punctuation for microsoft

jambonz · Dec 18, 2023 · 1fa05b5 · 1fa05b5
1 parent bcb4bf4
commit 1fa05b5
Show file tree

Hide file tree

Showing 3 changed files with 9 additions and 6 deletions.
diff --git a/lib/tasks/gather.js b/lib/tasks/gather.js
@@ -590,7 +590,8 @@ class TaskGather extends SttTask {
       return;
     }
 
-    evt = this.normalizeTranscription(evt, this.vendor, 1, this.language, this.shortUtterance);
+    evt = this.normalizeTranscription(evt, this.vendor, 1, this.language,
+      this.shortUtterance, this.data.recognizer.punctuation);
     if (evt.alternatives.length === 0) {
       this.logger.info({evt}, 'TaskGather:_onTranscription - got empty transcript, continue listening');
       return;

diff --git a/lib/tasks/transcribe.js b/lib/tasks/transcribe.js
@@ -305,7 +305,8 @@ class TaskTranscribe extends SttTask {
     }
     this.logger.debug({evt}, 'TaskTranscribe:_onTranscription - before normalization');
 
-    evt = this.normalizeTranscription(evt, this.vendor, channel, this.language);
+    evt = this.normalizeTranscription(evt, this.vendor, channel, this.language, undefined,
+      this.data.recognizer.punctuation);
     this.logger.debug({evt}, 'TaskTranscribe:_onTranscription');
     if (evt.alternatives.length === 0) {
       this.logger.info({evt}, 'TaskTranscribe:_onTranscription - got empty transcript, continue listening');

diff --git a/lib/utils/transcription-utils.js b/lib/utils/transcription-utils.js
@@ -338,14 +338,15 @@ const normalizeNuance = (evt, channel, language) => {
   };
 };
 
-const normalizeMicrosoft = (evt, channel, language) => {
+const normalizeMicrosoft = (evt, channel, language, punctuation) => {
   const copy = JSON.parse(JSON.stringify(evt));
   const nbest = evt.NBest;
   const language_code = evt.PrimaryLanguage?.Language || language;
   const alternatives = nbest ? nbest.map((n) => {
     return {
       confidence: n.Confidence,
-      transcript: n.Display
+      // remove all puntuation if needed
+      transcript: punctuation ? n.Display : n.Display.replace(/\p{P}/gu, '')
     };
   }) :
     [
@@ -400,14 +401,14 @@ const normalizeAssemblyAi = (evt, channel, language) => {
 };
 
 module.exports = (logger) => {
-  const normalizeTranscription = (evt, vendor, channel, language, shortUtterance) => {
+  const normalizeTranscription = (evt, vendor, channel, language, shortUtterance, punctuation) => {
 
     //logger.debug({ evt, vendor, channel, language }, 'normalizeTranscription');
     switch (vendor) {
       case 'deepgram':
         return normalizeDeepgram(evt, channel, language, shortUtterance);
       case 'microsoft':
-        return normalizeMicrosoft(evt, channel, language);
+        return normalizeMicrosoft(evt, channel, language, punctuation);
       case 'google':
         return normalizeGoogle(evt, channel, language);
       case 'aws':