@voice-ping/cognitive-services-speech
Version:
VoicePing Cognitive Services Speech SDK for JavaScript forked from Microsoft
154 lines (152 loc) • 9 kB
JavaScript
"use strict";
// Copyright (c) Microsoft Corporation. All rights reserved.
// Licensed under the MIT license.
var __extends = (this && this.__extends) || (function () {
var extendStatics = function (d, b) {
extendStatics = Object.setPrototypeOf ||
({ __proto__: [] } instanceof Array && function (d, b) { d.__proto__ = b; }) ||
function (d, b) { for (var p in b) if (b.hasOwnProperty(p)) d[p] = b[p]; };
return extendStatics(d, b);
};
return function (d, b) {
extendStatics(d, b);
function __() { this.constructor = d; }
d.prototype = b === null ? Object.create(b) : (__.prototype = b.prototype, new __());
};
})();
Object.defineProperty(exports, "__esModule", { value: true });
var Exports_1 = require("../sdk/Exports");
var Exports_2 = require("./Exports");
// tslint:disable-next-line:max-classes-per-file
var SpeechServiceRecognizer = /** @class */ (function (_super) {
__extends(SpeechServiceRecognizer, _super);
function SpeechServiceRecognizer(authentication, connectionFactory, audioSource, recognizerConfig, speechRecognizer) {
var _this = _super.call(this, authentication, connectionFactory, audioSource, recognizerConfig, speechRecognizer) || this;
_this.privSpeechRecognizer = speechRecognizer;
if (recognizerConfig.autoDetectSourceLanguages !== undefined) {
var sourceLanguages = recognizerConfig.autoDetectSourceLanguages.split(",");
_this.privSpeechContext.setSection("languageId", {
languages: sourceLanguages,
onSuccess: { action: "Recognize" },
onUnknown: { action: "None" }
});
_this.privSpeechContext.setSection("phraseOutput", {
interimResults: {
resultType: "Auto"
},
phraseResults: {
resultType: "Always"
}
});
}
return _this;
}
SpeechServiceRecognizer.prototype.processTypeSpecificMessages = function (connectionMessage) {
var result;
var resultProps = new Exports_1.PropertyCollection();
resultProps.setProperty(Exports_1.PropertyId.SpeechServiceResponse_JsonResult, connectionMessage.textBody);
var processed = false;
switch (connectionMessage.path.toLowerCase()) {
case "speech.hypothesis":
case "speech.fragment":
var hypothesis = Exports_2.SpeechHypothesis.fromJSON(connectionMessage.textBody);
var offset = hypothesis.Offset + this.privRequestSession.currentTurnAudioOffset;
result = new Exports_1.SpeechRecognitionResultCustom(this.privRequestSession.requestId, Exports_1.ResultReason.RecognizingSpeech, hypothesis.Text, hypothesis.Duration, offset, hypothesis.Language, hypothesis.LanguageDetectionConfidence, undefined, connectionMessage.textBody, resultProps, this.privRequestSession.requestId);
this.privRequestSession.onHypothesis(offset);
var ev = new Exports_1.SpeechRecognitionEventArgsCustom(result, hypothesis.Duration, this.privRequestSession.sessionId);
if (!!this.privSpeechRecognizer.recognizing) {
try {
this.privSpeechRecognizer.recognizing(this.privSpeechRecognizer, ev);
/* tslint:disable:no-empty */
}
catch (error) {
// Not going to let errors in the event handler
// trip things up.
}
}
processed = true;
break;
case "speech.phrase":
var simple = Exports_2.SimpleSpeechPhrase.fromJSON(connectionMessage.textBody);
var resultReason = Exports_2.EnumTranslation.implTranslateRecognitionResult(simple.RecognitionStatus);
this.privRequestSession.onPhraseRecognized(this.privRequestSession.currentTurnAudioOffset + simple.Offset + simple.Duration);
if (Exports_1.ResultReason.Canceled === resultReason) {
var cancelReason = Exports_2.EnumTranslation.implTranslateCancelResult(simple.RecognitionStatus);
this.cancelRecognitionLocal(cancelReason, Exports_2.EnumTranslation.implTranslateCancelErrorCode(simple.RecognitionStatus), undefined);
}
else {
if (!(this.privRequestSession.isSpeechEnded && resultReason === Exports_1.ResultReason.NoMatch && simple.RecognitionStatus !== Exports_2.RecognitionStatus.InitialSilenceTimeout)) {
if (this.privRecognizerConfig.parameters.getProperty(Exports_2.OutputFormatPropertyName) === Exports_1.OutputFormat[Exports_1.OutputFormat.Simple]) {
result = new Exports_1.SpeechRecognitionResultCustom(this.privRequestSession.requestId, resultReason, simple.DisplayText, simple.Duration, simple.Offset + this.privRequestSession.currentTurnAudioOffset, simple.Language, simple.LanguageDetectionConfidence, undefined, connectionMessage.textBody, resultProps, this.privRequestSession.requestId);
}
else {
var detailed = Exports_2.DetailedSpeechPhrase.fromJSON(connectionMessage.textBody);
result = new Exports_1.SpeechRecognitionResultCustom(this.privRequestSession.requestId, resultReason, detailed.RecognitionStatus === Exports_2.RecognitionStatus.Success ? detailed.NBest[0].Display : undefined, detailed.Duration, detailed.Offset + this.privRequestSession.currentTurnAudioOffset, detailed.Language, detailed.LanguageDetectionConfidence, undefined, connectionMessage.textBody, resultProps, this.privRequestSession.requestId);
}
var event_1 = new Exports_1.SpeechRecognitionEventArgsCustom(result, result.offset, this.privRequestSession.sessionId);
if (!!this.privSpeechRecognizer.recognized) {
try {
this.privSpeechRecognizer.recognized(this.privSpeechRecognizer, event_1);
/* tslint:disable:no-empty */
}
catch (error) {
// Not going to let errors in the event handler
// trip things up.
}
}
}
if (!!this.privSuccessCallback) {
try {
this.privSuccessCallback(result);
}
catch (e) {
if (!!this.privErrorCallback) {
this.privErrorCallback(e);
}
}
// Only invoke the call back once.
// and if it's successful don't invoke the
// error after that.
this.privSuccessCallback = undefined;
this.privErrorCallback = undefined;
}
}
processed = true;
break;
default:
break;
}
return processed;
};
// Cancels recognition.
SpeechServiceRecognizer.prototype.cancelRecognition = function (sessionId, requestId, cancellationReason, errorCode, error) {
var properties = new Exports_1.PropertyCollection();
properties.setProperty(Exports_2.CancellationErrorCodePropertyName, Exports_1.CancellationErrorCode[errorCode]);
if (!!this.privSpeechRecognizer.canceled) {
var cancelEvent = new Exports_1.SpeechRecognitionCanceledEventArgs(cancellationReason, error, errorCode, undefined, sessionId);
try {
this.privSpeechRecognizer.canceled(this.privSpeechRecognizer, cancelEvent);
/* tslint:disable:no-empty */
}
catch (_a) { }
}
if (!!this.privSuccessCallback) {
var result = new Exports_1.SpeechRecognitionResultCustom(requestId, Exports_1.ResultReason.Canceled, undefined, // Text
undefined, // Duration
undefined, // Offset
undefined, // Language
undefined, // Language Detection Confidence
error, undefined, // Json
properties, undefined);
try {
this.privSuccessCallback(result);
this.privSuccessCallback = undefined;
/* tslint:disable:no-empty */
}
catch (_b) { }
}
};
return SpeechServiceRecognizer;
}(Exports_2.ServiceRecognizerBase));
exports.SpeechServiceRecognizer = SpeechServiceRecognizer;
//# sourceMappingURL=SpeechServiceRecognizer.js.map