Feat/ambient sounds (#678)

* initial support for coaching mode in conference

* wip

* wip

* add support for answer verb

* wip

* wip

* wip

* wip

* wip

* updates to rename option to dub

* wip

* wip

* wip

* update verb-specs

* wip

* wip

* wip

* wip

* wip

* wip

* wip

* wip

* add option to boost audio signal in main channel

* wip

* wip

* wip

* wip

* wip

* wip

* for now, bypass use of streaming apis when generating tts audio for dub tracks

* add nested dub to dial

* wip

* add support for filler noise

* kill filler noise when gather killed

* wip

* wip

* while using sayOnTrack, we have to enclose the say command in double quotes

* disableTtsStreaming = false

* allow transcribe of b leg only on dial verb

* dub.say can either be text or object like say verb with text and synthesizer

* remove loop for sayOnTrack

* update speech-utils

* fixes for testing transcribe verb and support for dub and boostAudioSignal in lcc commands

* add dial.boostAudioSignal

* fix bug where session-level recognizer settings incorrectly overwrite verb-level settings

* update verb specs

* update dial to support array of dub verbs

* fix bug setting gain

* lint

* wip

* update speech-utils

* use new endpoint methods for mod_dub

---------

Co-authored-by: Dave Horton <daveh@beachdognet.com>

lib/session/call-session.js

@@ -14,6 +14,7 @@ const moment = require('moment');
 const assert = require('assert');
 const sessionTracker = require('./session-tracker');
 const makeTask = require('../tasks/make_task');
+const parseDecibels = require('../utils/parse-decibels');
 const { normalizeJambones } = require('@jambonz/verb-specifications');
 const listTaskNames = require('../utils/summarize-tasks');
 const HttpRequestor = require('../utils/http-requestor');
@@ -511,6 +512,18 @@ class CallSession extends Emitter {
     this.speechSynthesisVoice = this._origSynthesizerSettings.voice;
   }
 
+  enableFillerNoise(opts) {
+    this._fillerNoise = opts;
+  }
+
+  disableFillerNoise() {
+    this._fillerNoise = null;
+  }
+
+  get fillerNoise() {
+    return this._fillerNoise;
+  }
+
   async notifyRecordOptions(opts) {
     const {action} = opts;
     this.logger.debug({opts}, 'CallSession:notifyRecordOptions');
@@ -698,6 +711,8 @@ class CallSession extends Emitter {
     }
   }
   async disableBotMode() {
+    const task = this.backgroundTaskManager.getTask('bargeIn');
+    if (task) task.sticky = false;
     this.backgroundTaskManager.stop('bargeIn');
   }
 
@@ -1212,7 +1227,15 @@ class CallSession extends Emitter {
     this.callInfo.customerData = tag;
   }
 
-  async _lccMuteStatus(callSid, mute) {
+  async _lccConferenceParticipantAction(opts) {
+    const task = this.currentTask;
+    if (!task || TaskName.Conference !== task.name || !this.isInConference) {
+      return this.logger.info('CallSession:_lccConferenceParticipantState - invalid cmd, call is not in conference');
+    }
+    task.doConferenceParticipantAction(this, opts);
+  }
+
+  async _lccMuteStatus(mute, callSid) {
     // this whole thing requires us to be in a Dial or Conference verb
     const task = this.currentTask;
     if (!task || ![TaskName.Dial, TaskName.Conference].includes(task.name)) {
@@ -1344,6 +1367,45 @@ Duration=${duration} `
     task.whisper(tasks, callSid).catch((err) => this.logger.error(err, 'CallSession:_lccWhisper'));
   }
 
+  async _lccDub(opts, callSid) {
+    this.logger.debug({opts}, `CallSession:_lccDub on call_sid ${callSid}`);
+    const t = normalizeJambones(this.logger, [
+      {
+        verb: 'dub',
+        ...opts
+      }
+    ])
+      .map((tdata) => makeTask(this.logger, tdata));
+
+    const dubTask = t[0];
+    const ep = this.currentTask?.name === TaskName.Dial && callSid === this.currentTask?.callSid ?
+      this.currentTask.ep :
+      this.ep;
+
+    const {span, ctx} = this.rootSpan.startChildSpan(`verb:${dubTask.summary}`);
+    span.setAttributes({'verb.summary': dubTask.summary});
+    dubTask.span = span;
+    dubTask.ctx = ctx;
+    try {
+      await dubTask.exec(this, {ep});
+    } catch (err) {
+      this.logger.error(err, 'CallSession:_lccDub');
+    }
+    dubTask.span.end();
+  }
+
+  async _lccBoostAudioSignal(opts, callSid) {
+    const ep = this.currentTask?.name === TaskName.Dial && callSid === this.currentTask?.callSid ?
+      this.currentTask.ep :
+      this.ep;
+    const db = parseDecibels(opts);
+    this.logger.info(`_lccBoostAudioSignal: boosting audio signal by ${db} dB`);
+    const args = [ep.uuid, 'setGain', db];
+    const response = await ep.api('uuid_dub', args);
+    this.logger.info({response}, '_lccBoostAudioSignal: response from freeswitch');
+  }
+
+
   /**
    * perform call hangup by jambonz
    */
@@ -1374,7 +1436,7 @@ Duration=${duration} `
       await this._lccTranscribeStatus(opts);
     }
     else if (opts.mute_status) {
-      await this._lccMuteStatus(callSid, opts.mute_status === 'mute');
+      await this._lccMuteStatus(opts.mute_status === 'mute', callSid);
     }
     else if (opts.conf_hold_status) {
       await this._lccConfHoldStatus(opts);
@@ -1394,6 +1456,15 @@ Duration=${duration} `
     else if (opts.tag) {
       return this._lccTag(opts);
     }
+    else if (opts.conferenceParticipantAction) {
+      return this._lccConferenceParticipantState(opts);
+    }
+    else if (opts.dub) {
+      return this._lccDub(opts);
+    }
+    else if (opts.boostAudioSignal) {
+      return this._lccBoostAudioSignal(opts, callSid);
+    }
 
     // whisper may be the only thing we are asked to do, or it may that
     // we are doing a whisper after having muted, paused recording etc..
@@ -1597,12 +1668,16 @@ Duration=${duration} `
         this._lccCallDial(data);
         break;
 
+      case 'dub':
+        this._lccDub(data, call_sid);
+        break;
+
       case 'record':
         this.notifyRecordOptions(data);
         break;
 
       case 'mute:status':
-        this._lccMuteStatus(call_sid, data);
+        this._lccMuteStatus(data, call_sid);
         break;
 
       case 'conf:mute-status':
@@ -1613,6 +1688,10 @@ Duration=${duration} `
         this._lccConfHoldStatus(data);
         break;
 
+      case 'conf:participant-action':
+        this._lccConferenceParticipantAction(data);
+        break;
+
       case 'listen:status':
         this._lccListenStatus(data);
         break;
@@ -1639,6 +1718,13 @@ Duration=${duration} `
           });
         break;
 
+      case 'boostAudioSignal':
+        this._lccBoostAudioSignal(data, call_sid)
+          .catch((err) => {
+            this.logger.info({err, data}, 'CallSession:_onCommand - error boosting audio signal');
+          });
+        break;
+
       default:
         this.logger.info(`CallSession:_onCommand - invalid command ${command}`);
     }