add real testing of microsoft stt

lib/utils/speech-utils.js

@@ -2,6 +2,8 @@ const ttsGoogle = require('@google-cloud/text-to-speech');
 const sttGoogle = require('@google-cloud/speech').v1p1beta1;
 const Polly = require('aws-sdk/clients/polly');
 const AWS = require('aws-sdk');
+const { Deepgram } = require('@deepgram/sdk');
+const sdk = require('microsoft-cognitiveservices-speech-sdk');
 const bent = require('bent');
 const fs = require('fs');
 
@@ -32,6 +34,65 @@ const testGoogleStt = async(logger, credentials) => {
   }
 };
 
+const testDeepgramStt = async(logger, credentials) => {
+  const {api_key} = credentials;
+  const deepgram = new Deepgram(api_key);
+
+  const mimetype = 'audio/wav';
+  const source = {
+    buffer: fs.readFileSync(`${__dirname}/../../data/test_audio.wav`),
+    mimetype: mimetype
+  };
+
+  return new Promise((resolve, reject) => {
+    // Send the audio to Deepgram and get the response
+    deepgram.transcription
+      .preRecorded(source, {punctuate: true})
+      .then((response) => {
+        //logger.debug({response}, 'got transcript');
+        if (response?.results?.channels[0]?.alternatives?.length > 0) resolve(response);
+        else reject(new Error('no transcript returned'));
+        return;
+      })
+      .catch((err) => {
+        logger.info({err}, 'failed to get deepgram transcript');
+        reject(err);
+      });
+  });
+};
+
+const testMicrosoftStt = async(logger, credentials) => {
+  const {api_key, region} = credentials;
+
+  const speechConfig = sdk.SpeechConfig.fromSubscription(api_key, region);
+  const audioConfig = sdk.AudioConfig.fromWavFileInput(fs.readFileSync(`${__dirname}/../../data/test_audio.wav`));
+  speechConfig.speechRecognitionLanguage = 'en-US';
+  const speechRecognizer = new sdk.SpeechRecognizer(speechConfig, audioConfig);
+
+  return new Promise((resolve, reject) => {
+    speechRecognizer.recognizeOnceAsync((result) => {
+      switch (result.reason) {
+        case sdk.ResultReason.RecognizedSpeech:
+          resolve();
+          break;
+        case sdk.ResultReason.NoMatch:
+          reject('Speech could not be recognized.');
+          break;
+        case sdk.ResultReason.Canceled:
+          const cancellation = sdk.CancellationDetails.fromResult(result);
+          logger.info(`CANCELED: Reason=${cancellation.reason}`);
+          if (cancellation.reason == sdk.CancellationReason.Error) {
+            logger.info(`CANCELED: ErrorCode=${cancellation.ErrorCode}`);
+            logger.info(`CANCELED: ErrorDetails=${cancellation.errorDetails}`);
+          }
+          reject(cancellation.reason);
+          break;
+      }
+      speechRecognizer.close();
+    });
+  });
+};
+
 const testAwsTts = (logger, credentials) => {
   const polly = new Polly(credentials);
   return new Promise((resolve, reject) => {
@@ -89,11 +150,6 @@ const testMicrosoftTts = async(logger, credentials) => {
   }
 };
 
-const testMicrosoftStt = async(logger, credentials) => {
-  //TODO
-  return true;
-};
-
 const testWellSaidTts = async(logger, credentials) => {
   const {api_key} = credentials;
   try {
@@ -127,4 +183,5 @@ module.exports = {
   testMicrosoftTts,
   testMicrosoftStt,
   testWellSaidStt,
+  testDeepgramStt
 };