Should not synthesis without speak property (#3059)

* Should not synthesis without speak property * Add entry * Update CHANGELOG.md Co-Authored-By: TJ Durnford <[email protected]> Co-authored-by: TJ Durnford <[email protected]>
microsoft · Apr 2, 2020 · 1855c73 · 1855c73
1 parent 903f181
commit 1855c73
Show file tree

Hide file tree

Showing 6 changed files with 44 additions and 11 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -35,6 +35,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 -  Fixes [#2818](https://github.com/microsoft/BotFramework-WebChat/issues/2818). Fix user ID is not set when passing to embed as query parameter, by [@p-nagpal](https://github.com/p-nagpal) in PR [#3031](https://github.com/microsoft/BotFramework-WebChat/pull/3031)
 -  Fixes [#3026](https://github.com/microsoft/BotFramework-WebChat/issues/3026). Fix link `rel` attribute in the `renderMarkdown` function, by [@tdurnford](https://github.com/tdurnford) in PR [#3033](https://github.com/microsoft/BotFramework-WebChat/pull/3033)
 -  Fixes [#2933](https://github.com/microsoft/BotFramework-WebChat/issues/2933). Fix `text` should not be ignored in `messageBack` action in hero card, by [@geea-develop](https://github.com/geea-develop) and [@compulim](https://github.com/compulim) in PR [#3003](https://github.com/microsoft/BotFramework-WebChat/pull/3003)
+-  Fixes [#2953](https://github.com/microsoft/BotFramework-WebChat/issues/2953). Direct Line Speech should not synthesize when the `speak` property is falsy, by [@compulim](https://github.com/compulim) in PR [#3059](https://github.com/microsoft/BotFramework-WebChat/pull/3059)
 
 ### Changed
 

diff --git a/packages/directlinespeech/__tests__/constants.json b/packages/directlinespeech/__tests__/constants.json
@@ -0,0 +1,5 @@
+{
+  "timeouts": {
+    "test": 15000
+  }
+}
diff --git a/packages/directlinespeech/__tests__/sendSpeechActivity.js b/packages/directlinespeech/__tests__/sendSpeechActivity.js
@@ -4,14 +4,15 @@
 
 import 'global-agent/bootstrap';
 
+import { timeouts } from './constants.json';
 import createTestHarness from './utilities/createTestHarness';
 import MockAudioContext from './utilities/MockAudioContext';
 import recognizeActivityAsText from './utilities/recognizeActivityAsText';
 import subscribeAll from './utilities/observable/subscribeAll';
 import take from './utilities/observable/take';
 import waitForConnected from './utilities/waitForConnected';
 
-jest.setTimeout(15000);
+jest.setTimeout(timeouts.test);
 
 beforeEach(() => {
   global.AudioContext = MockAudioContext;
@@ -58,3 +59,21 @@ test('should echo back "Bellevue" when saying "bellview"', async () => {
     ]
   `);
 });
+
+test('should not synthesis when "speak" is empty', async () => {
+  const { directLine, sendTextAsSpeech } = await createTestHarness();
+
+  const connectedPromise = waitForConnected(directLine);
+  const activitiesPromise = subscribeAll(take(directLine.activity$, 1));
+
+  await connectedPromise;
+
+  // "Don't speak XXX" command will not send "speak" property on respond.
+  await sendTextAsSpeech("Don't speak anything.");
+
+  const activities = await activitiesPromise;
+  const activityUtterances = await Promise.all(activities.map(activity => recognizeActivityAsText(activity)));
+
+  expect(activityUtterances).toHaveProperty('length', 1);
+  expect(activityUtterances[0]).toBeFalsy();
+});
diff --git a/packages/directlinespeech/__tests__/utilities/createQueuedArrayBufferAudioSource.js b/packages/directlinespeech/__tests__/utilities/createQueuedArrayBufferAudioSource.js
@@ -17,6 +17,8 @@ import { EventSource } from 'microsoft-cognitiveservices-speech-sdk/distrib/lib/
 import { PromiseHelper } from 'microsoft-cognitiveservices-speech-sdk/distrib/lib/src/common/Promise';
 import { Stream } from 'microsoft-cognitiveservices-speech-sdk/distrib/lib/src/common/Stream';
 
+const CHUNK_SIZE = 4096;
+
 class QueuedArrayBufferAudioSource {
   constructor(audioFormat, audioSourceId = createNoDashGuid()) {
     this._audioFormat = audioFormat;
@@ -114,11 +116,15 @@ class QueuedArrayBufferAudioSource {
 
       const arrayBuffer = this._queue.shift();
 
-      stream.writeStreamChunk({
-        buffer: arrayBuffer,
-        isEnd: false,
-        timeReceived: Date.now()
-      });
+      const { byteLength } = arrayBuffer;
+
+      for (let i = 0; i < byteLength; i += CHUNK_SIZE) {
+        stream.writeStreamChunk({
+          buffer: arrayBuffer.slice(i, Math.min(i + CHUNK_SIZE, byteLength)),
+          isEnd: false,
+          timeReceived: Date.now()
+        });
+      }
 
       stream.close();
 

diff --git a/packages/directlinespeech/__tests__/utilities/recognizeActivityAsText.js b/packages/directlinespeech/__tests__/utilities/recognizeActivityAsText.js
@@ -2,9 +2,11 @@ import readCognitiveServicesAudioStreamAsWAVArrayBuffer from './readCognitiveSer
 import recognizeRiffWaveArrayBuffer from './recognizeRiffWaveArrayBuffer';
 
 export default async function recognizeActivityAsText(activity) {
-  const riffWAVBuffer = await readCognitiveServicesAudioStreamAsWAVArrayBuffer(
-    activity.channelData.speechSynthesisUtterance.audioStream
-  );
+  const { audioStream } = activity.channelData.speechSynthesisUtterance;
 
-  return await recognizeRiffWaveArrayBuffer(riffWAVBuffer);
+  if (audioStream) {
+    const riffWAVBuffer = await readCognitiveServicesAudioStreamAsWAVArrayBuffer(audioStream);
+
+    return await recognizeRiffWaveArrayBuffer(riffWAVBuffer);
+  }
 }
diff --git a/packages/directlinespeech/src/DirectLineSpeech.js b/packages/directlinespeech/src/DirectLineSpeech.js
@@ -41,7 +41,7 @@ export default class DirectLineSpeech {
             ...activity,
             channelData: {
               ...activity.channelData,
-              ...(audioStream ? { speechSynthesisUtterance: new SpeechSynthesisAudioStreamUtterance(audioStream) } : {})
+              speechSynthesisUtterance: new SpeechSynthesisAudioStreamUtterance(audioStream)
             },
             from: {
               ...activity.from,