try 5

The following vulnerabilities are fixed with an upgrade:
- https://snyk.io/vuln/SNYK-JS-XML2JS-5414874

.github/workflows/docker-publish.yml

@@ -2,13 +2,8 @@ name: Docker
 
 on:
   push:
-    # Publish `main` as Docker `latest` image.
-    branches:
-      - main
-
-    # Publish `v1.2.3` tags as releases.
     tags:
-      - v*
+      - '*'
 
 env:
   IMAGE_NAME: feature-server
@@ -20,32 +15,41 @@ jobs:
     if: github.event_name == 'push'
 
     steps:
-      - uses: actions/checkout@v3
+      - name: Checkout code
+        uses: actions/checkout@v3
 
-      - name: Build image
-        run: docker build . --file Dockerfile --tag $IMAGE_NAME
-
-      - name: Log into registry
-        run: echo "${{ secrets.GITHUB_TOKEN }}" | docker login ghcr.io -u ${{ github.actor }} --password-stdin
-
-      - name: Push image
+      - name: prepare tag 
+        id: prepare_tag
         run: |
-          IMAGE_ID=ghcr.io/${{ github.repository_owner }}/$IMAGE_NAME
+            IMAGE_ID=jambonz/$IMAGE_NAME
 
-          # Change all uppercase to lowercase
-          IMAGE_ID=$(echo $IMAGE_ID | tr '[A-Z]' '[a-z]')
+            # Strip git ref prefix from version
+            VERSION=$(echo "${{ github.ref }}" | sed -e 's,.*/\(.*\),\1,')
 
-          # Strip git ref prefix from version
-          VERSION=$(echo "${{ github.ref }}" | sed -e 's,.*/\(.*\),\1,')
+            # Strip "v" prefix from tag name
+            [[ "${{ github.ref }}" == "refs/tags/"* ]] && VERSION=$(echo $VERSION | sed -e 's/^v//')
 
-          # Strip "v" prefix from tag name
-          [[ "${{ github.ref }}" == "refs/tags/"* ]] && VERSION=$(echo $VERSION | sed -e 's/^v//')
+            # Use Docker `latest` tag convention
+            [ "$VERSION" == "main" ] && VERSION=latest
 
-          # Use Docker `latest` tag convention
-          [ "$VERSION" == "main" ] && VERSION=latest
+            echo IMAGE_ID=$IMAGE_ID
+            echo VERSION=$VERSION
 
-          echo IMAGE_ID=$IMAGE_ID
-          echo VERSION=$VERSION
+            echo "::set-output name=image_id::$IMAGE_ID"
+            echo "::set-output name=version::$VERSION"
 
-          docker tag $IMAGE_NAME $IMAGE_ID:$VERSION
-          docker push $IMAGE_ID:$VERSION
+      - name: Login to Docker Hub 
+        uses: docker/login-action@v2
+        with:
+          username: ${{ secrets.DOCKERHUB_USERNAME }}
+          password: ${{ secrets.DOCKERHUB_TOKEN }}
+
+      - name: Build and push Docker image
+        uses: docker/build-push-action@v4
+        with:
+          context: .
+          push: true
+          tags: ${{ steps.prepare_tag.outputs.image_id }}:${{ steps.prepare_tag.outputs.version }}
+          build-args: |
+            GITHUB_REPOSITORY=$GITHUB_REPOSITORY
+            GITHUB_REF=$GITHUB_REF

README.md

@@ -18,6 +18,7 @@ Configuration is provided via environment variables:
 |DRACHTIO_PORT| listening port of drachtio server for control connections (typically 9022)|yes|
 |DRACHTIO_SECRET| shared secret|yes|
 |ENABLE_METRICS| if 1, metrics will be generated|no|
+|ENCRYPTION_SECRET| secret for credential encryption(JWT_SECRET is deprecated) |yes|
 |GOOGLE_APPLICATION_CREDENTIALS| path to gcp service key file|yes|
 |HTTP_PORT| tcp port to listen on for API requests from jambonz-api-server|yes|
 |JAMBONES_GATHER_EARLY_HINTS_MATCH| if true and hints are provided, gather will opportunistically review interim transcripts if possible to reduce ASR latency |no|

app.js

@@ -8,6 +8,7 @@ assert.ok(process.env.DRACHTIO_SECRET, 'missing DRACHTIO_SECRET env var');
 assert.ok(process.env.JAMBONES_FREESWITCH, 'missing JAMBONES_FREESWITCH env var');
 assert.ok(process.env.JAMBONES_REDIS_HOST, 'missing JAMBONES_REDIS_HOST env var');
 assert.ok(process.env.JAMBONES_NETWORK_CIDR || process.env.K8S, 'missing JAMBONES_SUBNET env var');
+assert.ok(process.env.ENCRYPTION_SECRET || process.env.JWT_SECRET, 'missing ENCRYPTION_SECRET env var');
 
 const Srf = require('drachtio-srf');
 const srf = new Srf();

lib/http-routes/api/create-call.js

@@ -104,7 +104,7 @@ router.post('/', async(req, res) => {
       proxy: `sip:${sbcAddress}`,
       localSdp: ep.local.sdp
     });
-    if (target.auth) opts.auth = this.target.auth;
+    if (target.auth) opts.auth = target.auth;
 
 
     /**

lib/middleware.js

@@ -27,7 +27,11 @@ module.exports = function(srf, logger) {
 
   function initLocals(req, res, next) {
     const callId = req.get('Call-ID');
-    logger.info({callId}, 'new incoming call');
+    logger.info({
+      callId,
+      callingNumber: req.callingNumber,
+      calledNumber: req.calledNumber
+    }, 'new incoming call');
     if (!req.has('X-Account-Sid')) {
       logger.info('getAccountDetails - rejecting call due to missing X-Account-Sid header');
       return res.send(500);
@@ -42,7 +46,16 @@ module.exports = function(srf, logger) {
     }
     if (req.has('X-Authenticated-User')) req.locals.originatingUser = req.get('X-Authenticated-User');
     if (req.has('X-MS-Teams-Tenant-FQDN')) req.locals.msTeamsTenant = req.get('X-MS-Teams-Tenant-FQDN');
-
+    if (req.has('X-Cisco-Recording-Participant')) {
+      const ciscoParticipants = req.get('X-Cisco-Recording-Participant');
+      const regex = /sip:[\d]+@[\d]+\.[\d]+\.[\d]+\.[\d]+/g;
+      const sipURIs = ciscoParticipants.match(regex);
+      logger.info(`X-Cisco-Recording-Participant : ${sipURIs} `);
+      if (sipURIs && sipURIs.length > 0) {
+        req.locals.calledNumber = sipURIs[0];
+        req.locals.callingNumber = sipURIs[1];
+      }
+    }
     next();
   }
 
@@ -90,8 +103,10 @@ module.exports = function(srf, logger) {
           .find((p) => p.type === 'application/sdp')
           .content;
         const {sdp1, sdp2, ...metadata} = await parseSiprecPayload(req, logger);
-        req.locals.calledNumber = metadata.caller.number;
-        req.locals.callingNumber = metadata.callee.number;
+        if (!req.locals.calledNumber && !req.locals.calledNumber) {
+          req.locals.calledNumber = metadata.caller.number;
+          req.locals.callingNumber = metadata.callee.number;
+        }
         req.locals = {
           ...req.locals,
           siprec: {

lib/session/call-info.js

@@ -33,6 +33,23 @@ class CallInfo {
       this.callStatus = CallStatus.Trying;
       this.originatingSipIp = req.get('X-Forwarded-For');
       this.originatingSipTrunkName = req.get('X-Originating-Carrier');
+      const {siprec} = req.locals;
+      if (siprec) {
+        const caller = parseUri(req.locals.callingNumber);
+        const callee = parseUri(req.locals.calledNumber);
+        this.participants = [
+          {
+            participant: 'caller',
+            uriUser: caller.user,
+            uriHost: caller.host
+          },
+          {
+            participant: 'callee',
+            uriUser: callee.user,
+            uriHost: callee.host
+          }
+        ];
+      }
     }
     else if (opts.parentCallInfo) {
       // outbound call that is a child of an existing call

lib/session/call-session.js

@@ -511,12 +511,24 @@ class CallSession extends Emitter {
 
   async enableBotMode(gather, autoEnable) {
     try {
-      if (this.backgroundGatherTask) {
-        this.logger.info('CallSession:enableBotMode - bot mode currently enabled, ignoring request to start again');
-        return;
-      }
       const t = normalizeJambones(this.logger, [gather]);
-      this.backgroundGatherTask = makeTask(this.logger, t[0]);
+      const task = makeTask(this.logger, t[0]);
+
+      if (this.isBotModeEnabled) {
+        const currInput = this.backgroundGatherTask.input;
+        const newInput = task.input;
+        if (JSON.stringify(currInput) === JSON.stringify(newInput)) {
+          this.logger.info('CallSession:enableBotMode - bot mode currently enabled, ignoring request to start again');
+          return;
+        }
+        else {
+          this.logger.info({currInput, newInput},
+            'CallSession:enableBotMode - restarting background gather to apply new input type');
+          this.backgroundGatherTask.sticky = false;
+          this.disableBotMode();
+        }
+      }
+      this.backgroundGatherTask = task;
       this._bargeInEnabled = true;
       this.backgroundGatherTask
         .once('dtmf', this._clearTasks.bind(this, this.backgroundGatherTask))
@@ -528,13 +540,15 @@ class CallSession extends Emitter {
       const {span, ctx} = this.rootSpan.startChildSpan(`background-gather:${this.backgroundGatherTask.summary}`);
       this.backgroundGatherTask.span = span;
       this.backgroundGatherTask.ctx = ctx;
+      this.backgroundGatherTask.sticky = autoEnable;
       this.backgroundGatherTask.exec(this, resources)
         .then(() => {
           this.logger.info('CallSession:enableBotMode: gather completed');
           this.backgroundGatherTask && this.backgroundGatherTask.removeAllListeners();
           this.backgroundGatherTask && this.backgroundGatherTask.span.end();
+          const sticky = this.backgroundGatherTask?.sticky;
           this.backgroundGatherTask = null;
-          if (autoEnable && !this.callGone && !this._stopping && this._bargeInEnabled) {
+          if (sticky && !this.callGone && !this._stopping && this._bargeInEnabled) {
             this.logger.info('CallSession:enableBotMode: restarting background gather');
             setImmediate(() => this.enableBotMode(gather, true));
           }
@@ -636,7 +650,9 @@ class CallSession extends Emitter {
           return {
             speech_credential_sid: credential.speech_credential_sid,
             client_id: credential.client_id,
-            secret: credential.secret
+            secret: credential.secret,
+            nuance_tts_uri: credential.nuance_tts_uri,
+            nuance_stt_uri: credential.nuance_stt_uri
           };
         }
         else if ('deepgram' === vendor) {
@@ -698,7 +714,7 @@ class CallSession extends Emitter {
         let skip = false;
         this.currentTask = task;
         if (TaskName.Gather === task.name && this.isBotModeEnabled) {
-          if (this.backgroundGatherTask.updateTaskInProgress(task)) {
+          if (this.backgroundGatherTask.updateTaskInProgress(task) !== false) {
             this.logger.info(`CallSession:exec skipping #${stackNum}:${taskNum}: ${task.name}`);
             skip = true;
           }
@@ -762,7 +778,6 @@ class CallSession extends Emitter {
 
   trackTmpFile(path) {
     // TODO: don't add if its already in the list (should we make it a set?)
-    this.logger.debug(`adding tmp file to track ${path}`);
     this.tmpFiles.add(path);
   }
 
@@ -1137,14 +1152,14 @@ class CallSession extends Emitter {
   _injectTasks(newTasks) {
     const gatherPos = this.tasks.map((t) => t.name).indexOf(TaskName.Gather);
     const currentlyExecutingGather = this.currentTask?.name === TaskName.Gather;
-
+    /*
     this.logger.debug({
       currentTaskList: listTaskNames(this.tasks),
       newContent: listTaskNames(newTasks),
       currentlyExecutingGather,
       gatherPos
     }, 'CallSession:_injectTasks - starting');
-
+    */
     const killGather = () => {
       this.logger.debug('CallSession:_injectTasks - killing current gather because we have new content');
       this.currentTask.kill(this);
@@ -1153,10 +1168,11 @@ class CallSession extends Emitter {
     if (-1 === gatherPos) {
       /* no gather in the stack  simply append tasks */
       this.tasks.push(...newTasks);
+      /*
       this.logger.debug({
         updatedTaskList: listTaskNames(this.tasks)
       }, 'CallSession:_injectTasks - completed (simple append)');
-
+      */
       /* we do need to kill the current gather if we are executing one */
       if (currentlyExecutingGather) killGather();
       return;
@@ -1184,12 +1200,10 @@ class CallSession extends Emitter {
             this.replaceApplication(t);
           }
           else if (process.env.JAMBONES_INJECT_CONTENT) {
-            this.logger.debug({tasks: listTaskNames(t)}, 'CallSession:_onCommand - queueing tasks (injecting content)');
             this._injectTasks(t);
             this.logger.info({tasks: listTaskNames(this.tasks)}, 'CallSession:_onCommand - updated task list');
           }
           else {
-            this.logger.debug({tasks: listTaskNames(t)}, 'CallSession:_onCommand - queueing tasks');
             this.tasks.push(...t);
             this.logger.info({tasks: listTaskNames(this.tasks)}, 'CallSession:_onCommand - updated task list');
           }
@@ -1233,7 +1247,7 @@ class CallSession extends Emitter {
         this.logger.info(`CallSession:_onCommand - invalid command ${command}`);
     }
     if (this.wakeupResolver) {
-      this.logger.debug({resolution}, 'CallSession:_onCommand - got commands, waking up..');
+      //this.logger.debug({resolution}, 'CallSession:_onCommand - got commands, waking up..');
       this.wakeupResolver(resolution);
       this.wakeupResolver = null;
     }

lib/session/siprec-call-session.js

@@ -1,6 +1,7 @@
 const InboundCallSession = require('./inbound-call-session');
 const {createSipRecPayload} = require('../utils/siprec-utils');
 const {CallStatus} = require('../utils/constants');
+const {parseSiprecPayload} = require('../utils/siprec-utils');
 /**
  * @classdesc Subclass of InboundCallSession.  This represents a CallSession that is
  * established for an inbound SIPREC call.
@@ -16,6 +17,32 @@ class SipRecCallSession extends InboundCallSession {
     this.metadata = metadata;
   }
 
+  async _onReinvite(req, res) {
+    try {
+      this.logger.info(req.payload, 'SipRec Re-INVITE payload');
+      const {sdp1: reSdp1, sdp2: reSdp2, metadata: reMetadata} = await parseSiprecPayload(req, this.logger);
+      this.sdp1 = reSdp1;
+      this.sdp2 = reSdp2;
+      this.metadata = reMetadata;
+
+      if (this.ep && this.ep2) {
+        let remoteSdp = this.sdp1.replace(/sendonly/, 'sendrecv');
+        const newSdp1 = await this.ep.modify(remoteSdp);
+        remoteSdp = this.sdp2.replace(/sendonly/, 'sendrecv');
+        const newSdp2 = await this.ep2.modify(remoteSdp);
+        const combinedSdp = await createSipRecPayload(newSdp1, newSdp2, this.logger);
+        res.send(200, {body: combinedSdp});
+        this.logger.info({offer: req.body, answer: combinedSdp}, 'SipRec handling reINVITE');
+      }
+      else {
+        this.logger.info('got SipRec reINVITE but no endpoint and media has not been released');
+        res.send(488);
+      }
+    } catch (err) {
+      this.logger.error(err, 'Error handling reinvite');
+    }
+  }
+
   async answerSipRecCall() {
     try {
       this.ms = this.getMS();

lib/tasks/gather.js

@@ -54,7 +54,7 @@ class TaskGather extends Task {
 
     /* timeout of zero means no timeout */
     this.timeout = this.timeout === 0 ? 0 : (this.timeout || 15) * 1000;
-    this.interim = !!this.partialResultHook || this.bargein;
+    this.interim = !!this.partialResultHook || this.bargein || (this.timeout > 0);
     this.listenDuringPrompt = this.data.listenDuringPrompt === false ? false : true;
     this.minBargeinWordCount = this.data.minBargeinWordCount || 1;
     if (this.data.recognizer) {
@@ -70,6 +70,11 @@ class TaskGather extends Task {
       if (this.asrTimeout > 0) this.asrDtmfTerminationDigit = recognizer.asrDtmfTerminationDigit;
       this.isContinuousAsr = this.asrTimeout > 0;
 
+      if (Array.isArray(this.data.recognizer.hints) &&
+        0 == this.data.recognizer.hints.length && process.env.JAMBONES_GATHER_CLEAR_GLOBAL_HINTS_ON_EMPTY_HINTS) {
+        logger.debug('Gather: an empty hints array was supplied, so we will mask global hints');
+        this.maskGlobalSttHints = true;
+      }
       this.data.recognizer.hints = this.data.recognizer.hints || [];
       this.data.recognizer.altLanguages = this.data.recognizer.altLanguages || [];
     }
@@ -93,12 +98,17 @@ class TaskGather extends Task {
     this._sonioxTranscripts = [];
 
     this.parentTask = parentTask;
+    this.partialTranscriptsCount = 0;
   }
 
   get name() { return TaskName.Gather; }
 
   get needsStt() { return this.input.includes('speech'); }
 
+  get wantsSingleUtterance() {
+    return this.data.recognizer?.singleUtterance === true;
+  }
+
   get earlyMedia() {
     return (this.sayTask && this.sayTask.earlyMedia) ||
       (this.playTask && this.playTask.earlyMedia);
@@ -120,14 +130,17 @@ class TaskGather extends Task {
   }
 
   async exec(cs, {ep}) {
-    this.logger.debug('Gather:exec');
+    this.logger.debug({options: this.data}, 'Gather:exec');
     await super.exec(cs);
     const {updateSpeechCredentialLastUsed} = require('../utils/db-utils')(this.logger, cs.srf);
     const {getNuanceAccessToken, getIbmAccessToken} = cs.srf.locals.dbHelpers;
 
-    if (cs.hasGlobalSttHints) {
+    if (cs.hasGlobalSttHints && !this.maskGlobalSttHints) {
       const {hints, hintsBoost} = cs.globalSttHints;
-      this.data.recognizer.hints = this.data.recognizer.hints.concat(hints);
+      const setOfHints = new Set(this.data.recognizer.hints
+        .concat(hints)
+        .filter((h) => typeof h === 'string' && h.length > 0));
+      this.data.recognizer.hints = [...setOfHints];
       if (!this.data.recognizer.hintsBoost && hintsBoost) this.data.recognizer.hintsBoost = hintsBoost;
       this.logger.debug({hints: this.data.recognizer.hints, hintsBoost: this.data.recognizer.hintsBoost},
         'Gather:exec - applying global sttHints');
@@ -188,7 +201,6 @@ class TaskGather extends Task {
       throw new Error(`No speech-to-text service credentials for ${this.vendor} have been configured`);
     }
 
-    this.logger.info({sttCredentials: this.sttCredentials}, 'Gather:exec - sttCredentials');
     if (this.vendor === 'nuance' && this.sttCredentials.client_id) {
       /* get nuance access token */
       const {client_id, secret} = this.sttCredentials;
@@ -207,7 +219,6 @@ class TaskGather extends Task {
       this._startTimer();
       if (this.isContinuousAsr && 0 === this.timeout) this._startAsrTimer();
       if (this.input.includes('speech') && !this.listenDuringPrompt) {
-        this.logger.debug('Gather:exec - calling _initSpeech');
         this._initSpeech(cs, ep)
           .then(() => {
             if (this.killed) {
@@ -309,6 +320,7 @@ class TaskGather extends Task {
     const {timeout} = opts;
     this.timeout = timeout;
     this._startTimer();
+    return true;
   }
 
   _onDtmf(cs, ep, evt) {
@@ -348,7 +360,6 @@ class TaskGather extends Task {
 
   async _initSpeech(cs, ep) {
     const opts = this.setChannelVarsForStt(this, this.sttCredentials, this.data.recognizer);
-    this.logger.debug(opts, 'TaskGather:_initSpeech - channel vars');
     switch (this.vendor) {
       case 'google':
         this.bugname = 'google_transcribe';
@@ -550,8 +561,13 @@ class TaskGather extends Task {
     }
 
     evt = this.normalizeTranscription(evt, this.vendor, 1, this.language);
+    if (evt.alternatives.length === 0) {
+      this.logger.info({evt}, 'TaskGather:_onTranscription - got empty transcript, continue listening');
+      return;
+    }
 
-    if (this.earlyHintsMatch && evt.is_final === false) {
+    /* fast path: our first partial transcript exactly matches an early hint */
+    if (this.earlyHintsMatch && evt.is_final === false && this.partialTranscriptsCount++ === 0) {
       const transcript = evt.alternatives[0].transcript?.toLowerCase();
       const hints = this.data.recognizer?.hints || [];
       if (hints.find((h) => h.toLowerCase() === transcript)) {
@@ -625,6 +641,8 @@ class TaskGather extends Task {
         others do not.
       */
       //const isStableEnough = typeof evt.stability === 'undefined' || evt.stability > GATHER_STABILITY_THRESHOLD;
+      this._clearTimer();
+      this._startTimer();
       if (this.bargein && (words + bufferedWords) >= this.minBargeinWordCount) {
         if (!this.playComplete) {
           this.logger.debug({transcript: evt.alternatives[0].transcript}, 'killing audio due to speech');
@@ -653,7 +671,16 @@ class TaskGather extends Task {
       this._killAudio(cs);
     }
 
-    if (!this.resolved && !this.killed && !this._bufferedTranscripts.length) {
+    /**
+     * By default, Gather asks google for multiple utterances.
+     * The reason is that we can sometimes get an 'end_of_utterance' event without
+     * getting a transcription.  This can happen if someone coughs or mumbles.
+     * For that reason don't ask for a single utterance and we'll terminate the transcribe operation
+     * once we get a final transcript.
+     * However, if the usr has specified a singleUtterance, then we need to restart here
+     * since we dont have a final transcript yet.
+     */
+    if (!this.resolved && !this.killed && !this._bufferedTranscripts.length && this.wantsSingleUtterance) {
       this._startTranscribing(ep);
     }
   }
@@ -767,6 +794,10 @@ class TaskGather extends Task {
     if (this.resolved) return;
 
     this.resolved = true;
+    // Clear dtmf event
+    if (this.dtmfBargein) {
+      this.ep.removeAllListeners('dtmf');
+    }
     clearTimeout(this.interDigitTimer);
     this._clearTimer();
 

lib/tasks/listen.js

@@ -2,6 +2,7 @@ const Task = require('./task');
 const {TaskName, TaskPreconditions, ListenEvents, ListenStatus} = require('../utils/constants');
 const makeTask = require('./make_task');
 const moment = require('moment');
+const MAX_PLAY_AUDIO_QUEUE_SIZE = 10;
 
 class TaskListen extends Task {
   constructor(logger, opts, parentTask) {
@@ -20,6 +21,8 @@ class TaskListen extends Task {
     this.nested = parentTask instanceof Task;
 
     this.results = {};
+    this.playAudioQueue = [];
+    this.isPlayingAudioFromQueue = false;
 
     if (this.transcribe) this.transcribeTask = makeTask(logger, {'transcribe': opts.transcribe}, this);
   }
@@ -58,6 +61,7 @@ class TaskListen extends Task {
     super.kill(cs);
     this.logger.debug(`TaskListen:kill endpoint connected? ${this.ep && this.ep.connected}`);
     this._clearTimer();
+    this.playAudioQueue = [];
     if (this.ep && this.ep.connected) {
       this.logger.debug('TaskListen:kill closing websocket');
       try {
@@ -184,16 +188,36 @@ class TaskListen extends Task {
     this.notifyTaskDone();
   }
 
-  async _onPlayAudio(ep, evt) {
-    this.logger.info(`received play_audio event: ${JSON.stringify(evt)}`);
+  async _playAudio(ep, evt, logger) {
     try {
       const results = await ep.play(evt.file);
-      this.logger.debug(`Finished playing file, result: ${JSON.stringify(results)}`);
+      logger.debug(`Finished playing file, result: ${JSON.stringify(results)}`);
       ep.forkAudioSendText({type: 'playDone', data: Object.assign({id: evt.id}, results)});
+    } catch (err) {
+      logger.error({err}, 'Error playing file');
     }
-    catch (err) {
-      this.logger.error({err}, 'Error playing file');
+  }
+
+  async _onPlayAudio(ep, evt) {
+    this.logger.info(`received play_audio event: ${JSON.stringify(evt)}`);
+    if (!evt.queuePlay) {
+      this.playAudioQueue = [];
+      this._playAudio(ep, evt, this.logger);
+      this.isPlayingAudioFromQueue = false;
+      return;
     }
+
+    if (this.playAudioQueue.length <= MAX_PLAY_AUDIO_QUEUE_SIZE) {
+      this.playAudioQueue.push(evt);
+    }
+
+    if (this.isPlayingAudioFromQueue) return;
+
+    this.isPlayingAudioFromQueue = true;
+    while (this.playAudioQueue.length > 0) {
+      await this._playAudio(ep, this.playAudioQueue.shift(), this.logger);
+    }
+    this.isPlayingAudioFromQueue = false;
   }
 
   _onKillAudio(ep) {

lib/tasks/say.js

@@ -154,7 +154,6 @@ class TaskSay extends Task {
 
       const arr = this.text.map((t) => generateAudio(t));
       const filepath = (await Promise.all(arr)).filter((fp) => fp && fp.length);
-      this.logger.debug({filepath}, 'synthesized files for tts');
       this.notifyStatus({event: 'start-playback'});
 
       while (!this.killed && (this.loop === 'forever' || this.loop--) && this.ep?.connected) {

lib/tasks/transcribe.js

@@ -259,8 +259,11 @@ class TaskTranscribe extends Task {
     this.logger.debug({evt}, 'TaskTranscribe:_onTranscription - before normalization');
 
     evt = this.normalizeTranscription(evt, this.vendor, channel, this.language);
-
     this.logger.debug({evt}, 'TaskTranscribe:_onTranscription');
+    if (evt.alternatives.length === 0) {
+      this.logger.info({evt}, 'TaskTranscribe:_onTranscription - got empty transcript, continue listening');
+      return;
+    }
 
     if (evt.alternatives[0]?.transcript === '' && !cs.callGone && !this.killed) {
       if (['microsoft', 'deepgram'].includes(this.vendor)) {

lib/utils/amd-utils.js

@@ -267,7 +267,6 @@ module.exports = (logger) => {
       ep.addCustomEventListener(AzureTranscriptionEvents.Transcription, onTranscription.bind(null, cs, ep, task));
       ep.addCustomEventListener(AzureTranscriptionEvents.NoSpeechDetected, onNoSpeechDetected.bind(null, cs, ep, task));
     }
-    logger.debug({sttOpts}, 'startAmd: setting channel vars');
     await ep.set(sttOpts).catch((err) => logger.info(err, 'Error setting channel variables'));
 
     amd

lib/utils/db-utils.js

@@ -50,6 +50,8 @@ const speechMapper = (cred) => {
       const o = JSON.parse(decrypt(credential));
       obj.client_id = o.client_id;
       obj.secret = o.secret;
+      obj.nuance_tts_uri = o.nuance_tts_uri;
+      obj.nuance_stt_uri = o.nuance_stt_uri;
     }
     else if ('ibm' === obj.vendor) {
       const o = JSON.parse(decrypt(credential));

lib/utils/encrypt-decrypt.js

@@ -2,9 +2,9 @@ const crypto = require('crypto');
 const algorithm = process.env.LEGACY_CRYPTO ? 'aes-256-ctr' : 'aes-256-cbc';
 const iv = crypto.randomBytes(16);
 const secretKey = crypto.createHash('sha256')
-  .update(String(process.env.JWT_SECRET))
+  .update(process.env.ENCRYPTION_SECRET || process.env.JWT_SECRET)
   .digest('base64')
-  .substr(0, 32);
+  .substring(0, 32);
 
 const encrypt = (text) => {
   const cipher = crypto.createCipheriv(algorithm, secretKey, iv);
@@ -25,8 +25,8 @@ const decrypt = (data) => {
     throw err;
   }
   const decipher = crypto.createDecipheriv(algorithm, secretKey, Buffer.from(hash.iv, 'hex'));
-  const decrpyted = Buffer.concat([decipher.update(Buffer.from(hash.content, 'hex')), decipher.final()]);
-  return decrpyted.toString();
+  const decrypted = Buffer.concat([decipher.update(Buffer.from(hash.content, 'hex')), decipher.final()]);
+  return decrypted.toString();
 };
 
 module.exports = {

lib/utils/siprec-utils.js

@@ -47,8 +47,16 @@ const parseSiprecPayload = (req, logger) => {
       }
     }
 
-    if (!sdp || !meta) {
-      logger.info({payload: req.payload}, 'invalid SIPREC payload');
+
+    if (!meta && sdp) {
+      const arr = /^([^]+)(m=[^]+?)(m=[^]+?)$/.exec(sdp);
+      opts.sdp1 = `${arr[1]}${arr[2]}`;
+      opts.sdp2 = `${arr[1]}${arr[3]}\r\n`;
+      opts.sessionId = uuidv4();
+      logger.info({ payload: req.payload }, 'SIPREC payload with no metadata (e.g. Cisco NBR)');
+      resolve(opts);
+    } else if (!sdp || !meta) {
+      logger.info({ payload: req.payload }, 'invalid SIPREC payload');
       return reject(new Error('expected multipart SIPREC body'));
     }
 
@@ -242,7 +250,8 @@ const createSipRecPayload = (sdp1, sdp2, logger) => {
     .replace(/a=sendonly\r\n/g, '')
     .replace(/a=direction:both\r\n/g, '');
   */
-  return combinedSdp;
+
+  return combinedSdp.replace(/sendrecv/g, 'recvonly');
 };
 
 module.exports = { parseSiprecPayload, createSipRecPayload } ;

lib/utils/transcription-utils.js

@@ -29,6 +29,7 @@ const stickyVars = {
     'AZURE_SERVICE_ENDPOINT_ID',
     'AZURE_REQUEST_SNR',
     'AZURE_PROFANITY_OPTION',
+    'AZURE_SPEECH_ALTERNATIVE_LANGUAGE_CODES',
     'AZURE_SERVICE_ENDPOINT',
     'AZURE_INITIAL_SPEECH_TIMEOUT_MS',
     'AZURE_USE_OUTPUT_FORMAT_DETAILED',
@@ -335,55 +336,38 @@ module.exports = (logger) => {
     };
 
     if ('google' === vendor) {
+      const model = task.name === TaskName.Gather ? 'command_and_search' : 'latest_long';
       opts = {
         ...opts,
-        ...(sttCredentials &&
-          {GOOGLE_APPLICATION_CREDENTIALS: JSON.stringify(sttCredentials.credentials)}),
-        ...(rOpts.enhancedModel &&
-            {GOOGLE_SPEECH_USE_ENHANCED: 1}),
-        ...(rOpts.separateRecognitionPerChannel &&
-          {GOOGLE_SPEECH_SEPARATE_RECOGNITION_PER_CHANNEL: 1}),
-        ...(rOpts.profanityFilter &&
-          {GOOGLE_SPEECH_PROFANITY_FILTER: 1}),
-        ...(rOpts.punctuation &&
-          {GOOGLE_SPEECH_ENABLE_AUTOMATIC_PUNCTUATION: 1}),
-        ...(rOpts.words &&
-          {GOOGLE_SPEECH_ENABLE_WORD_TIME_OFFSETS: 1}),
-        ...((rOpts.singleUtterance ||  task.name === TaskName.Gather) &&
-          {GOOGLE_SPEECH_SINGLE_UTTERANCE: 1}),
-        ...(rOpts.diarization &&
-          {GOOGLE_SPEECH_SPEAKER_DIARIZATION: 1}),
+        ...(sttCredentials && {GOOGLE_APPLICATION_CREDENTIALS: JSON.stringify(sttCredentials.credentials)}),
+        ...(rOpts.separateRecognitionPerChannel && {GOOGLE_SPEECH_SEPARATE_RECOGNITION_PER_CHANNEL: 1}),
+        ...(rOpts.separateRecognitionPerChanne === false && {GOOGLE_SPEECH_SEPARATE_RECOGNITION_PER_CHANNEL: 0}),
+        ...(rOpts.profanityFilter && {GOOGLE_SPEECH_PROFANITY_FILTER: 1}),
+        ...(rOpts.punctuation && {GOOGLE_SPEECH_ENABLE_AUTOMATIC_PUNCTUATION: 1}),
+        ...(rOpts.words && {GOOGLE_SPEECH_ENABLE_WORD_TIME_OFFSETS: 1}),
+        ...(rOpts.singleUtterance && {GOOGLE_SPEECH_SINGLE_UTTERANCE: 1}),
+        ...(rOpts.diarization && {GOOGLE_SPEECH_SPEAKER_DIARIZATION: 1}),
         ...(rOpts.diarization && rOpts.diarizationMinSpeakers > 0 &&
           {GOOGLE_SPEECH_SPEAKER_DIARIZATION_MIN_SPEAKER_COUNT: rOpts.diarizationMinSpeakers}),
         ...(rOpts.diarization && rOpts.diarizationMaxSpeakers > 0 &&
           {GOOGLE_SPEECH_SPEAKER_DIARIZATION_MAX_SPEAKER_COUNT: rOpts.diarizationMaxSpeakers}),
-        ...(rOpts.enhancedModel === false &&
-          {GOOGLE_SPEECH_USE_ENHANCED: 0}),
-        ...(rOpts.separateRecognitionPerChannel === false &&
-          {GOOGLE_SPEECH_SEPARATE_RECOGNITION_PER_CHANNEL: 0}),
-        ...(rOpts.profanityFilter === false &&
-          {GOOGLE_SPEECH_PROFANITY_FILTER: 0}),
-        ...(rOpts.punctuation === false &&
-          {GOOGLE_SPEECH_ENABLE_AUTOMATIC_PUNCTUATION: 0}),
-        ...(rOpts.words  == false &&
-          {GOOGLE_SPEECH_ENABLE_WORD_TIME_OFFSETS: 0}),
-        ...((rOpts.singleUtterance === false || task.name === TaskName.Transcribe) &&
-          {GOOGLE_SPEECH_SINGLE_UTTERANCE: 0}),
-        ...(rOpts.diarization === false &&
-          {GOOGLE_SPEECH_SPEAKER_DIARIZATION: 0}),
+        ...(rOpts.enhancedModel !== false && {GOOGLE_SPEECH_USE_ENHANCED: 1}),
+        ...(rOpts.profanityFilter === false && {GOOGLE_SPEECH_PROFANITY_FILTER: 0}),
+        ...(rOpts.punctuation === false && {GOOGLE_SPEECH_ENABLE_AUTOMATIC_PUNCTUATION: 0}),
+        ...(rOpts.words  == false && {GOOGLE_SPEECH_ENABLE_WORD_TIME_OFFSETS: 0}),
+        ...(rOpts.diarization === false && {GOOGLE_SPEECH_SPEAKER_DIARIZATION: 0}),
         ...(rOpts.hints.length > 0 && typeof rOpts.hints[0] === 'string' &&
           {GOOGLE_SPEECH_HINTS: rOpts.hints.join(',')}),
         ...(rOpts.hints.length > 0 && typeof rOpts.hints[0] === 'object' &&
           {GOOGLE_SPEECH_HINTS: JSON.stringify(rOpts.hints)}),
-        ...(typeof rOpts.hintsBoost === 'number' &&
-          {GOOGLE_SPEECH_HINTS_BOOST: rOpts.hintsBoost}),
+        ...(typeof rOpts.hintsBoost === 'number' && {GOOGLE_SPEECH_HINTS_BOOST: rOpts.hintsBoost}),
         ...(rOpts.altLanguages.length > 0 &&
-          {GOOGLE_SPEECH_ALTERNATIVE_LANGUAGE_CODES: rOpts.altLanguages.join(',')}),
+          {GOOGLE_SPEECH_ALTERNATIVE_LANGUAGE_CODES: [...new Set(rOpts.altLanguages)].join(',')}),
         ...(rOpts.interactionType &&
           {GOOGLE_SPEECH_METADATA_INTERACTION_TYPE: rOpts.interactionType}),
-        ...{GOOGLE_SPEECH_MODEL: rOpts.model || (task.name === TaskName.Gather ? 'latest_short' : 'phone_call')},
-        ...(rOpts.naicsCode > 0 &&
-          {GOOGLE_SPEECH_METADATA_INDUSTRY_NAICS_CODE: rOpts.naicsCode}),
+        ...{GOOGLE_SPEECH_MODEL: rOpts.model || model},
+        ...(rOpts.naicsCode > 0 && {GOOGLE_SPEECH_METADATA_INDUSTRY_NAICS_CODE: rOpts.naicsCode}),
+        GOOGLE_SPEECH_METADATA_RECORDING_DEVICE_TYPE: 'phone_line',
       };
     }
     else if (['aws', 'polly'].includes(vendor)) {
@@ -407,7 +391,7 @@ module.exports = (logger) => {
         ...(rOpts.hints.length > 0 && typeof rOpts.hints[0] === 'object' &&
           {AZURE_SPEECH_HINTS: rOpts.hints.map((h) => h.phrase).join(',')}),
         ...(rOpts.altLanguages && rOpts.altLanguages.length > 0 &&
-          {AZURE_SERVICE_ENDPOINT_ID: rOpts.sttCredentials}),
+          {AZURE_SPEECH_ALTERNATIVE_LANGUAGE_CODES: [...new Set(rOpts.altLanguages)].join(',')}),
         ...(rOpts.requestSnr && {AZURE_REQUEST_SNR: 1}),
         ...(rOpts.profanityOption && {AZURE_PROFANITY_OPTION: rOpts.profanityOption}),
         ...(rOpts.azureServiceEndpoint && {AZURE_SERVICE_ENDPOINT: rOpts.azureServiceEndpoint}),
@@ -432,12 +416,9 @@ module.exports = (logger) => {
       const {nuanceOptions = {}} = rOpts;
       opts = {
         ...opts,
-        ...(sttCredentials.access_token) &&
-          {NUANCE_ACCESS_TOKEN: sttCredentials.access_token},
-        ...(sttCredentials.krypton_endpoint) &&
-          {NUANCE_KRYPTON_ENDPOINT: sttCredentials.krypton_endpoint},
-        ...(nuanceOptions.topic) &&
-          {NUANCE_TOPIC: nuanceOptions.topic},
+        ...(sttCredentials.access_token) && {NUANCE_ACCESS_TOKEN: sttCredentials.access_token},
+        ...(sttCredentials.nuance_stt_uri) && {NUANCE_KRYPTON_ENDPOINT: sttCredentials.nuance_stt_uri},
+        ...(nuanceOptions.topic) && {NUANCE_TOPIC: nuanceOptions.topic},
         ...(nuanceOptions.utteranceDetectionMode) &&
           {NUANCE_UTTERANCE_DETECTION_MODE: nuanceOptions.utteranceDetectionMode},
         ...(nuanceOptions.punctuation || rOpts.punctuation) && {NUANCE_PUNCTUATION: nuanceOptions.punctuation},
@@ -661,7 +642,7 @@ module.exports = (logger) => {
     if (recognizer.vendor === 'nuance') {
       const {clientId, secret, kryptonEndpoint} = recognizer.nuanceOptions || {};
       if (clientId && secret) return {client_id: clientId, secret};
-      if (kryptonEndpoint) return {krypton_endpoint: kryptonEndpoint};
+      if (kryptonEndpoint) return {nuance_stt_uri: kryptonEndpoint};
     }
     else if (recognizer.vendor === 'nvidia') {
       const {rivaUri} = recognizer.nvidiaOptions || {};

lib/utils/ws-requestor.js

@@ -219,7 +219,6 @@ class WsRequestor extends BaseRequestor {
   }
 
   _setHandlers(ws) {
-    this.logger.debug('WsRequestor:_setHandlers');
     ws
       .once('open', this._onOpen.bind(this, ws))
       .once('close', this._onClose.bind(this))
@@ -339,7 +338,7 @@ class WsRequestor extends BaseRequestor {
       this.logger.info({url: this.url}, `WsRequestor:_recvAck - ack to unknown msgid ${msgid}, discarding`);
       return;
     }
-    this.logger.debug({url: this.url}, `WsRequestor:_recvAck - received response to ${msgid}`);
+    //this.logger.debug({url: this.url}, `WsRequestor:_recvAck - received response to ${msgid}`);
     this.messagesInFlight.delete(msgid);
     const {success} = obj;
     success && success(data);

package.json

@@ -1,6 +1,6 @@
 {
   "name": "jambonz-feature-server",
-  "version": "v0.8.1",
+  "version": "v0.8.2",
   "main": "app.js",
   "engines": {
     "node": ">= 10.16.0"
@@ -19,7 +19,7 @@
   "bugs": {},
   "scripts": {
     "start": "node app",
-    "test": "NODE_ENV=test JAMBONES_HOSTING=1 HTTP_POOL=1 DRACHTIO_HOST=127.0.0.1 DRACHTIO_PORT=9060 DRACHTIO_SECRET=cymru JAMBONES_MYSQL_HOST=127.0.0.1 JAMBONES_MYSQL_PORT=3360 JAMBONES_MYSQL_USER=jambones_test JAMBONES_MYSQL_PASSWORD=jambones_test JAMBONES_MYSQL_DATABASE=jambones_test JAMBONES_REDIS_HOST=127.0.0.1 JAMBONES_REDIS_PORT=16379 JAMBONES_LOGLEVEL=error ENABLE_METRICS=0 HTTP_PORT=3000 JAMBONES_SBCS=172.38.0.10 JAMBONES_FREESWITCH=127.0.0.1:8022:JambonzR0ck$:docker-host JAMBONES_TIME_SERIES_HOST=127.0.0.1 JAMBONES_NETWORK_CIDR=172.38.0.0/16 node test/ ",
+    "test": "NODE_ENV=test JAMBONES_HOSTING=1 HTTP_POOL=1 ENCRYPTION_SECRET=foobar DRACHTIO_HOST=127.0.0.1 DRACHTIO_PORT=9060 DRACHTIO_SECRET=cymru JAMBONES_MYSQL_HOST=127.0.0.1 JAMBONES_MYSQL_PORT=3360 JAMBONES_MYSQL_USER=jambones_test JAMBONES_MYSQL_PASSWORD=jambones_test JAMBONES_MYSQL_DATABASE=jambones_test JAMBONES_REDIS_HOST=127.0.0.1 JAMBONES_REDIS_PORT=16379 JAMBONES_LOGLEVEL=error ENABLE_METRICS=0 HTTP_PORT=3000 JAMBONES_SBCS=172.38.0.10 JAMBONES_FREESWITCH=127.0.0.1:8022:JambonzR0ck$:docker-host JAMBONES_TIME_SERIES_HOST=127.0.0.1 JAMBONES_NETWORK_CIDR=172.38.0.0/16 node test/ ",
     "coverage": "./node_modules/.bin/nyc --reporter html --report-dir ./coverage npm run test",
     "jslint": "eslint app.js lib"
   },
@@ -27,8 +27,8 @@
     "@jambonz/db-helpers": "^0.7.4",
     "@jambonz/http-health-check": "^0.0.1",
     "@jambonz/realtimedb-helpers": "^0.7.0",
-    "@jambonz/speech-utils": "^0.0.2",
-    "@jambonz/stats-collector": "^0.1.6",
+    "@jambonz/speech-utils": "^0.0.12",
+    "@jambonz/stats-collector": "^0.1.8",
     "@jambonz/time-series": "^0.2.5",
     "@jambonz/verb-specifications": "^0.0.11",
     "@opentelemetry/api": "^1.4.0",
@@ -61,7 +61,7 @@
     "uuid-random": "^1.3.2",
     "verify-aws-sns-signature": "^0.1.0",
     "ws": "^8.9.0",
-    "xml2js": "^0.4.23"
+    "xml2js": "^0.5.0"
   },
   "devDependencies": {
     "clear-module": "^4.1.2",

test/create-test-db.js

@@ -18,30 +18,38 @@ test('creating schema', (t) => {
     if (err) return t.end(err);
     t.pass('schema and test data successfully created');
 
-    if (process.env.GCP_JSON_KEY && process.env.AWS_ACCESS_KEY_ID && process.env.AWS_SECRET_ACCESS_KEY) {
+    const sql = [];
+    if (process.env.GCP_JSON_KEY) {
       const google_credential = encrypt(process.env.GCP_JSON_KEY);
+      t.pass('adding google credentials');
+      sql.push(`UPDATE speech_credentials SET credential='${google_credential}' WHERE vendor='google';`);
+    }
+    if (process.env.AWS_ACCESS_KEY_ID && process.env.AWS_SECRET_ACCESS_KEY) {
       const aws_credential = encrypt(JSON.stringify({
         access_key_id: process.env.AWS_ACCESS_KEY_ID,
         secret_access_key: process.env.AWS_SECRET_ACCESS_KEY,
         aws_region: process.env.AWS_REGION
       }));
+      t.pass('adding aws credentials');
+      sql.push(`UPDATE speech_credentials SET credential='${aws_credential}' WHERE vendor='aws';`);
+    }
+    if (process.env.MICROSOFT_REGION && process.env.MICROSOFT_API_KEY) {
       const microsoft_credential = encrypt(JSON.stringify({
-        region: process.env.MICROSOFT_REGION || 'useast',
-        api_key: process.env.MICROSOFT_API_KEY || '1234567890'
+        region: process.env.MICROSOFT_REGION,
+        api_key: process.env.MICROSOFT_API_KEY
       }));
-      const cmd = `
-UPDATE speech_credentials SET credential='${google_credential}' WHERE vendor='google';
-UPDATE speech_credentials SET credential='${aws_credential}' WHERE vendor='aws';
-UPDATE speech_credentials SET credential='${microsoft_credential}' WHERE vendor='microsoft';
-`;
+      t.pass('adding microsoft credentials');
+      sql.push(`UPDATE speech_credentials SET credential='${microsoft_credential}' WHERE vendor='microsoft';`);
+    }
+    if (sql.length > 0) {
       const path = `${__dirname}/.creds.sql`;
-      fs.writeFileSync(path, cmd);
+      const cmd = sql.join('\n');
+      fs.writeFileSync(path, sql.join('\n'));
       exec(`mysql -h 127.0.0.1 -u root --protocol=tcp --port=3360  -D jambones_test < ${path}`, (err, stdout, stderr) => {
         console.log(stdout);
         console.log(stderr);
         if (err) return t.end(err);
         fs.unlinkSync(path)
-        fs.writeFileSync(`${__dirname}/credentials/gcp.json`, process.env.GCP_JSON_KEY);
         t.pass('set account-level speech credentials');
         t.end();
       });

test/remove-test-db.js

@@ -5,7 +5,6 @@ test('dropping jambones_test database', (t) => {
   exec(`mysql -h 127.0.0.1 -u root --protocol=tcp --port=3360 < ${__dirname}/db/remove_test_db.sql`, (err, stdout, stderr) => {
     if (err) return t.end(err);
     t.pass('database successfully dropped');
-    fs.unlinkSync(`${__dirname}/credentials/gcp.json`);
     t.end();
   });
 });