ben.carrasco/jambonz-feature-server
1
0
Fork 0
mirror of https://github.com/jambonz/jambonz-feature-server.git synced 2025-12-20 16:50:39 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix config verb can override recognizer *Options (#558)

Browse Source 
* fix config verb can override recognizer *Options

* wip

* clean up gather and transcribe

* fix failing testcase

* fix failing testcase

* fix failing testcase
This commit is contained in:
Hoan Luu Huu
2023-12-06 08:27:55 +07:00
committed by GitHub
parent 46755f909c
commit 59248c7638
6 changed files with 158 additions and 110 deletions
Download Patch File Download Diff File Expand all files Collapse all files

22
lib/session/call-session.js
View File

@@ -170,6 +170,18 @@ class CallSession extends Emitter {
return this.application.notifier;
}
/**
* syntheizer
*/
get synthesizer() {
return this._synthesizer;
}
set synthesizer(synth) {
this._synthesizer = synth;
}
/**
* default vendor to use for speech synthesis if not provided in the app
*/
@@ -252,6 +264,16 @@ class CallSession extends Emitter {
set fallbackSpeechRecognizerVendor(vendor) {
this.application.fallback_speech_recognizer_vendor = vendor;
}
/**
* recognizer
*/
get recognizer() {
return this._recognizer;
}
set recognizer(rec) {
this._recognizer = rec;
}
/**
* default vendor to use for speech recognition if not provided in the app
*/

2
lib/tasks/config.js
View File

@@ -109,6 +109,7 @@ class TaskConfig extends Task {
});
if (this.hasSynthesizer) {
cs.synthesizer = this.synthesizer;
cs.speechSynthesisVendor = this.synthesizer.vendor !== 'default'
? this.synthesizer.vendor
: cs.speechSynthesisVendor;
@@ -138,6 +139,7 @@ class TaskConfig extends Task {
this.logger.info({synthesizer: this.synthesizer}, 'Config: updated synthesizer');
}
if (this.hasRecognizer) {
cs.recognizer = this.recognizer;
cs.speechRecognizerVendor = this.recognizer.vendor !== 'default'
? this.recognizer.vendor
: cs.speechRecognizerVendor;

55
lib/tasks/gather.js
View File

@@ -103,7 +103,7 @@ class TaskGather extends SttTask {
async exec(cs, {ep}) {
this.logger.debug({options: this.data}, 'Gather:exec');
await super.exec(cs);
await super.exec(cs, {ep});
const {updateSpeechCredentialLastUsed} = require('../utils/db-utils')(this.logger, cs.srf);
if (cs.hasGlobalSttHints && !this.maskGlobalSttHints) {
@@ -141,59 +141,6 @@ class TaskGather extends SttTask {
this.interim = true;
this.logger.debug('Gather:exec - early hints match enabled');
}
this.ep = ep;
if ('default' === this.vendor || !this.vendor) {
this.vendor = cs.speechRecognizerVendor;
if (this.data.recognizer) this.data.recognizer.vendor = this.vendor;
}
if ('default' === this.language || !this.language) {
this.language = cs.speechRecognizerLanguage;
if (this.data.recognizer) this.data.recognizer.language = this.language;
}
if ('default' === this.label || !this.label) {
this.label = cs.speechRecognizerLabel;
if (this.data.recognizer) this.data.recognizer.label = this.label;
}
// Fallback options
if ('default' === this.fallbackVendor || !this.fallbackVendor) {
this.fallbackVendor = cs.fallbackSpeechRecognizerVendor;
if (this.data.recognizer) this.data.recognizer.fallbackVendor = this.fallbackVendor;
}
if ('default' === this.fallbackLanguage || !this.fallbackLanguage) {
this.fallbackLanguage = cs.fallbackSpeechRecognizerLanguage;
if (this.data.recognizer) this.data.recognizer.fallbackLanguage = this.fallbackLanguage;
}
if ('default' === this.fallbackLabel || !this.fallbackLabel) {
this.fallbackLabel = cs.fallbackSpeechRecognizerLabel;
if (this.data.recognizer) this.data.recognizer.fallbackLabel = this.fallbackLabel;
}
if (!this.data.recognizer.vendor) {
this.data.recognizer.vendor = this.vendor;
}
if (this.vendor === 'cobalt' && !this.data.recognizer.model) {
// By default, application saves cobalt model in language
this.data.recognizer.model = cs.speechRecognizerLanguage;
}
if (this.needsStt && !this.sttCredentials) {
try {
this.sttCredentials = await this._initSpeechCredentials(cs, this.vendor, this.label);
} catch (error) {
if (this.fallbackVendor && this.isHandledByPrimaryProvider) {
await this._fallback();
} else {
throw error;
}
}
}
/* when using cobalt model is required */
if (this.vendor === 'cobalt' && !this.data.recognizer.model) {
this.notifyError({ msg: 'ASR error', details:'Cobalt requires a model to be specified'});
throw new Error('Cobalt requires a model to be specified');
}
const startListening = async(cs, ep) => {
this._startTimer();
if (this.isContinuousAsr && 0 === this.timeout) this._startAsrTimer();

75
lib/tasks/stt-task.js
View File

@@ -52,8 +52,81 @@ class SttTask extends Task {
}
async exec(cs, {ep, ep2}) {
super.exec(cs);
this.ep = ep;
this.ep2 = ep2;
if ('default' === this.vendor || !this.vendor) {
this.vendor = cs.speechRecognizerVendor;
if (this.data.recognizer) this.data.recognizer.vendor = this.vendor;
}
if ('default' === this.language || !this.language) {
this.language = cs.speechRecognizerLanguage;
if (this.data.recognizer) this.data.recognizer.language = this.language;
}
if ('default' === this.label || !this.label) {
this.label = cs.speechRecognizerLabel;
if (this.data.recognizer) this.data.recognizer.label = this.label;
}
// Fallback options
if ('default' === this.fallbackVendor || !this.fallbackVendor) {
this.fallbackVendor = cs.fallbackSpeechRecognizerVendor;
if (this.data.recognizer) this.data.recognizer.fallbackVendor = this.fallbackVendor;
}
if ('default' === this.fallbackLanguage || !this.fallbackLanguage) {
this.fallbackLanguage = cs.fallbackSpeechRecognizerLanguage;
if (this.data.recognizer) this.data.recognizer.fallbackLanguage = this.fallbackLanguage;
}
if ('default' === this.fallbackLabel || !this.fallbackLabel) {
this.fallbackLabel = cs.fallbackSpeechRecognizerLabel;
if (this.data.recognizer) this.data.recognizer.fallbackLabel = this.fallbackLabel;
}
if (!this.data.recognizer.vendor) {
this.data.recognizer.vendor = this.vendor;
}
if (this.vendor === 'cobalt' && !this.data.recognizer.model) {
// By default, application saves cobalt model in language
this.data.recognizer.model = cs.speechRecognizerLanguage;
}
if (cs.recognizer) {
for (const k in cs.recognizer) {
if (Array.isArray(this.data.recognizer[k]) ||
Array.isArray(cs.recognizer[k]) ||
typeof this.data.recognizer[k] === 'object' ||
typeof cs.recognizer[k] === 'object'
) {
this.data.recognizer[k] = {
...this.data.recognizer[k],
...cs.recognizer[k]
};
} else {
this.data.recognizer[k] = cs.recognizer[k] || this.data.recognizer[k];
}
}
}
if (!this.sttCredentials) {
try {
this.sttCredentials = await this._initSpeechCredentials(this.cs, this.vendor, this.label);
} catch (error) {
if (this.fallbackVendor && this.isHandledByPrimaryProvider) {
await this._fallback();
} else {
throw error;
}
}
}
/* when using cobalt model is required */
if (this.vendor === 'cobalt' && !this.data.recognizer.model) {
this.notifyError({ msg: 'ASR error', details:'Cobalt requires a model to be specified'});
throw new Error('Cobalt requires a model to be specified');
}
}
async _initSpeechCredentials(cs, vendor, label) {
const {getNuanceAccessToken, getIbmAccessToken} = this.cs.srf.locals.dbHelpers;
const {getNuanceAccessToken, getIbmAccessToken} = cs.srf.locals.dbHelpers;
let credentials = cs.getSpeechCredentials(vendor, 'stt', label);
if (!credentials) {

56
lib/tasks/transcribe.js
View File

@@ -45,7 +45,7 @@ class TaskTranscribe extends SttTask {
get name() { return TaskName.Transcribe; }
async exec(cs, {ep, ep2}) {
super.exec(cs);
await super.exec(cs, {ep, ep2});
const {updateSpeechCredentialLastUsed} = require('../utils/db-utils')(this.logger, cs.srf);
if (cs.hasGlobalSttHints) {
@@ -63,60 +63,6 @@ class TaskTranscribe extends SttTask {
if (cs.hasGlobalSttPunctuation && !this.data.recognizer.punctuation) {
this.data.recognizer.punctuation = cs.globalSttPunctuation;
}
this.ep = ep;
this.ep2 = ep2;
if ('default' === this.vendor || !this.vendor) {
this.vendor = cs.speechRecognizerVendor;
if (this.data.recognizer) this.data.recognizer.vendor = this.vendor;
}
if ('default' === this.language || !this.language) {
this.language = cs.speechRecognizerLanguage;
if (this.data.recognizer) this.data.recognizer.language = this.language;
}
if ('default' === this.label || !this.label) {
this.label = cs.speechRecognizerLabel;
if (this.data.recognizer) this.data.recognizer.label = this.label;
}
// Fallback options
if ('default' === this.fallbackVendor || !this.fallbackVendor) {
this.fallbackVendor = cs.fallbackSpeechRecognizerVendor;
if (this.data.recognizer) this.data.recognizer.fallbackVendor = this.fallbackVendor;
}
if ('default' === this.fallbackLanguage || !this.fallbackLanguage) {
this.fallbackLanguage = cs.fallbackSpeechRecognizerLanguage;
if (this.data.recognizer) this.data.recognizer.fallbackLanguage = this.fallbackLanguage;
}
if ('default' === this.fallbackLabel || !this.fallbackLabel) {
this.fallbackLabel = cs.fallbackSpeechRecognizerLabel;
if (this.data.recognizer) this.data.recognizer.fallbackLabel = this.fallbackLabel;
}
if (!this.data.recognizer.vendor) {
this.data.recognizer.vendor = this.vendor;
}
if (this.vendor === 'cobalt' && !this.data.recognizer.model) {
// By default, application saves cobalt model in language
this.data.recognizer.model = cs.speechRecognizerLanguage;
}
if (!this.sttCredentials) {
try {
this.sttCredentials = await this._initSpeechCredentials(cs, this.vendor, this.label);
} catch (error) {
if (this.fallbackVendor && this.isHandledByPrimaryProvider) {
await this._fallback();
} else {
throw error;
}
}
}
/* when using cobalt model is required */
if (this.vendor === 'cobalt' && !this.data.recognizer.model) {
this.notifyError({ msg: 'ASR error', details:'Cobalt requires a model to be specified'});
throw new Error('Cobalt requires a model to be specified');
}
try {
await this._startTranscribing(cs, ep, 1);
if (this.separateRecognitionPerChannel && ep2) {