0.2.3

rimelabs support multiple model and languages

lib/add-file-to-cache.js

@@ -25,7 +25,7 @@ function getExtensionAndSampleRate(path) {
 }
 
 async function addFileToCache(client, logger, path,
-  {account_sid, vendor, language, voice, deploymentId, engine, text}) {
+  {account_sid, vendor, language, voice, deploymentId, engine, model, text}) {
   let key;
   logger = logger || noopLogger;
 
@@ -36,6 +36,7 @@ async function addFileToCache(client, logger, path,
       language: language || '',
       voice: voice || deploymentId,
       engine,
+      model,
       text,
     });
     const [extension, sampleRate] = getExtensionAndSampleRate(path);

lib/purge-tts-cache.js

@@ -12,7 +12,7 @@ const debug = require('debug')('jambonz:realtimedb-helpers');
  * @returns {object} result - {error, purgedCount}
  */
 async function purgeTtsCache(client, logger, {all, account_sid, vendor,
-  language, voice, deploymentId, engine, text} = {all: true}) {
+  language, voice, deploymentId, engine, model, text} = {all: true}) {
   logger = logger || noopLogger;
 
   let purgedCount = 0, error;
@@ -33,6 +33,7 @@ async function purgeTtsCache(client, logger, {all, account_sid, vendor,
         language: language || '',
         voice: voice || deploymentId,
         engine,
+        model,
         text,
       });
       purgedCount = await client.del(key);

lib/synth-audio.js

@@ -169,6 +169,8 @@ async function synthAudio(client, createHash, retrieveHash, logger, stats, { acc
     language: language || '',
     voice: voice || deploymentId,
     engine,
+    // model or model_id is used to identify the tts cache.
+    model: model || credentials.model_id,
     text
   });
 
@@ -924,7 +926,7 @@ const synthPlayHT = async(client, logger, {
 };
 
 const synthRimelabs = async(logger, {
-  credentials, options, stats, voice, text, renderForCaching, disableTtsStreaming
+  credentials, options, stats, language, voice, text, renderForCaching, disableTtsStreaming
 }) => {
   const {api_key, model_id, options: credOpts} = credentials;
   const opts = !!options && Object.keys(options).length !== 0 ? options : JSON.parse(credOpts || '{}');
@@ -935,6 +937,7 @@ const synthRimelabs = async(logger, {
     params += `{api_key=${api_key}`;
     params += `,model_id=${model_id}`;
     params += ',vendor=rimelabs';
+    params += `,language=${language}`;
     params += `,voice=${voice}`;
     params += ',write_cache_file=1';
     if (opts.speedAlpha) params += `,speed_alpha=${opts.speedAlpha}`;
@@ -960,6 +963,7 @@ const synthRimelabs = async(logger, {
       text,
       modelId: model_id,
       samplingRate: sampleRate,
+      lang: language,
       ...opts
     });
     return {

lib/utils.js

@@ -17,9 +17,16 @@ const debug = require('debug')('jambonz:realtimedb-helpers');
 //const nuanceClientMap = new Map();
 
 function makeSynthKey({
-  account_sid = '', vendor, language, voice, engine = '', text}) {
+  account_sid = '',
+  vendor,
+  language,
+  voice,
+  engine = '',
+  model = '',
+  text
+}) {
   const hash = crypto.createHash('sha1');
-  hash.update(`${language}:${vendor}:${voice}:${engine}:${text}`);
+  hash.update(`${language}:${vendor}:${voice}:${engine}:${model}:${text}`);
   const hexHashKey = hash.digest('hex');
   const accountKey = account_sid ? `:${account_sid}` : '';
   const key = `tts${accountKey}:${hexHashKey}`;

package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "@jambonz/speech-utils",
-  "version": "0.2.1",
+  "version": "0.2.3",
   "lockfileVersion": 2,
   "requires": true,
   "packages": {
     "": {
       "name": "@jambonz/speech-utils",
-      "version": "0.2.1",
+      "version": "0.2.3",
       "license": "MIT",
       "dependencies": {
         "@aws-sdk/client-polly": "^3.496.0",

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@jambonz/speech-utils",
-  "version": "0.2.1",
+  "version": "0.2.3",
   "description": "TTS-related speech utilities for jambonz",
   "main": "index.js",
   "author": "Dave Horton",

test/synth.js

@@ -5,7 +5,7 @@ const fs = require('fs');
 const {makeSynthKey} = require('../lib/utils');
 const logger = require('pino')();
 const bent = require('bent');
-const getJSON = bent('json')
+const getJSON = bent('json');
 
 process.on('unhandledRejection', (reason, p) => {
   console.log('Unhandled Rejection at: Promise', p, 'reason:', reason);
@@ -720,7 +720,7 @@ test('Cartesia speech synth tests', async(t) => {
   client.quit();
 });
 
-test('rimelabs speech synth tests', async(t) => {
+test('rimelabs speech synth tests mist', async(t) => {
   const fn = require('..');
   const {synthAudio, client} = fn(opts, logger);
 
@@ -730,7 +730,7 @@ test('rimelabs speech synth tests', async(t) => {
   }
   const text = 'Hi there and welcome to jambones!';
   try {
-    let opts = await synthAudio(stats, {
+    const opts = await synthAudio(stats, {
       vendor: 'rimelabs',
       credentials: {
         api_key: process.env.RIMELABS_API_KEY,
@@ -740,7 +740,7 @@ test('rimelabs speech synth tests', async(t) => {
           reduceLatency: false
         })
       },
-      language: 'en-US',
+      language: 'eng',
       voice: 'amber',
       text,
       renderForCaching: true
@@ -754,6 +754,40 @@ test('rimelabs speech synth tests', async(t) => {
   client.quit();
 });
 
+test('rimelabs speech synth tests mistv2', async(t) => {
+  const fn = require('..');
+  const {synthAudio, client} = fn(opts, logger);
+
+  if (!process.env.RIMELABS_API_KEY) {
+    t.pass('skipping rimelabs speech synth tests since RIMELABS_API_KEY is not provided');
+    return t.end();
+  }
+  const text = 'Hi there and welcome to jambones!';
+  try {
+    const opts = await synthAudio(stats, {
+      vendor: 'rimelabs',
+      credentials: {
+        api_key: process.env.RIMELABS_API_KEY,
+        model_id: 'mistv2',
+        options: JSON.stringify({
+          speedAlpha: 1.0,
+          reduceLatency: false
+        })
+      },
+      language: 'spa',
+      voice: 'pablo',
+      text,
+      renderForCaching: true
+    });
+    t.ok(!opts.servedFromCache, `successfully synthesized rimelabs mistv2 audio to ${opts.filePath}`);
+
+  } catch (err) {
+    console.error(JSON.stringify(err));
+    t.end(err);
+  }
+  client.quit();
+});
+
 test('whisper speech synth tests', async(t) => {
   const fn = require('..');
   const {synthAudio, client} = fn(opts, logger);
@@ -853,11 +887,11 @@ test('TTS Cache tests', async(t) => {
     // save some random tts keys to cache
     const minRecords = 8;
     for (const i in Array(minRecords).fill(0)) {
-      await client.set(makeSynthKey({vendor: i, language: i, voice: i, engine: i, text: i}), i);
+      await client.set(makeSynthKey({vendor: i, language: i, voice: i, engine: i, model: i, text: i}), i);
     }
     const count = await getTtsSize();
     t.ok(count >= minRecords, 'getTtsSize worked.');
-    
+
     const {purgedCount} = await purgeTtsCache();
     t.ok(purgedCount >= minRecords, `successfully purged at least ${minRecords} tts records from cache`);