From e09c763d3a78bad869e4c4887044a28b12e22884 Mon Sep 17 00:00:00 2001
From: Dave Horton <daveh@beachdognet.com>
Date: Tue, 18 Feb 2025 16:30:59 -0500
Subject: [PATCH] #1088 ignore UtteranceEnd if we have unprocessed words
 (#1089)

* #1088 ignore UtteranceEnd if we have unprocessed words

* wip
---
 lib/tasks/gather.js | 29 ++++++++++++++++++++++++-----
 1 file changed, 24 insertions(+), 5 deletions(-)

diff --git a/lib/tasks/gather.js b/lib/tasks/gather.js
index 3e4d6731..6a1bb41a 100644
--- a/lib/tasks/gather.js
+++ b/lib/tasks/gather.js
@@ -785,10 +785,16 @@ class TaskGather extends SttTask {
         this.logger.debug('Gather:_onTranscription - got UtteranceEnd event from deepgram but no buffered transcripts');
       }
       else {
-        this.logger.debug('Gather:_onTranscription - got UtteranceEnd event from deepgram, return buffered transcript');
-        evt = this.consolidateTranscripts(this._bufferedTranscripts, 1, this.language, this.vendor);
-        this._bufferedTranscripts = [];
-        this._resolve('speech', evt);
+        const utteranceTime = evt.last_word_end;
+        if (utteranceTime && this._dgTimeOfLastUnprocessedWord && utteranceTime < this._dgTimeOfLastUnprocessedWord) {
+          this.logger.debug('Gather:_onTranscription - got UtteranceEnd with unprocessed words, continue listening');
+        }
+        else {
+          this.logger.debug('Gather:_onTranscription - got UtteranceEnd from deepgram, return buffered transcript');
+          evt = this.consolidateTranscripts(this._bufferedTranscripts, 1, this.language, this.vendor);
+          this._bufferedTranscripts = [];
+          this._resolve('speech', evt);
+        }
       }
       return;
     }
@@ -922,8 +928,21 @@ class TaskGather extends SttTask {
         if (originalEvent.is_final && evt.alternatives[0].transcript !== '') {
           this.logger.debug({evt}, 'Gather:_onTranscription - buffering a completed (partial) deepgram transcript');
           this._bufferedTranscripts.push(evt);
+          this._dgTimeOfLastUnprocessedWord = null;
+        }
+        if (evt.alternatives[0].transcript === '') {
+          emptyTranscript = true;
+        }
+        else if (!originalEvent.is_final) {
+          /* Deepgram: we have unprocessed words-save last word end time so we can later compare to UtteranceEnd */
+          const words = originalEvent.channel.alternatives[0].words;
+          if (words?.length > 0) {
+            this._dgTimeOfLastUnprocessedWord = words.slice(-1)[0].end;
+            this.logger.debug(
+              `TaskGather:_onTranscription - saving word end time: ${this._dgTimeOfLastUnprocessedWord}`);
+          }
+
         }
-        if (evt.alternatives[0].transcript === '') emptyTranscript = true;
       }
 
       if (!emptyTranscript) {