0.0.34

Elevenlabs streaming

.github/workflows/ci.yml

@@ -14,6 +14,9 @@ jobs:
           node-version: lts/*
       - run: npm install
       - run: npm run jslint
+      - run: sudo apt update && sudo apt install -y squid
+      - run: sudo cp test/squid.conf /etc/squid/squid.conf
+      - run: sudo systemctl start squid
       - run: npm test
         env:
           AWS_ACCESS_KEY_ID: ${{ secrets.AWS_ACCESS_KEY_ID }}
@@ -29,3 +32,5 @@ jobs:
           ELEVENLABS_API_KEY: ${{ secrets.ELEVENLABS_API_KEY }}
           ELEVENLABS_VOICE_ID: ${{ secrets.ELEVENLABS_VOICE_ID }}
           ELEVENLABS_MODEL_ID: ${{ secrets.ELEVENLABS_MODEL_ID }}
+          JAMBONES_HTTP_PROXY_IP: 127.0.0.1
+          JAMBONES_HTTP_PROXY_PORT: 3128

README.md

@@ -1,3 +1,3 @@
-# speech-utils
+# speech-utils ![CI](https://github.com/jambonz/speech-utils/workflows/CI/badge.svg)
 TTS-related speech utilities for jambonz.
 

index.js

@@ -2,13 +2,11 @@ const {noopLogger} = require('./lib/utils');
 
 module.exports = (opts, logger) => {
   logger = logger || noopLogger;
-  let client = opts.redis_client;
   const {
-    client: redisClient,
+    client,
     createHash,
     retrieveHash
   } = require('@jambonz/realtimedb-helpers')(opts, logger);
-  client = opts.redis_client || redisClient;
 
   return {
     client,

lib/get-aws-sts-token.js

@@ -27,7 +27,7 @@ async function getAwsAuthToken(
     const credentials = {
       accessKeyId: data.Credentials.AccessKeyId,
       secretAccessKey: data.Credentials.SecretAccessKey,
-      sessionToken: data.Credentials.SessionToken
+      securityToken: data.Credentials.SessionToken
     };
 
     /* expire 10 minutes before the hour, so we don't lose the use of it during a call */

lib/synth-audio.js

@@ -76,7 +76,8 @@ const trimTrailingSilence = (buffer) => {
  * the synthesized audio, and a variable indicating whether it was served from cache
  */
 async function synthAudio(client, logger, stats, { account_sid,
-  vendor, language, voice, gender, text, engine, salt, model, credentials, deploymentId, disableTtsCache, options
+  vendor, language, voice, gender, text, engine, salt, model, credentials, deploymentId,
+  disableTtsCache, renderForCaching, options
 }) {
   let audioBuffer;
   let servedFromCache = false;
@@ -84,7 +85,7 @@ async function synthAudio(client, logger, stats, { account_sid,
   logger = logger || noopLogger;
 
   assert.ok(['google', 'aws', 'polly', 'microsoft',
-    'wellsaid', 'nuance', 'nvidia', 'ibm', 'elevenlabs', 'whisper'].includes(vendor) ||
+    'wellsaid', 'nuance', 'nvidia', 'ibm', 'elevenlabs', 'whisper', 'deepgram'].includes(vendor) ||
   vendor.startsWith('custom'),
   `synthAudio supported vendors are google, aws, microsoft, nuance, nvidia and wellsaid, not ${vendor}`);
   if ('google' === vendor) {
@@ -194,11 +195,22 @@ async function synthAudio(client, logger, stats, { account_sid,
         audioBuffer = await synthWellSaid(logger, {credentials, stats, language, voice, text, filePath});
         break;
       case 'elevenlabs':
-        audioBuffer = await synthElevenlabs(logger, {credentials, options, stats, language, voice, text, filePath});
+        audioBuffer = await synthElevenlabs(logger, {
+          credentials, options, stats, language, voice, text, renderForCaching, filePath
+        });
+        if (typeof audioBuffer === 'object' && audioBuffer.filePath) {
+          return audioBuffer;
+        }
+        else {
+          audioBuffer = await synthElevenlabs(logger, {credentials, options, stats, language, voice, text, filePath});
+        }
         break;
       case 'whisper':
         audioBuffer = await synthWhisper(logger, {credentials, stats, voice, text});
         break;
+      case 'deepgram':
+        audioBuffer = await synthDeepgram(logger, {credentials, stats, model, text});
+        break;
       case vendor.startsWith('custom') ? vendor : 'cant_match_value':
         ({ audioBuffer, filePath } = await synthCustomVendor(logger,
           {credentials, stats, language, voice, text, filePath}));
@@ -400,6 +412,12 @@ const synthMicrosoft = async(logger, {
     speechConfig.speechSynthesisOutputFormat = trimSilence ?
       SpeechSynthesisOutputFormat.Raw8Khz16BitMonoPcm :
       SpeechSynthesisOutputFormat.Audio16Khz32KBitRateMonoMp3;
+
+    if (process.env.JAMBONES_HTTP_PROXY_IP && process.env.JAMBONES_HTTP_PROXY_PORT) {
+      logger.debug(
+        `synthMicrosoft: using proxy ${process.env.JAMBONES_HTTP_PROXY_IP}:${process.env.JAMBONES_HTTP_PROXY_PORT}`);
+      speechConfig.setProxy(process.env.JAMBONES_HTTP_PROXY_IP, process.env.JAMBONES_HTTP_PROXY_PORT);
+    }
     const synthesizer = new SpeechSynthesizer(speechConfig);
 
     if (content.startsWith('<speak>')) {
@@ -585,9 +603,29 @@ const synthCustomVendor = async(logger, {credentials, stats, language, voice, te
   }
 };
 
-const synthElevenlabs = async(logger, {credentials, options, stats, language, voice, text}) => {
+const synthElevenlabs = async(logger, {credentials, options, stats, language, voice, text, renderForCaching}) => {
   const {api_key, model_id, options: credOpts} = credentials;
   const opts = !!options && Object.keys(options).length !== 0 ? options : JSON.parse(credOpts || '{}');
+
+  /* if the env is set to stream then bag out, unless we are specifically rendering to generate a cache file */
+  if (process.env.JAMBONES_ELEVENLABS_STREAMING && !renderForCaching) {
+    let params = '';
+    params += `{api_key=${api_key}`;
+    params += `,model_id=${model_id}`;
+    params += `,optimize_streaming_latency=${opts.optimize_streaming_latency || 2}`;
+    if (opts.voice_settings?.similarity_boost) params += `,similarity_boost=${opts.voice_settings.similarity_boost}`;
+    if (opts.voice_settings?.stability) params += `,stability=${opts.voice_settings.stability}`;
+    if (opts.voice_settings?.style) params += `,style=${opts.voice_settings.style}`;
+    if (opts.voice_settings?.use_speaker_boost === false) params += ',use_speaker_boost=false';
+    params += '}';
+
+    return {
+      filePath: `say:${params}${text.replace(/\n/g, ' ')}`,
+      servedFromCache: false,
+      rtt: 0
+    };
+  }
+
   const optimize_streaming_latency = opts.optimize_streaming_latency ?
     `?optimize_streaming_latency=${opts.optimize_streaming_latency}` : '';
   try {
@@ -634,8 +672,27 @@ const synthWhisper = async(logger, {credentials, stats, voice, text}) => {
     stats.increment('tts.count', ['vendor:openai', 'accepted:no']);
     throw err;
   }
-}
-;
+};
+
+const synthDeepgram = async(logger, {credentials, stats, model, text}) => {
+  const {api_key} = credentials;
+  try {
+    const post = bent('https://api.beta.deepgram.com', 'POST', 'buffer', {
+      'Authorization': `Token ${api_key}`,
+      'Accept': 'audio/mpeg',
+      'Content-Type': 'application/json'
+    });
+    const mp3 = await post(`/v1/speak?model=${model}`, {
+      text
+    });
+    return mp3;
+  } catch (err) {
+    logger.info({err}, 'synth Deepgram returned error');
+    stats.increment('tts.count', ['vendor:deepgram', 'accepted:no']);
+    throw err;
+  }
+};
+
 const getFileExtFromMime = (mime) => {
   switch (mime) {
     case 'audio/wav':

package.json

@@ -1,6 +1,6 @@
 {
   "name": "@jambonz/speech-utils",
-  "version": "0.0.29",
+  "version": "0.0.34",
   "description": "TTS-related speech utilities for jambonz",
   "main": "index.js",
   "author": "Dave Horton",
@@ -24,26 +24,26 @@
   },
   "homepage": "https://github.com/jambonz/speech-utils#readme",
   "dependencies": {
-    "@aws-sdk/client-polly": "^3.359.0",
-    "@aws-sdk/client-sts": "^3.458.0",
-    "@google-cloud/text-to-speech": "^4.2.1",
-    "@grpc/grpc-js": "^1.8.13",
+    "@aws-sdk/client-polly": "^3.496.0",
+    "@aws-sdk/client-sts": "^3.496.0",
+    "@google-cloud/text-to-speech": "^5.0.2",
+    "@grpc/grpc-js": "^1.9.14",
     "@jambonz/realtimedb-helpers": "^0.8.7",
     "bent": "^7.3.12",
     "debug": "^4.3.4",
-    "form-urlencoded": "^6.1.0",
+    "form-urlencoded": "^6.1.4",
     "google-protobuf": "^3.21.2",
     "ibm-watson": "^8.0.0",
-    "microsoft-cognitiveservices-speech-sdk": "1.32.0",
-    "openai": "^4.16.2",
-    "undici": "^5.21.0"
+    "microsoft-cognitiveservices-speech-sdk": "1.34.0",
+    "openai": "^4.25.0",
+    "undici": "^6.4.0"
   },
   "devDependencies": {
-    "config": "^3.3.9",
-    "eslint": "^8.33.0",
+    "config": "^3.3.10",
+    "eslint": "^8.56.0",
     "eslint-plugin-promise": "^6.1.1",
     "nyc": "^15.1.0",
-    "pino": "^7.2.0",
-    "tape": "^5.1.1"
+    "pino": "^8.17.0",
+    "tape": "^5.7.3"
   }
 }

test/aws.js

@@ -21,12 +21,12 @@ test('AWS - create and cache auth token', async(t) => {
   try {
     let obj = await getAwsAuthToken(process.env.AWS_ACCESS_KEY_ID, process.env.AWS_SECRET_ACCESS_KEY, process.env.AWS_REGION);
     //console.log({obj}, 'received auth token from AWS');
-    t.ok(obj.sessionToken && !obj.servedFromCache, 'successfullY generated auth token from AWS');
+    t.ok(obj.securityToken && !obj.servedFromCache, 'successfullY generated auth token from AWS');
 
     await sleep(250);
     obj = await getAwsAuthToken(process.env.AWS_ACCESS_KEY_ID, process.env.AWS_SECRET_ACCESS_KEY, process.env.AWS_REGION);
     //console.log({obj}, 'received auth token from AWS - second request');
-    t.ok(obj.sessionToken && obj.servedFromCache, 'successfully received access token from cache');
+    t.ok(obj.securityToken && obj.servedFromCache, 'successfully received access token from cache');
  
     await client.flushall();
     t.end();

test/squid.conf

@@ -0,0 +1,85 @@
+#
+# Recommended minimum configuration:
+#
+
+# Example rule allowing access from your local networks.
+# Adapt to list your (internal) IP networks from where browsing
+# should be allowed
+acl localnet src 0.0.0.1-0.255.255.255	# RFC 1122 "this" network (LAN)
+acl localnet src 10.0.0.0/8		# RFC 1918 local private network (LAN)
+acl localnet src 100.64.0.0/10		# RFC 6598 shared address space (CGN)
+acl localnet src 169.254.0.0/16 	# RFC 3927 link-local (directly plugged) machines
+acl localnet src 172.16.0.0/12		# RFC 1918 local private network (LAN)
+acl localnet src 192.168.0.0/16		# RFC 1918 local private network (LAN)
+acl localnet src fc00::/7       	# RFC 4193 local private network range
+acl localnet src fe80::/10      	# RFC 4291 link-local (directly plugged) machines
+
+acl SSL_ports port 443
+acl Safe_ports port 80		# http
+acl Safe_ports port 21		# ftp
+acl Safe_ports port 443		# https
+acl Safe_ports port 70		# gopher
+acl Safe_ports port 210		# wais
+acl Safe_ports port 1025-65535	# unregistered ports
+acl Safe_ports port 280		# http-mgmt
+acl Safe_ports port 488		# gss-http
+acl Safe_ports port 591		# filemaker
+acl Safe_ports port 777		# multiling http
+
+#
+# Recommended minimum Access Permission configuration:
+#
+# Deny requests to certain unsafe ports
+http_access deny !Safe_ports
+
+# Deny CONNECT to other than secure SSL ports
+http_access allow CONNECT !SSL_ports
+
+# Only allow cachemgr access from localhost
+http_access allow localhost manager
+http_access deny manager
+
+# This default configuration only allows localhost requests because a more
+# permissive Squid installation could introduce new attack vectors into the
+# network by proxying external TCP connections to unprotected services.
+http_access allow localhost
+
+# The two deny rules below are unnecessary in this default configuration
+# because they are followed by a "deny all" rule. However, they may become
+# critically important when you start allowing external requests below them.
+
+# Protect web applications running on the same server as Squid. They often
+# assume that only local users can access them at "localhost" ports.
+http_access deny to_localhost
+
+# Protect cloud servers that provide local users with sensitive info about
+# their server via certain well-known link-local (a.k.a. APIPA) addresses.
+http_access deny to_linklocal
+
+#
+# INSERT YOUR OWN RULE(S) HERE TO ALLOW ACCESS FROM YOUR CLIENTS
+#
+
+# For example, to allow access from your local networks, you may uncomment the
+# following rule (and/or add rules that match your definition of "local"):
+# http_access allow localnet
+
+# And finally deny all other access to this proxy
+http_access deny all
+
+# Squid normally listens to port 3128
+http_port 3128
+
+# Uncomment and adjust the following to add a disk cache directory.
+#cache_dir ufs /usr/local/var/cache/squid 100 16 256
+
+# Leave coredumps in the first cache dir
+coredump_dir /usr/local/var/cache/squid
+
+#
+# Add any of your own refresh_pattern entries above these.
+#
+refresh_pattern ^ftp:		1440	20%	10080
+refresh_pattern ^gopher:	1440	0%	1440
+refresh_pattern -i (/cgi-bin/|\?) 0	0%	0
+refresh_pattern .		0	20%	4320

test/synth.js

@@ -182,7 +182,9 @@ test('Azure speech synth tests', async(t) => {
       text: longText,
     });
     t.ok(!opts.servedFromCache, `successfully synthesized microsoft audio to ${opts.filePath}`);
-
+    if (process.env.JAMBONES_HTTP_PROXY_IP && process.env.JAMBONES_HTTP_PROXY_PORT) {
+      t.pass('successfully used proxy to reach microsoft tts service');
+    }
 
     opts = await synthAudio(stats, {
       vendor: 'microsoft',
@@ -512,6 +514,33 @@ test('whisper speech synth tests', async(t) => {
   client.quit();
 })
 
+test('Deepgram speech synth tests', async(t) => {
+  const fn = require('..');
+  const {synthAudio, client} = fn(opts, logger);
+
+  if (!process.env.DEEPGRAM_API_KEY) {
+    t.pass('skipping Deepgram speech synth tests since DEEPGRAM_API_KEY');
+    return t.end();
+  }
+  const text = 'Hi there and welcome to jambones!';
+  try {
+    let opts = await synthAudio(stats, {
+      vendor: 'deepgram',
+      credentials: {
+        api_key: process.env.DEEPGRAM_API_KEY
+      },
+      model: 'alpha-asteria-en-v2',
+      text,
+    });
+    t.ok(!opts.servedFromCache, `successfully synthesized deepgram audio to ${opts.filePath}`);
+
+  } catch (err) {
+    console.error(JSON.stringify(err));
+    t.end(err);
+  }
+  client.quit();
+})
+
 test('TTS Cache tests', async(t) => {
   const fn = require('..');
   const {purgeTtsCache, getTtsSize, client} = fn(opts, logger);