mirror of
https://github.com/jambonz/jambonz-api-server.git
synced 2026-07-24 04:52:03 +00:00
TTS/STT languages and voices for each provider (#281)
* wip * wip * wip * wip * wip * wip * add testcase
This commit is contained in:
Hoan Luu Huu
+191
-1
@@ -8,6 +8,30 @@ const fs = require('fs');
|
||||
const { AssemblyAI } = require('assemblyai');
|
||||
const {decrypt, obscureKey} = require('./encrypt-decrypt');
|
||||
|
||||
const TtsGoogleLanguagesVoices = require('./speech-data/tts-google');
|
||||
const TtsAwsLanguagesVoices = require('./speech-data/tts-aws');
|
||||
const TtsMicrosoftLanguagesVoices = require('./speech-data/tts-microsoft');
|
||||
const TtsWellsaidLanguagesVoices = require('./speech-data/tts-wellsaid');
|
||||
const TtsNuanceLanguagesVoices = require('./speech-data/tts-nuance');
|
||||
const TtsIbmLanguagesVoices = require('./speech-data/tts-ibm');
|
||||
const TtsNvidiaLanguagesVoices = require('./speech-data/tts-nvidia');
|
||||
const TtsElevenlabsLanguagesVoices = require('./speech-data/tts-elevenlabs');
|
||||
const TtsWhisperLanguagesVoices = require('./speech-data/tts-whisper');
|
||||
|
||||
const TtsModelDeepgram = require('./speech-data/tts-model-deepgram');
|
||||
const TtsModelElevenLabs = require('./speech-data/tts-model-elevenlabs');
|
||||
const TtsModelWhisper = require('./speech-data/tts-model-whisper');
|
||||
|
||||
const SttGoogleLanguagesVoices = require('./speech-data/stt-google');
|
||||
const SttAwsLanguagesVoices = require('./speech-data/stt-aws');
|
||||
const SttMicrosoftLanguagesVoices = require('./speech-data/stt-microsoft');
|
||||
const SttNuanceLanguagesVoices = require('./speech-data/stt-nuance');
|
||||
const SttDeepgramLanguagesVoices = require('./speech-data/stt-deepgram');
|
||||
const SttIbmLanguagesVoices = require('./speech-data/stt-ibm');
|
||||
const SttNvidiaLanguagesVoices = require('./speech-data/stt-nvidia');
|
||||
const SttCobaltLanguagesVoices = require('./speech-data/stt-cobalt');
|
||||
const SttSonioxLanguagesVoices = require('./speech-data/stt-soniox');
|
||||
const SttAssemblyaiLanguagesVoices = require('./speech-data/stt-assemblyai');
|
||||
|
||||
const testSonioxStt = async(logger, credentials) => {
|
||||
const api_key = credentials;
|
||||
@@ -419,6 +443,171 @@ function decryptCredential(obj, credential, logger, isObscureKey = true) {
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
*
|
||||
* @param {*} logger logger
|
||||
* @param {*} vendor vendor
|
||||
* @param {*} credential STT/TTS vendor credential, can be null
|
||||
* @returns List of language and coresponding voices for specific vendor follow below format
|
||||
* {
|
||||
"tts": [
|
||||
{
|
||||
code: "ar-XA",
|
||||
name: "Arabic",
|
||||
voices: [
|
||||
{ value: "ar-XA-Standard-A", name: "Standard-A (Female)" },
|
||||
]
|
||||
}
|
||||
],
|
||||
"stt": [
|
||||
{ name: "Afrikaans (South Africa)", code: "af-ZA" },
|
||||
]
|
||||
}
|
||||
*/
|
||||
async function getLanguagesAndVoicesForVendor(logger, vendor, credential) {
|
||||
switch (vendor) {
|
||||
case 'google':
|
||||
return await getLanguagesVoicesForGoogle(credential);
|
||||
case 'aws':
|
||||
return await getLanguagesVoicesForAws(credential);
|
||||
case 'microsoft':
|
||||
return await getLanguagesVoicesForMicrosoft(credential);
|
||||
case 'wellsaid':
|
||||
return await getLanguagesVoicesForWellsaid(credential);
|
||||
case 'nuance':
|
||||
return await getLanguagesVoicesForNuane(credential);
|
||||
case 'deepgram':
|
||||
return await getLanguagesVoicesForDeepgram(credential);
|
||||
case 'ibm':
|
||||
return await getLanguagesVoicesForIbm(credential);
|
||||
case 'nvidia':
|
||||
return await getLanguagesVoicesForNvida(credential);
|
||||
case 'cobalt':
|
||||
return await getLanguagesVoicesForCobalt(credential);
|
||||
case 'soniox':
|
||||
return await getLanguagesVoicesForSoniox(credential);
|
||||
case 'elevenlabs':
|
||||
return await getLanguagesVoicesForElevenlabs(credential);
|
||||
case 'assemblyai':
|
||||
return await getLanguagesVoicesForAssemblyAI(credential);
|
||||
case 'whisper':
|
||||
return await getLanguagesVoicesForWhisper(credential);
|
||||
default:
|
||||
logger.info(`invalid vendor ${vendor}, return empty result`);
|
||||
throw new Error(`Invalid vendor ${vendor}`);
|
||||
}
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForGoogle(credential) {
|
||||
return tranform(TtsGoogleLanguagesVoices, SttGoogleLanguagesVoices);
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForAws(credential) {
|
||||
return tranform(TtsAwsLanguagesVoices, SttAwsLanguagesVoices);
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForMicrosoft(credential) {
|
||||
return tranform(TtsMicrosoftLanguagesVoices, SttMicrosoftLanguagesVoices);
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForWellsaid(credential) {
|
||||
return tranform(TtsWellsaidLanguagesVoices);
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForNuane(credential) {
|
||||
return tranform(TtsNuanceLanguagesVoices, SttNuanceLanguagesVoices);
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForDeepgram(credential) {
|
||||
return tranform(undefined, SttDeepgramLanguagesVoices, TtsModelDeepgram);
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForIbm(credential) {
|
||||
return tranform(TtsIbmLanguagesVoices, SttIbmLanguagesVoices);
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForNvida(credential) {
|
||||
return tranform(TtsNvidiaLanguagesVoices, SttNvidiaLanguagesVoices);
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForCobalt(credential) {
|
||||
return tranform(undefined, SttCobaltLanguagesVoices);
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForSoniox(credential) {
|
||||
if (credential) {
|
||||
|
||||
} else {
|
||||
return tranform(undefined, SttSonioxLanguagesVoices);
|
||||
}
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForElevenlabs(credential) {
|
||||
if (credential) {
|
||||
const get = bent('https://api.elevenlabs.io', 'GET', 'json', {
|
||||
'xi-api-key' : credential.api_key
|
||||
});
|
||||
|
||||
const [langResp, voiceResp] = await Promise.all([get('/v1/models'), get('/v1/voices')]);
|
||||
|
||||
const model = langResp.find((m) => m.model_id === credential.model_id);
|
||||
const languages = model ? model.languages.map((l) => {
|
||||
return {
|
||||
value: l.language_id,
|
||||
name: l.name
|
||||
};
|
||||
}).sort((a, b) => a.name.localeCompare(b.name)) : [];
|
||||
|
||||
if (languages && languages.length > 0) {
|
||||
const voices = voiceResp ? voiceResp.voices.map((v) => {
|
||||
let name = `${v.name}${v.category !== 'premade' ? ` (${v.category})` : ''} -
|
||||
${v.labels.accent ? ` ${v.labels.accent},` : ''}
|
||||
${v.labels.description ? ` ${v.labels.description},` : ''}
|
||||
${v.labels.age ? ` ${v.labels.age},` : ''}
|
||||
${v.labels.gender ? ` ${v.labels.gender},` : ''}
|
||||
${v.labels['use case'] ? ` ${v.labels['use case']},` : ''}
|
||||
`;
|
||||
const lastIndex = name.lastIndexOf(',');
|
||||
if (lastIndex !== -1) {
|
||||
name = name.substring(0, lastIndex);
|
||||
}
|
||||
return {
|
||||
value: v.voice_id,
|
||||
name
|
||||
};
|
||||
}).sort((a, b) => a.name.localeCompare(b.name)) : [];
|
||||
languages[0].voices = voices;
|
||||
}
|
||||
return tranform(languages, undefined, TtsModelElevenLabs);
|
||||
} else {
|
||||
return tranform(TtsElevenlabsLanguagesVoices, undefined, TtsModelElevenLabs);
|
||||
}
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForAssemblyAI(credential) {
|
||||
if (credential) {
|
||||
|
||||
} else {
|
||||
return tranform(undefined, SttAssemblyaiLanguagesVoices);
|
||||
}
|
||||
}
|
||||
|
||||
async function getLanguagesVoicesForWhisper(credential) {
|
||||
if (credential) {
|
||||
|
||||
} else {
|
||||
return tranform(TtsWhisperLanguagesVoices, undefined, TtsModelWhisper);
|
||||
}
|
||||
}
|
||||
|
||||
function tranform(tts, stt, models) {
|
||||
return {
|
||||
...(tts && {tts}),
|
||||
...(stt && {stt}),
|
||||
...(models && {models})
|
||||
};
|
||||
}
|
||||
|
||||
module.exports = {
|
||||
testGoogleTts,
|
||||
testGoogleStt,
|
||||
@@ -439,5 +628,6 @@ module.exports = {
|
||||
testDeepgramTTS,
|
||||
getSpeechCredential,
|
||||
decryptCredential,
|
||||
testWhisper
|
||||
testWhisper,
|
||||
getLanguagesAndVoicesForVendor
|
||||
};
|
||||
|
||||
Reference in New Issue
Block a user