Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

✨ Improve Web-Client browser-detection #848

Merged
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions jovo-clients/jovo-client-web/package.json
Original file line number Diff line number Diff line change
Expand Up @@ -19,6 +19,7 @@
"author": "jovotech",
"license": "Apache-2.0",
"dependencies": {
"detect-browser": "^5.2.0",
"jovo-platform-web": "^3.2.2",
"lodash.defaultsdeep": "^4.6.1",
"lodash.get": "^4.4.2",
Expand Down
35 changes: 16 additions & 19 deletions jovo-clients/jovo-client-web/src/Client.ts
Original file line number Diff line number Diff line change
Expand Up @@ -67,6 +67,22 @@ export interface Config {
}

export class Client extends EventEmitter {
get isInitialized(): boolean {
return this.initialized;
}

get isPlayingAudio(): boolean {
return this.$audioPlayer.isPlaying || this.$speechSynthesizer.isSpeaking;
}

get isRecordingInput(): boolean {
return this.$audioRecorder.isRecording || this.$speechRecognizer.isRecording;
}

get isUsingSpeechRecognition(): boolean {
return this.useSpeechRecognition;
}

static getDefaultConfig(): Config {
return {
version: '3.2.1',
Expand Down Expand Up @@ -96,13 +112,10 @@ export class Client extends EventEmitter {
readonly $speechRecognizer: SpeechRecognizer;
readonly $speechSynthesizer: SpeechSynthesizer;
readonly $store: Store;

readonly config: Config;

readonly $actionHandler: ActionHandler;
readonly $repromptHandler: RepromptHandler;
readonly $ssmlHandler: SSMLHandler;

private useSpeechRecognition = true;
private isInputProcessOngoing = false;
private initialized = false;
Expand Down Expand Up @@ -173,22 +186,6 @@ export class Client extends EventEmitter {
});
}

get isInitialized(): boolean {
return this.initialized;
}

get isPlayingAudio(): boolean {
return this.$audioPlayer.isPlaying || this.$speechSynthesizer.isSpeaking;
}

get isRecordingInput(): boolean {
return this.$audioRecorder.isRecording || this.$speechRecognizer.isRecording;
}

get isUsingSpeechRecognition(): boolean {
return this.useSpeechRecognition;
}

addListener(event: ClientEvent.Request, listener: ClientRequestListener): this;
addListener(event: ClientEvent.Response, listener: ClientResponseListener): this;
addListener(event: ClientEvent.Action, listener: ClientActionListener): this;
Expand Down
3 changes: 0 additions & 3 deletions jovo-clients/jovo-client-web/src/core/RepromptHandler.ts
Original file line number Diff line number Diff line change
Expand Up @@ -16,13 +16,10 @@ export class RepromptHandler {
maxAttempts: 1,
};
}

private actions: Action[] = [];
private attempts = 0;
private hasAddedEvents = false;

private useSpeechRecognition = true;

private timeoutFn = this.onInputTimeout.bind(this);
private endFn = this.onInputEnd.bind(this);

Expand Down
38 changes: 18 additions & 20 deletions jovo-clients/jovo-client-web/src/standalone/AudioPlayer.ts
Original file line number Diff line number Diff line change
Expand Up @@ -23,26 +23,6 @@ export interface AudioPlayerConfig {
}

export class AudioPlayer extends EventEmitter {
static getDefaultConfig(): AudioPlayerConfig {
return {
enabled: true,
};
}

readonly config: AudioPlayerConfig;

private $volume = 1.0;
private audio: HTMLAudioElement | null = null;
private isAudioPlaying: boolean = false;
private initialized = false;

constructor(config?: DeepPartial<AudioPlayerConfig>) {
super();

const defaultConfig = AudioPlayer.getDefaultConfig();
this.config = config ? _defaultsDeep(config, defaultConfig) : defaultConfig;
}

get isInitialized(): boolean {
return this.initialized;
}
Expand Down Expand Up @@ -74,6 +54,24 @@ export class AudioPlayer extends EventEmitter {
return !!this.audio && !this.audio.ended;
}

static getDefaultConfig(): AudioPlayerConfig {
return {
enabled: true,
};
}
readonly config: AudioPlayerConfig;
private $volume = 1.0;
private audio: HTMLAudioElement | null = null;
private isAudioPlaying: boolean = false;
private initialized = false;

constructor(config?: DeepPartial<AudioPlayerConfig>) {
super();

const defaultConfig = AudioPlayer.getDefaultConfig();
this.config = config ? _defaultsDeep(config, defaultConfig) : defaultConfig;
}

addListener(event: AudioPlayerVoidEvents, listener: VoidListener): this;
addListener(event: AudioPlayerEvent.Play, listener: AudioPlayerPlayListener): this;
addListener(event: AudioPlayerEvent.Error, listener: ErrorListener): this;
Expand Down
59 changes: 27 additions & 32 deletions jovo-clients/jovo-client-web/src/standalone/AudioRecorder.ts
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
import { EventEmitter } from 'events';
import _defaultsDeep from 'lodash.defaultsdeep';
import { DeepPartial, OSHelper, VoidListener } from '..';
import { DeepPartial, OSDetector, VoidListener } from '..';

interface AudioRecorderNodes {
inputStream?: MediaStreamAudioSourceNode;
Expand Down Expand Up @@ -62,6 +62,30 @@ export interface AudioRecorderConfig {
}

export class AudioRecorder extends EventEmitter {
get isInitialized(): boolean {
return this.initialized;
}

get isRecording(): boolean {
return this.recording;
}

get startDetectionEnabled(): boolean {
return !!(
this.config.startDetection.enabled &&
this.config.startDetection.threshold &&
this.config.startDetection.timeoutInMs
);
}

get silenceDetectionEnabled(): boolean {
return !!(
this.config.silenceDetection.enabled &&
this.config.silenceDetection.threshold &&
this.config.silenceDetection.timeoutInMs
);
}

static getDefaultConfig(): AudioRecorderConfig {
return {
sampleRate: 16000,
Expand Down Expand Up @@ -89,19 +113,14 @@ export class AudioRecorder extends EventEmitter {
},
};
}

readonly config: AudioRecorderConfig;

private readonly audioNodes: AudioRecorderNodes;
private audioCtx: AudioContext | null;
private audioStream: MediaStream | null;

private initialized = false;

private recording = false;
private recordingStartedAt?: Date;
private startThresholdPassed = false;

private chunks: Float32Array[] = [];
private chunkLength = 0;

Expand All @@ -117,30 +136,6 @@ export class AudioRecorder extends EventEmitter {
this.audioStream = null;
}

get isInitialized(): boolean {
return this.initialized;
}

get isRecording(): boolean {
return this.recording;
}

get startDetectionEnabled(): boolean {
return !!(
this.config.startDetection.enabled &&
this.config.startDetection.threshold &&
this.config.startDetection.timeoutInMs
);
}

get silenceDetectionEnabled(): boolean {
return !!(
this.config.silenceDetection.enabled &&
this.config.silenceDetection.threshold &&
this.config.silenceDetection.timeoutInMs
);
}

addListener(event: AudioRecorderVoidEvents, listener: VoidListener): this;
addListener(event: AudioRecorderEvent.Stop, listener: AudioRecorderStopListener): this;
addListener(
Expand Down Expand Up @@ -219,7 +214,7 @@ export class AudioRecorder extends EventEmitter {
}
this.checkForBrowserCompatibility();

if (OSHelper.isWindows) {
if (OSDetector.isWindows()) {
if (!this.audioStream) {
this.audioStream = await navigator.mediaDevices.getUserMedia({
audio: this.config.audioConstraints,
Expand Down Expand Up @@ -307,7 +302,7 @@ export class AudioRecorder extends EventEmitter {
this.audioNodes.inputGain?.disconnect();
this.audioNodes.inputStream?.disconnect();

if (this.audioStream && !OSHelper.isWindows) {
if (this.audioStream && !OSDetector.isWindows) {
this.audioStream.getTracks().forEach((track) => {
track.stop();
});
Expand Down
63 changes: 31 additions & 32 deletions jovo-clients/jovo-client-web/src/standalone/SpeechRecognizer.ts
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
import { EventEmitter } from 'events';
import _defaultsDeep from 'lodash.defaultsdeep';
import { DeepPartial, ErrorListener, VoidListener } from '..';
import { BrowserDetector, DeepPartial, ErrorListener, VoidListener } from '..';

// TODO maybe rename SpeechRecognized to Processing to have almost identical events as the AudioRecorder
export enum SpeechRecognizerEvent {
Expand Down Expand Up @@ -42,8 +42,36 @@ export interface SpeechRecognizerConfig extends SpeechRecognitionConfig {
}

export class SpeechRecognizer extends EventEmitter {
get isRecording(): boolean {
return this.recording;
}

get isAvailable(): boolean {
return !!this.recognition;
}

get startDetectionEnabled(): boolean {
return !!(
this.config.continuous &&
this.config.interimResults &&
this.config.startDetection.enabled &&
this.config.startDetection.timeoutInMs
);
}

get silenceDetectionEnabled(): boolean {
return !!(
this.config.continuous &&
this.config.interimResults &&
this.config.silenceDetection.enabled &&
this.config.silenceDetection.timeoutInMs
);
}

static isSupported(): boolean {
return !!(window.SpeechRecognition || window.webkitSpeechRecognition);
return (
!!(window.SpeechRecognition || window.webkitSpeechRecognition) && BrowserDetector.isChrome()
);
}

static getDefaultConfig(): SpeechRecognizerConfig {
Expand All @@ -66,12 +94,9 @@ export class SpeechRecognizer extends EventEmitter {
}

readonly config: SpeechRecognizerConfig;

private readonly recognition: SpeechRecognition | null = null;

private recording = false;
private lastRecognitionEvent: SpeechRecognitionEvent | null = null;

private timeoutId?: number;
private ignoreNextEnd = false;

Expand All @@ -82,38 +107,12 @@ export class SpeechRecognizer extends EventEmitter {
const defaultConfig = SpeechRecognizer.getDefaultConfig();
this.config = config ? _defaultsDeep(config, defaultConfig) : defaultConfig;

if (window.SpeechRecognition) {
if (SpeechRecognizer.isSupported()) {
this.recognition = new window.SpeechRecognition();
this.setupSpeechRecognition(this.recognition);
}
}

get isRecording(): boolean {
return this.recording;
}

get isAvailable(): boolean {
return !!this.recognition;
}

get startDetectionEnabled(): boolean {
return !!(
this.config.continuous &&
this.config.interimResults &&
this.config.startDetection.enabled &&
this.config.startDetection.timeoutInMs
);
}

get silenceDetectionEnabled(): boolean {
return !!(
this.config.continuous &&
this.config.interimResults &&
this.config.silenceDetection.enabled &&
this.config.silenceDetection.timeoutInMs
);
}

addListener(
event: SpeechRecognizerEvent.SpeechRecognized,
listener: SpeechRecognizerSpeechRecognizedListener,
Expand Down
43 changes: 20 additions & 23 deletions jovo-clients/jovo-client-web/src/standalone/SpeechSynthesizer.ts
Original file line number Diff line number Diff line change
Expand Up @@ -25,29 +25,6 @@ export interface SpeechSynthesizerConfig {
}

export class SpeechSynthesizer extends EventEmitter {
static getDefaultConfig(): SpeechSynthesizerConfig {
return {
enabled: true,
language: 'en',
};
}

readonly config: SpeechSynthesizerConfig;

private $volume = 1.0;
private readonly synthesis: SpeechSynthesis | null;

private isSpeakingUtterance = false;

constructor(config?: DeepPartial<SpeechSynthesizerConfig>) {
super();

const defaultConfig = SpeechSynthesizer.getDefaultConfig();
this.config = config ? _defaultsDeep(config, defaultConfig) : defaultConfig;

this.synthesis = window.speechSynthesis || null;
}

get volume(): number {
return this.$volume;
}
Expand Down Expand Up @@ -76,6 +53,26 @@ export class SpeechSynthesizer extends EventEmitter {
return !!this.synthesis && this.synthesis.speaking;
}

static getDefaultConfig(): SpeechSynthesizerConfig {
return {
enabled: true,
language: 'en',
};
}
readonly config: SpeechSynthesizerConfig;
private $volume = 1.0;
private readonly synthesis: SpeechSynthesis | null;
private isSpeakingUtterance = false;

constructor(config?: DeepPartial<SpeechSynthesizerConfig>) {
super();

const defaultConfig = SpeechSynthesizer.getDefaultConfig();
this.config = config ? _defaultsDeep(config, defaultConfig) : defaultConfig;

this.synthesis = window.speechSynthesis || null;
}

addListener(event: SpeechSynthesizerVoidEvents, listener: VoidListener): this;
addListener(event: SpeechSynthesizerEvent.Speak, listener: SpeechSynthesizerSpeakListener): this;
addListener(event: SpeechSynthesizerEvent.Error, listener: ErrorListener): this;
Expand Down
Loading