876, - serialized_end=6024, + serialized_start=5999, + serialized_end=6147, ) @@ -2761,8 +2763,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=6026, - serialized_end=6122, + serialized_start=6149, + serialized_end=6245, ) @@ -2818,8 +2820,8 @@ syntax="proto3", extension_ranges=[], oneofs=[], - serialized_start=6125, - serialized_end=6285, + serialized_start=6248, + serialized_end=6408, ) @@ -2937,8 +2939,8 @@ fields=[], ) ], - serialized_start=6288, - serialized_end=6567, + serialized_start=6411, + serialized_end=6690, ) _ANNOTATEVIDEOREQUEST.fields_by_name["features"].enum_type = _FEATURE @@ -3093,11 +3095,11 @@ DESCRIPTOR.message_types_by_name[ "ShotChangeDetectionConfig" ] = _SHOTCHANGEDETECTIONCONFIG +DESCRIPTOR.message_types_by_name["ObjectTrackingConfig"] = _OBJECTTRACKINGCONFIG +DESCRIPTOR.message_types_by_name["FaceDetectionConfig"] = _FACEDETECTIONCONFIG DESCRIPTOR.message_types_by_name[ "ExplicitContentDetectionConfig" ] = _EXPLICITCONTENTDETECTIONCONFIG -DESCRIPTOR.message_types_by_name["FaceDetectionConfig"] = _FACEDETECTIONCONFIG -DESCRIPTOR.message_types_by_name["ObjectTrackingConfig"] = _OBJECTTRACKINGCONFIG DESCRIPTOR.message_types_by_name["TextDetectionConfig"] = _TEXTDETECTIONCONFIG DESCRIPTOR.message_types_by_name["VideoSegment"] = _VIDEOSEGMENT DESCRIPTOR.message_types_by_name["LabelSegment"] = _LABELSEGMENT @@ -3165,11 +3167,11 @@ specified via ``input_uri``. If set, ``input_uri`` should be unset. features: - Requested video annotation features. + Required. Requested video annotation features. video_context: Additional video context and/or feature-specific parameters. output_uri: - Optional location where the output (in JSON format) should be + Optional. Location where the output (in JSON format) should be stored. Currently, only `Google Cloud Storage `__ URIs are supported, which must be specified in the following format: @@ -3178,7 +3180,7 @@ ]). For more information, see `Request URIs `__. location_id: - Optional cloud region where annotation should take place. + Optional. Cloud region where annotation should take place. Supported cloud regions: ``us-east1``, ``us-west1``, ``europe- west1``, ``asia-east1``. If no region is specified, a region will be determined based on video file location. @@ -3285,24 +3287,24 @@ ) _sym_db.RegisterMessage(ShotChangeDetectionConfig) -ExplicitContentDetectionConfig = _reflection.GeneratedProtocolMessageType( - "ExplicitContentDetectionConfig", +ObjectTrackingConfig = _reflection.GeneratedProtocolMessageType( + "ObjectTrackingConfig", (_message.Message,), dict( - DESCRIPTOR=_EXPLICITCONTENTDETECTIONCONFIG, + DESCRIPTOR=_OBJECTTRACKINGCONFIG, __module__="google.cloud.videointelligence_v1.proto.video_intelligence_pb2", - __doc__="""Config for EXPLICIT\_CONTENT\_DETECTION. + __doc__="""Config for OBJECT\_TRACKING. Attributes: model: - Model to use for explicit content detection. Supported values: + Model to use for object tracking. Supported values: "builtin/stable" (the default if unset) and "builtin/latest". """, - # @@protoc_insertion_point(class_scope:google.cloud.videointelligence.v1.ExplicitContentDetectionConfig) + # @@protoc_insertion_point(class_scope:google.cloud.videointelligence.v1.ObjectTrackingConfig) ), ) -_sym_db.RegisterMessage(ExplicitContentDetectionConfig) +_sym_db.RegisterMessage(ObjectTrackingConfig) FaceDetectionConfig = _reflection.GeneratedProtocolMessageType( "FaceDetectionConfig", @@ -3326,24 +3328,24 @@ ) _sym_db.RegisterMessage(FaceDetectionConfig) -ObjectTrackingConfig = _reflection.GeneratedProtocolMessageType( - "ObjectTrackingConfig", +ExplicitContentDetectionConfig = _reflection.GeneratedProtocolMessageType( + "ExplicitContentDetectionConfig", (_message.Message,), dict( - DESCRIPTOR=_OBJECTTRACKINGCONFIG, + DESCRIPTOR=_EXPLICITCONTENTDETECTIONCONFIG, __module__="google.cloud.videointelligence_v1.proto.video_intelligence_pb2", - __doc__="""Config for OBJECT\_TRACKING. + __doc__="""Config for EXPLICIT\_CONTENT\_DETECTION. Attributes: model: - Model to use for object tracking. Supported values: + Model to use for explicit content detection. Supported values: "builtin/stable" (the default if unset) and "builtin/latest". """, - # @@protoc_insertion_point(class_scope:google.cloud.videointelligence.v1.ObjectTrackingConfig) + # @@protoc_insertion_point(class_scope:google.cloud.videointelligence.v1.ExplicitContentDetectionConfig) ), ) -_sym_db.RegisterMessage(ObjectTrackingConfig) +_sym_db.RegisterMessage(ExplicitContentDetectionConfig) TextDetectionConfig = _reflection.GeneratedProtocolMessageType( "TextDetectionConfig", @@ -3635,15 +3637,23 @@ segment_presence_label_annotations: Presence label annotations on video level or user specified segment level. There is exactly one element for each unique - label. This will eventually get publicly exposed and the - restriction will be removed. + label. Compared to the existing topical + ``segment_label_annotations``, this field presents more fine- + grained, segment-level labels detected in video content and is + made available only when the client sets + ``LabelDetectionConfig.model`` to "builtin/latest" in the + request. shot_label_annotations: Topical label annotations on shot level. There is exactly one element for each unique label. shot_presence_label_annotations: Presence label annotations on shot level. There is exactly one - element for each unique label. This will eventually get - publicly exposed and the restriction will be removed. + element for each unique label. Compared to the existing + topical ``shot_label_annotations``, this field presents more + fine-grained, shot-level labels detected in video content and + is made available only when the client sets + ``LabelDetectionConfig.model`` to "builtin/latest" in the + request. frame_label_annotations: Label annotations on frame level. There is exactly one element for each unique label. @@ -3757,13 +3767,13 @@ Attributes: language_code: - *Required* The language of the supplied audio as a `BCP-47 - `__ language - tag. Example: "en-US". See `Language Support + Required. *Required* The language of the supplied audio as a + `BCP-47 `__ + language tag. Example: "en-US". See `Language Support `__ for a list of the currently supported language codes. max_alternatives: - *Optional* Maximum number of recognition hypotheses to be + Optional. Maximum number of recognition hypotheses to be returned. Specifically, the maximum number of ``SpeechRecognitionAlternative`` messages within each ``SpeechTranscription``. The server may return fewer than @@ -3771,16 +3781,16 @@ of ``0`` or ``1`` will return a maximum of one. If omitted, will return a maximum of one. filter_profanity: - *Optional* If set to ``true``, the server will attempt to + Optional. If set to ``true``, the server will attempt to filter out profanities, replacing all but the initial character in each filtered word with asterisks, e.g. "f\*\*\*". If set to ``false`` or omitted, profanities won't be filtered out. speech_contexts: - *Optional* A means to provide context to assist the speech + Optional. A means to provide context to assist the speech recognition. enable_automatic_punctuation: - *Optional* If 'true', adds punctuation to recognition result + Optional. If 'true', adds punctuation to recognition result hypotheses. This feature is only available in select languages. Setting this for requests in other languages has no effect at all. The default 'false' value does not add @@ -3789,11 +3799,11 @@ users. In the future this may be exclusively available as a premium feature." audio_tracks: - *Optional* For file formats, such as MXF or MKV, supporting + Optional. For file formats, such as MXF or MKV, supporting multiple audio tracks, specify up to two tracks. Default: track 0. enable_speaker_diarization: - *Optional* If 'true', enables speaker detection for each + Optional. If 'true', enables speaker detection for each recognized word in the top alternative of the recognition result using a speaker\_tag provided in the WordInfo. Note: When this is true, we send all the words from the beginning of @@ -3802,13 +3812,13 @@ as our models learn to identify the speakers in the conversation over time. diarization_speaker_count: - *Optional* If set, specifies the estimated number of speakers + Optional. If set, specifies the estimated number of speakers in the conversation. If not set, defaults to '2'. Ignored unless enable\_speaker\_diarization is set to true. enable_word_confidence: - *Optional* If ``true``, the top result includes a list of - words and the confidence for those words. If ``false``, no - word-level confidence information is returned. The default is + Optional. If ``true``, the top result includes a list of words + and the confidence for those words. If ``false``, no word- + level confidence information is returned. The default is ``false``. """, # @@protoc_insertion_point(class_scope:google.cloud.videointelligence.v1.SpeechTranscriptionConfig) @@ -3828,7 +3838,7 @@ Attributes: phrases: - *Optional* A list of strings containing words and phrases + Optional. A list of strings containing words and phrases "hints" so that the speech recognition is more likely to recognize them. This can be used to improve the accuracy for specific words and phrases, for example, if specific commands @@ -4124,6 +4134,26 @@ DESCRIPTOR._options = None +_ANNOTATEVIDEOREQUEST.fields_by_name["features"]._options = None +_ANNOTATEVIDEOREQUEST.fields_by_name["output_uri"]._options = None +_ANNOTATEVIDEOREQUEST.fields_by_name["location_id"]._options = None +_SPEECHTRANSCRIPTIONCONFIG.fields_by_name["language_code"]._options = None +_SPEECHTRANSCRIPTIONCONFIG.fields_by_name["max_alternatives"]._options = None +_SPEECHTRANSCRIPTIONCONFIG.fields_by_name["filter_profanity"]._options = None +_SPEECHTRANSCRIPTIONCONFIG.fields_by_name["speech_contexts"]._options = None +_SPEECHTRANSCRIPTIONCONFIG.fields_by_name[ + "enable_automatic_punctuation" +]._options = None +_SPEECHTRANSCRIPTIONCONFIG.fields_by_name["audio_tracks"]._options = None +_SPEECHTRANSCRIPTIONCONFIG.fields_by_name["enable_speaker_diarization"]._options = None +_SPEECHTRANSCRIPTIONCONFIG.fields_by_name["diarization_speaker_count"]._options = None +_SPEECHTRANSCRIPTIONCONFIG.fields_by_name["enable_word_confidence"]._options = None +_SPEECHCONTEXT.fields_by_name["phrases"]._options = None +_SPEECHTRANSCRIPTION.fields_by_name["language_code"]._options = None +_SPEECHRECOGNITIONALTERNATIVE.fields_by_name["confidence"]._options = None +_SPEECHRECOGNITIONALTERNATIVE.fields_by_name["words"]._options = None +_WORDINFO.fields_by_name["confidence"]._options = None +_WORDINFO.fields_by_name["speaker_tag"]._options = None _VIDEOINTELLIGENCESERVICE = _descriptor.ServiceDescriptor( name="VideoIntelligenceService", @@ -4133,8 +4163,8 @@ serialized_options=_b( "\312A videointelligence.googleapis.com\322A.https://www.googleapis.com/auth/cloud-platform" ), - serialized_start=7008, - serialized_end=7258, + serialized_start=7131, + serialized_end=7451, methods=[ _descriptor.MethodDescriptor( name="AnnotateVideo", @@ -4144,7 +4174,7 @@ input_type=_ANNOTATEVIDEOREQUEST, output_type=google_dot_longrunning_dot_operations__pb2._OPERATION, serialized_options=_b( - '\202\323\344\223\002\030"\023/v1/videos:annotate:\001*' + '\202\323\344\223\002\030"\023/v1/videos:annotate:\001*\332A\022input_uri,features\312A.\n\025AnnotateVideoResponse\022\025AnnotateVideoProgress' ), ) ], diff --git a/videointelligence/synth.metadata b/videointelligence/synth.metadata index 36eb59e39aec..f548f7349615 100644 --- a/videointelligence/synth.metadata +++ b/videointelligence/synth.metadata @@ -1,5 +1,5 @@ { - "updateTime": "2019-10-08T16:21:52.884418Z", + "updateTime": "2019-10-09T12:39:48.814274Z", "sources": [ { "generator": { @@ -12,8 +12,8 @@ "git": { "name": "googleapis", "remote": "https://github.com/googleapis/googleapis.git", - "sha": "8fca1cdaf063d914042a678848331d7d6211ee13", - "internalRef": "273514424" + "sha": "2dec8f98383214ad4fafa7680eb0cc46d6531976", + "internalRef": "273619851" } }, {