googleapis/google/cloud/language/v2/language_service.proto

// Copyright 2025 Google LLC
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//     http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

syntax = "proto3";

package google.cloud.language.v2;

import "google/api/annotations.proto";
import "google/api/client.proto";
import "google/api/field_behavior.proto";

option go_package = "cloud.google.com/go/language/apiv2/languagepb;languagepb";
option java_multiple_files = true;
option java_outer_classname = "LanguageServiceProto";
option java_package = "com.google.cloud.language.v2";

// Provides text analysis operations such as sentiment analysis and entity
// recognition.
service LanguageService {
  option (google.api.default_host) = "language.googleapis.com";
  option (google.api.oauth_scopes) =
      "https://www.googleapis.com/auth/cloud-language,"
      "https://www.googleapis.com/auth/cloud-platform";

  // Analyzes the sentiment of the provided text.
  rpc AnalyzeSentiment(AnalyzeSentimentRequest)
      returns (AnalyzeSentimentResponse) {
    option (google.api.http) = {
      post: "/v2/documents:analyzeSentiment"
      body: "*"
    };
    option (google.api.method_signature) = "document,encoding_type";
    option (google.api.method_signature) = "document";
  }

  // Finds named entities (currently proper names and common nouns) in the text
  // along with entity types, probability, mentions for each entity, and
  // other properties.
  rpc AnalyzeEntities(AnalyzeEntitiesRequest)
      returns (AnalyzeEntitiesResponse) {
    option (google.api.http) = {
      post: "/v2/documents:analyzeEntities"
      body: "*"
    };
    option (google.api.method_signature) = "document,encoding_type";
    option (google.api.method_signature) = "document";
  }

  // Classifies a document into categories.
  rpc ClassifyText(ClassifyTextRequest) returns (ClassifyTextResponse) {
    option (google.api.http) = {
      post: "/v2/documents:classifyText"
      body: "*"
    };
    option (google.api.method_signature) = "document";
  }

  // Moderates a document for harmful and sensitive categories.
  rpc ModerateText(ModerateTextRequest) returns (ModerateTextResponse) {
    option (google.api.http) = {
      post: "/v2/documents:moderateText"
      body: "*"
    };
    option (google.api.method_signature) = "document";
  }

  // A convenience method that provides all features in one call.
  rpc AnnotateText(AnnotateTextRequest) returns (AnnotateTextResponse) {
    option (google.api.http) = {
      post: "/v2/documents:annotateText"
      body: "*"
    };
    option (google.api.method_signature) = "document,features,encoding_type";
    option (google.api.method_signature) = "document,features";
  }
}

// Represents the input to API methods.
message Document {
  // The document types enum.
  enum Type {
    // The content type is not specified.
    TYPE_UNSPECIFIED = 0;

    // Plain text
    PLAIN_TEXT = 1;

    // HTML
    HTML = 2;
  }

  // Required. If the type is not set or is `TYPE_UNSPECIFIED`,
  // returns an `INVALID_ARGUMENT` error.
  Type type = 1;

  // The source of the document: a string containing the content or a
  // Google Cloud Storage URI.
  oneof source {
    // The content of the input in string format.
    // Cloud audit logging exempt since it is based on user data.
    string content = 2;

    // The Google Cloud Storage URI where the file content is located.
    // This URI must be of the form: gs://bucket_name/object_name. For more
    // details, see https://cloud.google.com/storage/docs/reference-uris.
    // NOTE: Cloud Storage object versioning is not supported.
    string gcs_content_uri = 3;
  }

  // Optional. The language of the document (if not specified, the language is
  // automatically detected). Both ISO and BCP-47 language codes are
  // accepted.<br>
  // [Language
  // Support](https://cloud.google.com/natural-language/docs/languages) lists
  // currently supported languages for each API method. If the language (either
  // specified by the caller or automatically detected) is not supported by the
  // called API method, an `INVALID_ARGUMENT` error is returned.
  string language_code = 4 [(google.api.field_behavior) = OPTIONAL];
}

// Represents a sentence in the input document.
message Sentence {
  // The sentence text.
  TextSpan text = 1;

  // For calls to [AnalyzeSentiment][] or if
  // [AnnotateTextRequest.Features.extract_document_sentiment][google.cloud.language.v2.AnnotateTextRequest.Features.extract_document_sentiment]
  // is set to true, this field will contain the sentiment for the sentence.
  Sentiment sentiment = 2;
}

// Represents the text encoding that the caller uses to process the output.
// Providing an `EncodingType` is recommended because the API provides the
// beginning offsets for various outputs, such as tokens and mentions, and
// languages that natively use different text encodings may access offsets
// differently.
enum EncodingType {
  // If `EncodingType` is not specified, encoding-dependent information (such as
  // `begin_offset`) will be set at `-1`.
  NONE = 0;

  // Encoding-dependent information (such as `begin_offset`) is calculated based
  // on the UTF-8 encoding of the input. C++ and Go are examples of languages
  // that use this encoding natively.
  UTF8 = 1;

  // Encoding-dependent information (such as `begin_offset`) is calculated based
  // on the UTF-16 encoding of the input. Java and JavaScript are examples of
  // languages that use this encoding natively.
  UTF16 = 2;

  // Encoding-dependent information (such as `begin_offset`) is calculated based
  // on the UTF-32 encoding of the input. Python is an example of a language
  // that uses this encoding natively.
  UTF32 = 3;
}

// Represents a phrase in the text that is a known entity, such as
// a person, an organization, or location. The API associates information, such
// as probability and mentions, with entities.
message Entity {
  // The type of the entity. The table
  // below lists the associated fields for entities that have different
  // metadata.
  enum Type {
    // Unknown
    UNKNOWN = 0;

    // Person
    PERSON = 1;

    // Location
    LOCATION = 2;

    // Organization
    ORGANIZATION = 3;

    // Event
    EVENT = 4;

    // Artwork
    WORK_OF_ART = 5;

    // Consumer product
    CONSUMER_GOOD = 6;

    // Other types of entities
    OTHER = 7;

    // Phone number
    //
    // The metadata lists the phone number, formatted according to local
    // convention, plus whichever additional elements appear in the text:
    //
    // * `number` - the actual number, broken down into sections as per local
    // convention
    // * `national_prefix` - country code, if detected
    // * `area_code` - region or area code, if detected
    // * `extension` - phone extension (to be dialed after connection), if
    // detected
    PHONE_NUMBER = 9;

    // Address
    //
    // The metadata identifies the street number and locality plus whichever
    // additional elements appear in the text:
    //
    // * `street_number` - street number
    // * `locality` - city or town
    // * `street_name` - street/route name, if detected
    // * `postal_code` - postal code, if detected
    // * `country` - country, if detected
    // * `broad_region` - administrative area, such as the state, if detected
    // * `narrow_region` - smaller administrative area, such as county, if
    // detected
    // * `sublocality` - used in Asian addresses to demark a district within a
    // city, if detected
    ADDRESS = 10;

    // Date
    //
    // The metadata identifies the components of the date:
    //
    // * `year` - four digit year, if detected
    // * `month` - two digit month number, if detected
    // * `day` - two digit day number, if detected
    DATE = 11;

    // Number
    //
    // The metadata is the number itself.
    NUMBER = 12;

    // Price
    //
    // The metadata identifies the `value` and `currency`.
    PRICE = 13;
  }

  // The representative name for the entity.
  string name = 1;

  // The entity type.
  Type type = 2;

  // Metadata associated with the entity.
  //
  // For the metadata
  // associated with other entity types, see the Type table below.
  map<string, string> metadata = 3;

  // The mentions of this entity in the input document. The API currently
  // supports proper noun mentions.
  repeated EntityMention mentions = 5;

  // For calls to [AnalyzeEntitySentiment][] or if
  // [AnnotateTextRequest.Features.extract_entity_sentiment][google.cloud.language.v2.AnnotateTextRequest.Features.extract_entity_sentiment]
  // is set to true, this field will contain the aggregate sentiment expressed
  // for this entity in the provided document.
  Sentiment sentiment = 6;
}

// Represents the feeling associated with the entire text or entities in
// the text.
message Sentiment {
  // A non-negative number in the [0, +inf) range, which represents
  // the absolute magnitude of sentiment regardless of score (positive or
  // negative).
  float magnitude = 1;

  // Sentiment score between -1.0 (negative sentiment) and 1.0
  // (positive sentiment).
  float score = 2;
}

// Represents a mention for an entity in the text. Currently, proper noun
// mentions are supported.
message EntityMention {
  // The supported types of mentions.
  enum Type {
    // Unknown
    TYPE_UNKNOWN = 0;

    // Proper name
    PROPER = 1;

    // Common noun (or noun compound)
    COMMON = 2;
  }

  // The mention text.
  TextSpan text = 1;

  // The type of the entity mention.
  Type type = 2;

  // For calls to [AnalyzeEntitySentiment][] or if
  // [AnnotateTextRequest.Features.extract_entity_sentiment][google.cloud.language.v2.AnnotateTextRequest.Features.extract_entity_sentiment]
  // is set to true, this field will contain the sentiment expressed for this
  // mention of the entity in the provided document.
  Sentiment sentiment = 3;

  // Probability score associated with the entity.
  //
  // The score shows the probability of the entity mention being the entity
  // type. The score is in (0, 1] range.
  float probability = 4;
}

// Represents a text span in the input document.
message TextSpan {
  // The content of the text span, which is a substring of the document.
  string content = 1;

  // The API calculates the beginning offset of the content in the original
  // document according to the
  // [EncodingType][google.cloud.language.v2.EncodingType] specified in the API
  // request.
  int32 begin_offset = 2;
}

// Represents a category returned from the text classifier.
message ClassificationCategory {
  // The name of the category representing the document.
  string name = 1;

  // The classifier's confidence of the category. Number represents how certain
  // the classifier is that this category represents the given text.
  float confidence = 2;

  // Optional. The classifier's severity of the category. This is only present
  // when the ModerateTextRequest.ModelVersion is set to MODEL_VERSION_2, and
  // the corresponding category has a severity score.
  float severity = 3 [(google.api.field_behavior) = OPTIONAL];
}

// The sentiment analysis request message.
message AnalyzeSentimentRequest {
  // Required. Input document.
  Document document = 1 [(google.api.field_behavior) = REQUIRED];

  // The encoding type used by the API to calculate sentence offsets.
  EncodingType encoding_type = 2;
}

// The sentiment analysis response message.
message AnalyzeSentimentResponse {
  // The overall sentiment of the input document.
  Sentiment document_sentiment = 1;

  // The language of the text, which will be the same as the language specified
  // in the request or, if not specified, the automatically-detected language.
  // See [Document.language][] field for more details.
  string language_code = 2;

  // The sentiment for all the sentences in the document.
  repeated Sentence sentences = 3;

  // Whether the language is officially supported. The API may still return a
  // response when the language is not supported, but it is on a best effort
  // basis.
  bool language_supported = 4;
}

// The entity analysis request message.
message AnalyzeEntitiesRequest {
  // Required. Input document.
  Document document = 1 [(google.api.field_behavior) = REQUIRED];

  // The encoding type used by the API to calculate offsets.
  EncodingType encoding_type = 2;
}

// The entity analysis response message.
message AnalyzeEntitiesResponse {
  // The recognized entities in the input document.
  repeated Entity entities = 1;

  // The language of the text, which will be the same as the language specified
  // in the request or, if not specified, the automatically-detected language.
  // See [Document.language][] field for more details.
  string language_code = 2;

  // Whether the language is officially supported. The API may still return a
  // response when the language is not supported, but it is on a best effort
  // basis.
  bool language_supported = 3;
}

// The document classification request message.
message ClassifyTextRequest {
  // Required. Input document.
  Document document = 1 [(google.api.field_behavior) = REQUIRED];
}

// The document classification response message.
message ClassifyTextResponse {
  // Categories representing the input document.
  repeated ClassificationCategory categories = 1;

  // The language of the text, which will be the same as the language specified
  // in the request or, if not specified, the automatically-detected language.
  // See [Document.language][] field for more details.
  string language_code = 2;

  // Whether the language is officially supported. The API may still return a
  // response when the language is not supported, but it is on a best effort
  // basis.
  bool language_supported = 3;
}

// The document moderation request message.
message ModerateTextRequest {
  // The model version to use for ModerateText.
  enum ModelVersion {
    // The default model version.
    MODEL_VERSION_UNSPECIFIED = 0;

    // Use the v1 model, this model is used by default when not provided.
    // The v1 model only returns probability (confidence) score for each
    // category.
    MODEL_VERSION_1 = 1;

    // Use the v2 model.
    // The v2 model only returns probability (confidence) score for each
    // category, and returns severity score for a subset of the categories.
    MODEL_VERSION_2 = 2;
  }

  // Required. Input document.
  Document document = 1 [(google.api.field_behavior) = REQUIRED];

  // Optional. The model version to use for ModerateText.
  ModelVersion model_version = 2 [(google.api.field_behavior) = OPTIONAL];
}

// The document moderation response message.
message ModerateTextResponse {
  // Harmful and sensitive categories representing the input document.
  repeated ClassificationCategory moderation_categories = 1;

  // The language of the text, which will be the same as the language specified
  // in the request or, if not specified, the automatically-detected language.
  // See [Document.language][] field for more details.
  string language_code = 2;

  // Whether the language is officially supported. The API may still return a
  // response when the language is not supported, but it is on a best effort
  // basis.
  bool language_supported = 3;
}

// The request message for the text annotation API, which can perform multiple
// analysis types in one call.
message AnnotateTextRequest {
  // All available features.
  // Setting each one to true will enable that specific analysis for the input.
  message Features {
    // Optional. Extract entities.
    bool extract_entities = 1 [(google.api.field_behavior) = OPTIONAL];

    // Optional. Extract document-level sentiment.
    bool extract_document_sentiment = 2
        [(google.api.field_behavior) = OPTIONAL];

    // Optional. Classify the full document into categories.
    bool classify_text = 4 [(google.api.field_behavior) = OPTIONAL];

    // Optional. Moderate the document for harmful and sensitive categories.
    bool moderate_text = 5 [(google.api.field_behavior) = OPTIONAL];
  }

  // Required. Input document.
  Document document = 1 [(google.api.field_behavior) = REQUIRED];

  // Required. The enabled features.
  Features features = 2 [(google.api.field_behavior) = REQUIRED];

  // The encoding type used by the API to calculate offsets.
  EncodingType encoding_type = 3;
}

// The text annotations response message.
message AnnotateTextResponse {
  // Sentences in the input document. Populated if the user enables
  // [AnnotateTextRequest.Features.extract_document_sentiment][google.cloud.language.v2.AnnotateTextRequest.Features.extract_document_sentiment].
  repeated Sentence sentences = 1;

  // Entities, along with their semantic information, in the input document.
  // Populated if the user enables
  // [AnnotateTextRequest.Features.extract_entities][google.cloud.language.v2.AnnotateTextRequest.Features.extract_entities]
  // or
  // [AnnotateTextRequest.Features.extract_entity_sentiment][google.cloud.language.v2.AnnotateTextRequest.Features.extract_entity_sentiment].
  repeated Entity entities = 2;

  // The overall sentiment for the document. Populated if the user enables
  // [AnnotateTextRequest.Features.extract_document_sentiment][google.cloud.language.v2.AnnotateTextRequest.Features.extract_document_sentiment].
  Sentiment document_sentiment = 3;

  // The language of the text, which will be the same as the language specified
  // in the request or, if not specified, the automatically-detected language.
  // See [Document.language][] field for more details.
  string language_code = 4;

  // Categories identified in the input document.
  repeated ClassificationCategory categories = 5;

  // Harmful and sensitive categories identified in the input document.
  repeated ClassificationCategory moderation_categories = 6;

  // Whether the language is officially supported by all requested features.
  // The API may still return a response when the language is not supported, but
  // it is on a best effort basis.
  bool language_supported = 7;
}