initial changes for openai stt (#1127)

* initial changes for openai stt

* wip

* wip

* wip

* wip

* wip

* make minBargeinWordCount work for openai

* wip

* wip

* wip

* wip

* wip

* wip

* wip

* wipp

* wip

* wip

* wip

* openai stt: support for prompt templates

* lint

* wip

* support openai semantic_vad

* wip

* transcribe supports openai stt

* sip

* wip

* wip

* refactor list of stt vendors that dont need to be restarted after a final transcript

* cleanup

* wip

* cleanup

* wip

* wip

* wip

* remove credentials from log

* comment

lib/tasks/transcribe.js

@@ -14,6 +14,7 @@ const {
   TranscribeStatus,
   AssemblyAiTranscriptionEvents,
   VoxistTranscriptionEvents,
+  OpenAITranscriptionEvents,
   VerbioTranscriptionEvents,
   SpeechmaticsTranscriptionEvents
 } = require('../utils/constants.json');
@@ -330,6 +331,20 @@ class TaskTranscribe extends SttTask {
           this._onSpeechmaticsError.bind(this, cs, ep));
         break;
 
+      case 'openai':
+        this.bugname = `${this.bugname_prefix}openai_transcribe`;
+        this.addCustomEventListener(
+          ep, OpenAITranscriptionEvents.Transcription, this._onTranscription.bind(this, cs, ep, channel));
+        this.addCustomEventListener(ep, OpenAITranscriptionEvents.Connect,
+          this._onVendorConnect.bind(this, cs, ep));
+        this.addCustomEventListener(ep, OpenAITranscriptionEvents.ConnectFailure,
+          this._onVendorConnectFailure.bind(this, cs, ep));
+        this.addCustomEventListener(ep, OpenAITranscriptionEvents.Error,
+          this._onOpenAIErrror.bind(this, cs, ep));
+
+        this.modelSupportsConversationTracking = opts.OPENAI_MODEL !== 'whisper-1';
+        break;
+
       default:
         if (this.vendor.startsWith('custom:')) {
           this.bugname = `${this.bugname_prefix}${this.vendor}_transcribe`;
@@ -365,6 +380,25 @@ class TaskTranscribe extends SttTask {
   async _transcribe(ep) {
     this.logger.debug(
       `TaskTranscribe:_transcribe - starting transcription vendor ${this.vendor} bugname ${this.bugname}`);
+
+    /* special feature for openai: we can provide a prompt that includes recent conversation history */
+    let prompt;
+    if (this.vendor === 'openai') {
+      if (this.modelSupportsConversationTracking) {
+        prompt = this.formatOpenAIPrompt(this.cs, {
+          prompt: this.data.recognizer?.openaiOptions?.prompt,
+          hintsTemplate: this.data.recognizer?.openaiOptions?.promptTemplates?.hintsTemplate,
+          // eslint-disable-next-line max-len
+          conversationHistoryTemplate: this.data.recognizer?.openaiOptions?.promptTemplates?.conversationHistoryTemplate,
+          hints: this.data.recognizer?.hints,
+        });
+        this.logger.debug({prompt}, 'Gather:_startTranscribing - created an openai prompt');
+      }
+      else if (this.data.recognizer?.hints?.length > 0) {
+        prompt = this.data.recognizer?.hints.join(', ');
+      }
+    }
+
     await ep.startTranscription({
       vendor: this.vendor,
       interim: this.interim ? true : false,
@@ -456,8 +490,9 @@ class TaskTranscribe extends SttTask {
         this._startAsrTimer(channel);
 
         /* some STT engines will keep listening after a final response, so no need to restart */
-        if (!['soniox', 'aws', 'microsoft', 'deepgram', 'google', 'speechmatics']
-          .includes(this.vendor)) this._startTranscribing(cs, ep, channel);
+        if (!this.doesVendorContinueListeningAfterFinalTranscript(this.vendor)) {
+          this._startTranscribing(cs, ep, channel);
+        }
       }
       else {
         if (this.vendor === 'soniox') {
@@ -480,9 +515,7 @@ class TaskTranscribe extends SttTask {
         this.logger.debug({evt}, 'TaskTranscribe:_onTranscription - sending final transcript');
         this._resolve(channel, evt);
 
-        /* some STT engines will keep listening after a final response, so no need to restart */
-        if (!['soniox', 'aws', 'microsoft', 'deepgram', 'google', 'speechmatics'].includes(this.vendor) &&
-          !this.vendor.startsWith('custom:')) {
+        if (!this.doesVendorContinueListeningAfterFinalTranscript(this.vendor)) {
           this.logger.debug('TaskTranscribe:_onTranscription - restarting transcribe');
           this._startTranscribing(cs, ep, channel);
         }
@@ -733,6 +766,12 @@ class TaskTranscribe extends SttTask {
     this._onVendorError(cs, _ep, {error: JSON.stringify(e)});
   }
 
+  async _onOpenAIErrror(cs, _ep, evt) {
+    // eslint-disable-next-line no-unused-vars
+    const {message, ...e} = evt;
+    this._onVendorError(cs, _ep, {error: JSON.stringify(e)});
+  }
+
   _startAsrTimer(channel) {
     if (this.vendor === 'deepgram') return; // no need
     assert(this.isContinuousAsr);