@aristech-org/stt-client/dist/generated/stt_service.d.ts

A Node.js client library for the Aristech Speech-to-Text API

import { BinaryReader, BinaryWriter } from "@bufbuild/protobuf/wire";
import { type CallOptions, ChannelCredentials, Client, type ClientDuplexStream, type ClientOptions, type ClientUnaryCall, type handleBidiStreamingCall, type handleUnaryCall, Metadata, type ServiceError, type UntypedServiceImplementation } from "@grpc/grpc-js";
import { Duration } from "./google/protobuf/duration.js";
export declare const protobufPackage = "ari.stt.v1";
export declare enum EndpointingType {
    /** LM - Endpointing that considers the language model. */
    LM = 0,
    /** VAD - VAD based endpointing. */
    VAD = 1,
    UNRECOGNIZED = -1
}
export declare function endpointingTypeFromJSON(object: any): EndpointingType;
export declare function endpointingTypeToJSON(object: EndpointingType): string;
export declare enum ModelType {
    /** CORE_STT - STT-Core models */
    CORE_STT = 0,
    /** GRAMMAR_STT - Grammar only models */
    GRAMMAR_STT = 1,
    /** MULTITASK_STT - Multitask models */
    MULTITASK_STT = 2,
    /** DIARIZATION - Speaker diarization model type. */
    DIARIZATION = 3,
    UNRECOGNIZED = -1
}
export declare function modelTypeFromJSON(object: any): ModelType;
export declare function modelTypeToJSON(object: ModelType): string;
export declare enum GrammarType {
    /**
     * JSGF - JSGF grammar type.
     * Example grammar: `jsgf:<yes_no> = yes | no;`
     */
    JSGF = 0,
    /**
     * SRGS - SRGS grammar type.
     * Example grammar: `srgs:$yes_no = yes | no;`
     */
    SRGS = 3,
    /**
     * KWS - Keyword / Keyphrase spotting grammar type.
     * Example grammar: `kws:oh mighty computer|hey computer`
     */
    KWS = 1,
    /**
     * PHRASE_LIST - A simple json phrase list grammar type.
     * Example grammar: `["yes", "yeah", "yep", "why not", "no", "nope"]`
     */
    PHRASE_LIST = 2,
    UNRECOGNIZED = -1
}
export declare function grammarTypeFromJSON(object: any): GrammarType;
export declare function grammarTypeToJSON(object: GrammarType): string;
/** The top-level message sent by the client for the `StreamingRecognize` method. */
export interface StreamingRecognitionRequest {
    /**
     * The configuration for the stream.
     * This is the first message that must be sent.
     */
    config?: RecognitionConfig | undefined;
    /** The audio data to be recognized. */
    audioContent?: Uint8Array | undefined;
}
/** The top-level message returned from the `StreamingRecognize` method. */
export interface StreamingRecognitionResponse {
    /** List of results that are currently available. */
    chunks: SpeechRecognitionChunk[];
    /**
     * A short id that is used in the stt-server logs to differentiate between
     * different client requests. Be aware that this id should only be used for
     * debugging purposes because it is not collision safe.
     */
    clientId: string;
    /** The language identified by the server, e.g. en, de, etc. */
    language: string;
}
/**
 * The `RecognitionConfig` message provides information to the recognizer that
 * specifies how to process the request.
 */
export interface RecognitionConfig {
    /**
     * Specifies what kind of audio is being sent and how the recognizer should
     * process it.
     */
    specification: RecognitionSpec | undefined;
}
/**
 * The `RecognitionSpec` message provides information to the recognizer that
 * specifies how to process the request.
 */
export interface RecognitionSpec {
    /** At the moment only LINEAR16 is supported. */
    audioEncoding: RecognitionSpec_AudioEncoding;
    /** 8000, 16000, 48000 only for pcm. */
    sampleRateHertz: number;
    /** [language[_territory]] e.g. en, en-IN, de. */
    locale: string;
    /** load a specific graph for the locale specific model (e.g. yes_no). */
    graph: string;
    /**
     * Allows to specify a grammar to be used for the recognition.
     * To specify a JSGF grammar for example set grammar to `jsgf:public <yes_no>
     * = yes | no;` To spot a keyword / phrase, set grammar to `kws:oh mighty
     * computer` You can also specify a json string to narrow the possible words
     * to appear `["oh one two three four five six seven eight nine zero",
     * "[unk]"]`.
     */
    grammar: string;
    /**
     * If set true, tentative hypotheses may be returned as they become available
     * (final=false flag) If false or omitted, only final=true result(s) are
     * returned. Makes sense only for StreamingRecognize requests.
     */
    partialResults: boolean;
    /** Decode as single utterance. */
    singleUtterance: boolean;
    /** Specifies how text should be normalized. */
    normalization: NormalizationSpec | undefined;
    /**
     * When set, the recognizer opts out of MBR decoding and produces phoneme
     * infos.
     */
    phones: boolean;
    /**
     * Instead of picking a model based on the locale, this field can be used to
     * specify a specific model directly.
     * To specify a graph model directly use model:graph e.g.
     * generic-model-de-0.21:ja_nein
     */
    model: string;
    /**
     * For models that use endpointing (e.g. STT-Core models) this field can be
     * used to specify the endpointing configuration.
     */
    endpointing: EndpointSpec | undefined;
    /**
     * For models that use voice activity detection (VAD) this field can be used
     * to specify the VAD configuration.
     */
    vad: VadSpec | undefined;
    /**
     * Some models allow to specify a prompt that can be used to give the model
     * some context on what was said before or to steer the model to use
     * particular spellings or styles.
     */
    prompt: string;
}
export declare enum RecognitionSpec_AudioEncoding {
    /** AUDIO_ENCODING_UNSPECIFIED - If not specified, defaults to LINEAR16_PCM. */
    AUDIO_ENCODING_UNSPECIFIED = 0,
    /** LINEAR16_PCM - 16-bit signed little-endian (Linear PCM) */
    LINEAR16_PCM = 1,
    UNRECOGNIZED = -1
}
export declare function recognitionSpec_AudioEncodingFromJSON(object: any): RecognitionSpec_AudioEncoding;
export declare function recognitionSpec_AudioEncodingToJSON(object: RecognitionSpec_AudioEncoding): string;
/**
 * The `NormalizationSpec` message provides information to the recognizer that
 * specifies which normalizer to use.
 */
export interface NormalizationSpec {
    /**
     * Whether to strip unknown label in the resulting text
     * Note: The unknown label can still be accessed via the `words` field
     */
    stripUnk: boolean;
    /**
     * Allows to specify which nlp functions should be applied to the text
     * before it is returned. By specifying this field, the default nlp
     * configuration is overwritten.
     */
    nlp: NLPSpec | undefined;
}
export interface NLPSpec {
    /** The server config name of the server that provides the nlp functions. */
    serverConfig: string;
    /** Which nlp functions should be applied to the text before it is returned. */
    functions: NLPFunctionSpec[];
    /** Whether to apply the nlp functions to the partial results. */
    partialResults: boolean;
    /** Optional global argument. */
    args: string;
    /** Specifies which field should be used as input for the nlp functions. */
    inputField: NLPSpec_NlpInputField;
}
export declare enum NLPSpec_NlpInputField {
    /** UNSPECIFIED - If not specified, defaults to TEXT. */
    UNSPECIFIED = 0,
    /** TEXT - The text field is used as input for the nlp processing. */
    TEXT = 1,
    /** TAGGED_TEXT - Use the tagged_text field as input for the nlp processing. */
    TAGGED_TEXT = 2,
    /** SLOTTED_TEXT - Use the slotted_text field as input for the nlp processing. */
    SLOTTED_TEXT = 3,
    UNRECOGNIZED = -1
}
export declare function nLPSpec_NlpInputFieldFromJSON(object: any): NLPSpec_NlpInputField;
export declare function nLPSpec_NlpInputFieldToJSON(object: NLPSpec_NlpInputField): string;
/**
 * The `NLPFunction` message provides information to the recognizer that
 * specifies which nlp function to use.
 */
export interface NLPFunctionSpec {
    /** The id of the nlp function (e.g. `ner-de`). */
    id: string;
    /** Optional additional parameters for the nlp function (e.g. `ANONYMIZE`). */
    args: string[];
}
/** Endpointing configuration for LM based endpointing. */
export interface EndpointSpec {
    /**
     * How many seconds of non-speech before the endpointer triggers to
     * clean up the buffer.
     */
    silenceTimeout: number;
    /**
     * How many seconds of non-speech after some speech with high
     * probability for an endpoint before the endpointer triggers.
     */
    trailingSilenceHighProbability: number;
    /**
     * How many seconds of non-speech after some speech with low probability
     * for an endpoint before the endpointer triggers.
     */
    trailingSilenceOkProbability: number;
    /**
     * How many seconds of non-speech without the endpointer reaching a
     * final state before the endpointer triggers.
     */
    trailingSilenceNoEndpoint: number;
    /**
     * After how many seconds of audio to trigger and endpoint regardless
     * of anything else.
     */
    utteranceTimeout: number;
}
/**
 * Endpointing configuration for Voice activity detection (VAD) based
 * endpointing.
 */
export interface VadSpec {
    /**
     * The threshold between 0 and 1.0 to determine if a frame is speech or
     * non-speech. A higher threshold will result in less false positives but also
     * some speech might be cut off.
     */
    threshold: number;
    /**
     * Amount of trailing silence before an utterance is considered after a speech
     * to non-speech transition.
     */
    trailingSilence: number;
    /**
     * The minimum duration of speech in seconds before trying to perform a
     * partial recognition.
     */
    minSpeech: number;
}
/**
 * The `SpeechRecognitionChunk` message contains the result of a single
 * utterance.
 */
export interface SpeechRecognitionChunk {
    /** The transcription alternatives. */
    alternatives: SpeechRecognitionAlternative[];
    /** This flag indicates if the transcription is final or not. */
    final: boolean;
    /** This flag shows that the received chunk is the end of an utterance. */
    endOfUtterance: boolean;
}
/**
 * The `SpeechRecognitionAlternative` message contains one alternative of a
 * transcription.
 */
export interface SpeechRecognitionAlternative {
    /** The raw recognized text. */
    text: string;
    /**
     * When the model is composed of multiple nested language models, this field
     * contains the recognized text including xml tags that indicate which
     * language model produced which part of the text. e.g. "i live in <address>
     * <number> 21 </number> <street> jumpstreet </street> <city> heidelberg
     * </city> </address>"
     */
    slottedText: string;
    /** The tagged recognized text. */
    taggedText: string;
    /** The nlp result. */
    nlpText: string;
    /** The overall confidence of the recognition result. */
    confidence: number;
    /**
     * Word level infos such as start and end time offsets, word level
     * confidences, or phoneme infos.
     */
    words: WordInfo[];
}
/** The `WordInfo` message contains the word level information. */
export interface WordInfo {
    /** The word's start time, in seconds. */
    startTime: Duration | undefined;
    /** The word's end time, in seconds. */
    endTime: Duration | undefined;
    /** The word. */
    word: string;
    /** The confidence of the word in the range [0.0, 1.0]. */
    confidence: number;
    /** Phoneme infos. */
    phones: PhoneInfo[];
    /**
     * Speech recognition slot the word belongs to.
     * For nested slots, the slots are joined with a dot and ordered from outer to
     * inner
     * e.g.: "i live in <address> <number> 21 </number> <street> jumpstreet
     * </street> heidelberg </address>"
     * will have the following slots:
     * i -> ''
     * live -> ''
     * in -> ''
     * 21 -> 'address.number'
     * jumpstreet -> 'address.street'
     * heidelberg -> 'address'
     */
    slot: string;
}
/** The `PhoneInfo` message contains the phoneme level information. */
export interface PhoneInfo {
    /** The phone's start time, in seconds. */
    startTime: Duration | undefined;
    /** The phone's end time, in seconds. */
    endTime: Duration | undefined;
    /** The phone. */
    phone: string;
}
/** The `ModelsRequest` message currently contains no information. */
export interface ModelsRequest {
}
/** The `ModelsResponse` message contains the list of supported models. */
export interface ModelsResponse {
    /** List of supported models. */
    model: Model[];
}
/** The `Model` message contains the information about a single model. */
export interface Model {
    /**
     * The model id.
     * e.g. generic-model-de-0.21
     */
    id: string;
    /**
     * Alias that can also be used to refer to the model instead of the id.
     * e.g. generic-de or german-large
     */
    alias: string[];
    /**
     * The human readable model name (for display purposes).
     * e.g. German Generic Model (Large)
     */
    name: string;
    /** The model description. */
    description: string;
    /** The model version. */
    version: string;
    /** The model type. */
    type: ModelType;
    /** The locale(s) supported by the model. */
    locale: string[];
    /** Which grammar types are supported by the model. */
    grammarType: GrammarType[];
    /** The NLP preconfiguration for this model (if any). */
    nlp: NLPSpec | undefined;
    /** The slots the model potentially outputs. */
    slots: string[];
    /** Examples of what the model can recognize. */
    examples: string[];
    /** The supported endpointing modes. */
    endpointing: EndpointingType[];
}
/** The `NLPFunctionsRequest` message currently contains no information. */
export interface NLPFunctionsRequest {
}
/**
 * The `NLPFunctionsResponse` message contains the list of supported nlp
 * servers and the corresponding functions.
 */
export interface NLPFunctionsResponse {
    /** List of supported nlp servers. */
    server: NLPFunctionServer[];
}
export interface NLPFunctionServer {
    /** The nlp server configuration name (to be used in `NLPSpec`). */
    serverConfig: string;
    /** The nlp functions supported by the nlp server. */
    functions: NLPFunction[];
}
/**
 * The `NLPFunction` message contains the information about a single nlp
 * function.
 */
export interface NLPFunction {
    /** The nlp function id. */
    id: string;
    /** The nlp function name. */
    name: string;
    /** The nlp function description. */
    description: string;
}
/** The `Graph` message contains the information about a single graph. */
export interface Graph {
    /** The name of the graph */
    name: string;
}
/** The `AccountInfoRequest` message currently contains no information. */
export interface AccountInfoRequest {
}
/** The `AccountInfoResponse` message contains the account information. */
export interface AccountInfoResponse {
    /** The account token. */
    token: string;
    /** The account display name. */
    displayName: string;
    /** How many requests were made with this account. */
    totalRequests: number;
    /** How many seconds of audio this account has booked. */
    bookedSeconds: number;
    /** How many seconds of audio this account has used. */
    usedSeconds: number;
    /** Expiration date of the account as unix timestamp (-1 for unlimited). */
    expirationDate: number;
    /** Whether the account is blocked. */
    blocked: boolean;
}
/** The `NLPProcessRequest` message contains the text to be processed. */
export interface NLPProcessRequest {
    /** The text to be processed. */
    text: string;
    /** The nlp specification. */
    nlp: NLPSpec | undefined;
}
/** The `NLPProcessResponse` message contains the processed text. */
export interface NLPProcessResponse {
    /** The processed text. */
    text: string;
}
export declare const StreamingRecognitionRequest: MessageFns<StreamingRecognitionRequest>;
export declare const StreamingRecognitionResponse: MessageFns<StreamingRecognitionResponse>;
export declare const RecognitionConfig: MessageFns<RecognitionConfig>;
export declare const RecognitionSpec: MessageFns<RecognitionSpec>;
export declare const NormalizationSpec: MessageFns<NormalizationSpec>;
export declare const NLPSpec: MessageFns<NLPSpec>;
export declare const NLPFunctionSpec: MessageFns<NLPFunctionSpec>;
export declare const EndpointSpec: MessageFns<EndpointSpec>;
export declare const VadSpec: MessageFns<VadSpec>;
export declare const SpeechRecognitionChunk: MessageFns<SpeechRecognitionChunk>;
export declare const SpeechRecognitionAlternative: MessageFns<SpeechRecognitionAlternative>;
export declare const WordInfo: MessageFns<WordInfo>;
export declare const PhoneInfo: MessageFns<PhoneInfo>;
export declare const ModelsRequest: MessageFns<ModelsRequest>;
export declare const ModelsResponse: MessageFns<ModelsResponse>;
export declare const Model: MessageFns<Model>;
export declare const NLPFunctionsRequest: MessageFns<NLPFunctionsRequest>;
export declare const NLPFunctionsResponse: MessageFns<NLPFunctionsResponse>;
export declare const NLPFunctionServer: MessageFns<NLPFunctionServer>;
export declare const NLPFunction: MessageFns<NLPFunction>;
export declare const Graph: MessageFns<Graph>;
export declare const AccountInfoRequest: MessageFns<AccountInfoRequest>;
export declare const AccountInfoResponse: MessageFns<AccountInfoResponse>;
export declare const NLPProcessRequest: MessageFns<NLPProcessRequest>;
export declare const NLPProcessResponse: MessageFns<NLPProcessResponse>;
export type SttServiceService = typeof SttServiceService;
export declare const SttServiceService: {
    /** Transcribe a stream of audio. */
    readonly streamingRecognize: {
        readonly path: "/ari.stt.v1.SttService/StreamingRecognize";
        readonly requestStream: true;
        readonly responseStream: true;
        readonly requestSerialize: (value: StreamingRecognitionRequest) => Buffer<ArrayBuffer>;
        readonly requestDeserialize: (value: Buffer) => StreamingRecognitionRequest;
        readonly responseSerialize: (value: StreamingRecognitionResponse) => Buffer<ArrayBuffer>;
        readonly responseDeserialize: (value: Buffer) => StreamingRecognitionResponse;
    };
    /** List all supported models. */
    readonly models: {
        readonly path: "/ari.stt.v1.SttService/Models";
        readonly requestStream: false;
        readonly responseStream: false;
        readonly requestSerialize: (value: ModelsRequest) => Buffer<ArrayBuffer>;
        readonly requestDeserialize: (value: Buffer) => ModelsRequest;
        readonly responseSerialize: (value: ModelsResponse) => Buffer<ArrayBuffer>;
        readonly responseDeserialize: (value: Buffer) => ModelsResponse;
    };
    /** List all available nlp server configs and corresponding functions. */
    readonly nlpFunctions: {
        readonly path: "/ari.stt.v1.SttService/NLPFunctions";
        readonly requestStream: false;
        readonly responseStream: false;
        readonly requestSerialize: (value: NLPFunctionsRequest) => Buffer<ArrayBuffer>;
        readonly requestDeserialize: (value: Buffer) => NLPFunctionsRequest;
        readonly responseSerialize: (value: NLPFunctionsResponse) => Buffer<ArrayBuffer>;
        readonly responseDeserialize: (value: Buffer) => NLPFunctionsResponse;
    };
    readonly accountInfo: {
        readonly path: "/ari.stt.v1.SttService/AccountInfo";
        readonly requestStream: false;
        readonly responseStream: false;
        readonly requestSerialize: (value: AccountInfoRequest) => Buffer<ArrayBuffer>;
        readonly requestDeserialize: (value: Buffer) => AccountInfoRequest;
        readonly responseSerialize: (value: AccountInfoResponse) => Buffer<ArrayBuffer>;
        readonly responseDeserialize: (value: Buffer) => AccountInfoResponse;
    };
    /** Processes the given text with the given nlp pipeline. */
    readonly nlpProcess: {
        readonly path: "/ari.stt.v1.SttService/NLPProcess";
        readonly requestStream: false;
        readonly responseStream: false;
        readonly requestSerialize: (value: NLPProcessRequest) => Buffer<ArrayBuffer>;
        readonly requestDeserialize: (value: Buffer) => NLPProcessRequest;
        readonly responseSerialize: (value: NLPProcessResponse) => Buffer<ArrayBuffer>;
        readonly responseDeserialize: (value: Buffer) => NLPProcessResponse;
    };
};
export interface SttServiceServer extends UntypedServiceImplementation {
    /** Transcribe a stream of audio. */
    streamingRecognize: handleBidiStreamingCall<StreamingRecognitionRequest, StreamingRecognitionResponse>;
    /** List all supported models. */
    models: handleUnaryCall<ModelsRequest, ModelsResponse>;
    /** List all available nlp server configs and corresponding functions. */
    nlpFunctions: handleUnaryCall<NLPFunctionsRequest, NLPFunctionsResponse>;
    accountInfo: handleUnaryCall<AccountInfoRequest, AccountInfoResponse>;
    /** Processes the given text with the given nlp pipeline. */
    nlpProcess: handleUnaryCall<NLPProcessRequest, NLPProcessResponse>;
}
export interface SttServiceClient extends Client {
    /** Transcribe a stream of audio. */
    streamingRecognize(): ClientDuplexStream<StreamingRecognitionRequest, StreamingRecognitionResponse>;
    streamingRecognize(options: Partial<CallOptions>): ClientDuplexStream<StreamingRecognitionRequest, StreamingRecognitionResponse>;
    streamingRecognize(metadata: Metadata, options?: Partial<CallOptions>): ClientDuplexStream<StreamingRecognitionRequest, StreamingRecognitionResponse>;
    /** List all supported models. */
    models(request: ModelsRequest, callback: (error: ServiceError | null, response: ModelsResponse) => void): ClientUnaryCall;
    models(request: ModelsRequest, metadata: Metadata, callback: (error: ServiceError | null, response: ModelsResponse) => void): ClientUnaryCall;
    models(request: ModelsRequest, metadata: Metadata, options: Partial<CallOptions>, callback: (error: ServiceError | null, response: ModelsResponse) => void): ClientUnaryCall;
    /** List all available nlp server configs and corresponding functions. */
    nlpFunctions(request: NLPFunctionsRequest, callback: (error: ServiceError | null, response: NLPFunctionsResponse) => void): ClientUnaryCall;
    nlpFunctions(request: NLPFunctionsRequest, metadata: Metadata, callback: (error: ServiceError | null, response: NLPFunctionsResponse) => void): ClientUnaryCall;
    nlpFunctions(request: NLPFunctionsRequest, metadata: Metadata, options: Partial<CallOptions>, callback: (error: ServiceError | null, response: NLPFunctionsResponse) => void): ClientUnaryCall;
    accountInfo(request: AccountInfoRequest, callback: (error: ServiceError | null, response: AccountInfoResponse) => void): ClientUnaryCall;
    accountInfo(request: AccountInfoRequest, metadata: Metadata, callback: (error: ServiceError | null, response: AccountInfoResponse) => void): ClientUnaryCall;
    accountInfo(request: AccountInfoRequest, metadata: Metadata, options: Partial<CallOptions>, callback: (error: ServiceError | null, response: AccountInfoResponse) => void): ClientUnaryCall;
    /** Processes the given text with the given nlp pipeline. */
    nlpProcess(request: NLPProcessRequest, callback: (error: ServiceError | null, response: NLPProcessResponse) => void): ClientUnaryCall;
    nlpProcess(request: NLPProcessRequest, metadata: Metadata, callback: (error: ServiceError | null, response: NLPProcessResponse) => void): ClientUnaryCall;
    nlpProcess(request: NLPProcessRequest, metadata: Metadata, options: Partial<CallOptions>, callback: (error: ServiceError | null, response: NLPProcessResponse) => void): ClientUnaryCall;
}
export declare const SttServiceClient: {
    new (address: string, credentials: ChannelCredentials, options?: Partial<ClientOptions>): SttServiceClient;
    service: typeof SttServiceService;
    serviceName: string;
};
type Builtin = Date | Function | Uint8Array | string | number | boolean | undefined;
export type DeepPartial<T> = T extends Builtin ? T : T extends globalThis.Array<infer U> ? globalThis.Array<DeepPartial<U>> : T extends ReadonlyArray<infer U> ? ReadonlyArray<DeepPartial<U>> : T extends {} ? {
    [K in keyof T]?: DeepPartial<T[K]>;
} : Partial<T>;
type KeysOfUnion<T> = T extends T ? keyof T : never;
export type Exact<P, I extends P> = P extends Builtin ? P : P & {
    [K in keyof P]: Exact<P[K], I[K]>;
} & {
    [K in Exclude<keyof I, KeysOfUnion<P>>]: never;
};
export interface MessageFns<T> {
    encode(message: T, writer?: BinaryWriter): BinaryWriter;
    decode(input: BinaryReader | Uint8Array, length?: number): T;
    fromJSON(object: any): T;
    toJSON(message: T): unknown;
    create<I extends Exact<DeepPartial<T>, I>>(base?: I): T;
    fromPartial<I extends Exact<DeepPartial<T>, I>>(object: I): T;
}
export {};