Adds support for word time offset

GoogleCloudPlatform · Jul 31, 2017 · f48fff1 · f48fff1
1 parent fd0ef5c
commit f48fff1
Show file tree

Hide file tree

Showing 3 changed files with 24 additions and 11 deletions.
diff --git a/speech/cloud-client/pom.xml b/speech/cloud-client/pom.xml
@@ -34,7 +34,7 @@
   </properties>
 
   <!-- FIXME(lesv) - temp to fix an issue w/ GA Datastore -->
-<!--
+  <!--
   <dependencyManagement>
     <dependencies>
       <dependency>
@@ -45,13 +45,12 @@
     </dependencies>
   </dependencyManagement>
  -->
-
   <dependencies>
     <!-- [START dependencies] -->
     <dependency>
       <groupId>com.google.cloud</groupId>
       <artifactId>google-cloud</artifactId>
-      <version>0.20.3-alpha</version>
+      <version>0.21.1-alpha</version>
     </dependency>
     <!-- [END dependencies] -->
 

diff --git a/speech/cloud-client/src/main/java/com/example/speech/Recognize.java b/speech/cloud-client/src/main/java/com/example/speech/Recognize.java
@@ -15,10 +15,9 @@
 */
 
 package com.example.speech;
-
-import com.google.api.gax.grpc.ApiStreamObserver;
-import com.google.api.gax.grpc.OperationFuture;
-import com.google.api.gax.grpc.StreamingCallable;
+import com.google.api.gax.rpc.ApiStreamObserver;
+import com.google.api.gax.rpc.OperationFuture;
+import com.google.api.gax.rpc.StreamingCallable;
 import com.google.cloud.speech.v1.LongRunningRecognizeMetadata;
 import com.google.cloud.speech.v1.LongRunningRecognizeResponse;
 import com.google.cloud.speech.v1.RecognitionAudio;
@@ -32,9 +31,10 @@
 import com.google.cloud.speech.v1.StreamingRecognitionResult;
 import com.google.cloud.speech.v1.StreamingRecognizeRequest;
 import com.google.cloud.speech.v1.StreamingRecognizeResponse;
+import com.google.cloud.speech.v1.WordInfo;
 import com.google.common.util.concurrent.SettableFuture;
+import com.google.longrunning.Operation;
 import com.google.protobuf.ByteString;
-
 import java.io.IOException;
 import java.nio.file.Files;
 import java.nio.file.Path;
@@ -169,8 +169,9 @@ public static void asyncRecognizeFile(String fileName) throws Exception, IOExcep
         .build();
 
     // Use non-blocking call for getting file transcription
-    OperationFuture<LongRunningRecognizeResponse, LongRunningRecognizeMetadata> response =
+    OperationFuture<LongRunningRecognizeResponse, LongRunningRecognizeMetadata, Operation> response =
         speech.longRunningRecognizeAsync(config, audio);
+
     while (!response.isDone()) {
       System.out.println("Waiting for response...");
       Thread.sleep(10000);
@@ -202,13 +203,14 @@ public static void asyncRecognizeGcs(String gcsUri) throws Exception, IOExceptio
         .setEncoding(AudioEncoding.FLAC)
         .setLanguageCode("en-US")
         .setSampleRateHertz(16000)
+        .setEnableWordTimeOffsets(true)
         .build();
     RecognitionAudio audio = RecognitionAudio.newBuilder()
         .setUri(gcsUri)
         .build();
 
     // Use non-blocking call for getting file transcription
-    OperationFuture<LongRunningRecognizeResponse, LongRunningRecognizeMetadata> response =
+    OperationFuture<LongRunningRecognizeResponse, LongRunningRecognizeMetadata, Operation> response =
         speech.longRunningRecognizeAsync(config, audio);
     while (!response.isDone()) {
       System.out.println("Waiting for response...");
@@ -220,7 +222,12 @@ public static void asyncRecognizeGcs(String gcsUri) throws Exception, IOExceptio
     for (SpeechRecognitionResult result: results) {
       List<SpeechRecognitionAlternative> alternatives = result.getAlternativesList();
       for (SpeechRecognitionAlternative alternative: alternatives) {
-        System.out.printf("Transcription: %s%n", alternative.getTranscript());
+        System.out.printf("Transcription: %s\n",alternative.getTranscript());
+        for (WordInfo wordInfo: alternative.getWordsList()) {
+          System.out.println(wordInfo.getWord());
+          System.out.printf("\t%s ns - %s ns\n",
+              wordInfo.getStartTime().getNanos(), wordInfo.getEndTime().getNanos());
+        }
       }
     }
     speech.close();

diff --git a/speech/cloud-client/src/test/java/com/example/speech/RecognizeIT.java b/speech/cloud-client/src/test/java/com/example/speech/RecognizeIT.java
@@ -83,6 +83,13 @@ public void testAsyncRecognizeGcs() throws Exception {
     assertThat(got).contains("how old is the Brooklyn Bridge");
   }
 
+  @Test
+  public void testAsyncWordoffset() throws Exception {
+    Recognize.asyncRecognizeGcs(gcsPath);
+    String got = bout.toString();
+    assertThat(got).contains("\t0.0 sec -");
+  }
+
   @Test
   public void testStreamRecognize() throws Exception {
     Recognize.streamingRecognizeFile(fileName);