splits word-time-offset samples into their own files

speech/README.md

@@ -22,16 +22,53 @@ Configure your project using [Application Default Credentials][adc]
 
 ## Usage
 
+To run the Speech Samples:
+
+    $ php speech.php
+
+    Cloud Speech
+
+    Usage:
+      command [options] [arguments]
+
+    Options:
+      -h, --help            Display this help message
+      -q, --quiet           Do not output any message
+      -V, --version         Display this application version
+          --ansi            Force ANSI output
+          --no-ansi         Disable ANSI output
+      -n, --no-interaction  Do not ask any interactive question
+      -v|vv|vvv, --verbose  Increase the verbosity of messages: 1 for normal output, 2 for more verbose output and 3 for debug
+
+    Available commands:
+      help                    Displays help for a command
+      list                    Lists commands
+      transcribe              Transcribe an audio file using Google Cloud Speech API
+      transcribe-async        Transcribe an audio file asynchronously using Google Cloud Speech API
+      transcribe-async-gcs    Transcribe audio asynchronously from a Storage Object using Google Cloud Speech API
+      transcribe-async-words  Transcribe an audio file asynchronously and print word time offsets using Google Cloud Speech API
+      transcribe-gcs          Transcribe audio from a Storage Object using Google Cloud Speech API
+      transcribe-stream       Transcribe a stream of audio using Google Cloud Speech API
+      transcribe-words        Transcribe an audio file and print word time offsets using Google Cloud Speech API
+
 Once you have a speech sample in the proper format, send it through the speech
 API using the transcribe command:
 
 ```sh
 php speech.php transcribe test/data/audio32KHz.raw --encoding LINEAR16 --sample-rate 32000
-php speech.php transcribe test/data/audio32KHz.flac --encoding FLAC --sample-rate 32000 --async
+php speech.php transcribe-async test/data/audio32KHz.flac --encoding FLAC --sample-rate 32000
+php speech.php transcribe-words test/data/audio32KHz.flac --encoding FLAC --sample-rate 32000
 
 ```
 ## Troubleshooting
 
+If you get the following error, set the environment variable `GCLOUD_PROJECT` to your project ID:
+
+```
+[Google\Cloud\Core\Exception\GoogleException]
+No project ID was provided, and we were unable to detect a default project ID.
+```
+
 If you have not set a timezone you may get an error from php. This can be resolved by:
 
   1. Finding where the php.ini is stored by running php -i | grep 'Configuration File'

speech/composer.json

@@ -14,8 +14,10 @@
             "src/streaming_recognize.php",
             "src/transcribe_async.php",
             "src/transcribe_async_gcs.php",
+            "src/transcribe_async_words.php",
             "src/transcribe_sync.php",
-            "src/transcribe_sync_gcs.php"
+            "src/transcribe_sync_gcs.php",
+            "src/transcribe_sync_words.php"
         ]
     },
     "require-dev": {

speech/speech.php

@@ -21,90 +21,178 @@
 use Symfony\Component\Console\Application;
 use Symfony\Component\Console\Command\Command;
 use Symfony\Component\Console\Input\InputArgument;
+use Symfony\Component\Console\Input\InputDefinition;
 use Symfony\Component\Console\Input\InputInterface;
 use Symfony\Component\Console\Input\InputOption;
 use Symfony\Component\Console\Output\OutputInterface;
 
+$inputDefinition = new InputDefinition([
+    new InputArgument('audio-file', InputArgument::REQUIRED, 'The audio file to transcribe'),
+    new InputOption('encoding', null, InputOption::VALUE_REQUIRED,
+        'The encoding of the audio file. This is required if the encoding is ' .
+        'unable to be determined. '
+    ),
+    new InputOption('language-code', null, InputOption::VALUE_REQUIRED,
+        'The language code for the language used in the source file. ',
+        'en-US'
+    ),
+    new InputOption('sample-rate', null, InputOption::VALUE_REQUIRED,
+        'The sample rate of the audio file in hertz. This is required ' .
+        'if the sample rate is unable to be determined. '
+    ),
+    new InputOption('sample-rate', null, InputOption::VALUE_REQUIRED,
+        'The sample rate of the audio file in hertz. This is required ' .
+        'if the sample rate is unable to be determined. '
+    ),
+]);
+
 $application = new Application('Cloud Speech');
 $application->add(new Command('transcribe'))
-    ->setDescription('Transcribe Audio using Google Cloud Speech API')
+    ->setDefinition($inputDefinition)
+    ->setDescription('Transcribe an audio file using Google Cloud Speech API')
     ->setHelp(<<<EOF
-The <info>%command.name%</info> command transcribes audio using the Google Cloud Speech API.
+The <info>%command.name%</info> command transcribes audio from a file using the
+Google Cloud Speech API.
 
 <info>php %command.full_name% audio_file.wav</info>
 
 EOF
     )
-    ->addArgument(
-        'audio-file',
-        InputArgument::REQUIRED,
-        'The audio file to transcribe'
-    )
-    ->addOption(
-        'encoding',
-        null,
-        InputOption::VALUE_REQUIRED,
-        'The encoding of the audio file. This is required if the encoding is ' .
-        'unable to be determined. '
-    )
-    ->addOption(
-        'language-code',
-        null,
-        InputOption::VALUE_REQUIRED,
-        'The language code for the language used in the source file. ',
-        'en-US'
-    )
-    ->addOption(
-        'sample-rate',
-        null,
-        InputOption::VALUE_REQUIRED,
-        'The sample rate of the audio file in hertz. This is required ' .
-        'if the sample rate is unable to be determined. '
+    ->setCode(function (InputInterface $input, OutputInterface $output) {
+        $audioFile = $input->getArgument('audio-file');
+        $languageCode = $input->getOption('language-code');
+        transcribe_sync($audioFile, $languageCode, [
+            'encoding' => $input->getOption('encoding'),
+            'sampleRateHertz' => $input->getOption('sample-rate'),
+        ]);
+    });
+
+$application->add(new Command('transcribe-gcs'))
+    ->setDefinition($inputDefinition)
+    ->setDescription('Transcribe audio from a Storage Object using Google Cloud Speech API')
+    ->setHelp(<<<EOF
+The <info>%command.name%</info> command transcribes audio from a Cloud Storage
+Object using the Google Cloud Speech API.
+
+<info>php %command.full_name% gs://my-bucket/audio_file.wav</info>
+
+EOF
     )
-    ->addOption(
-        'async',
-        null,
-        InputOption::VALUE_NONE,
-        'Run the transcription asynchronously. '
+    ->setCode(function (InputInterface $input, OutputInterface $output) {
+        $audioFile = $input->getArgument('audio-file');
+        $languageCode = $input->getOption('language-code');
+        if (!preg_match('/^gs:\/\/([a-z0-9\._\-]+)\/(\S+)$/', $audioFile, $matches)) {
+            throw new \Exception('Invalid file name. Must be gs://[bucket]/[audiofile]');
+        }
+        list($bucketName, $objectName) = array_slice($matches, 1);
+        transcribe_sync_gcs($bucketName, $objectName, $languageCode, [
+            'encoding' => $input->getOption('encoding'),
+            'sampleRateHertz' => $input->getOption('sample-rate'),
+        ]);
+    });
+
+$application->add(new Command('transcribe-words'))
+    ->setDefinition($inputDefinition)
+    ->setDescription('Transcribe an audio file and print word time offsets using Google Cloud Speech API')
+    ->setHelp(<<<EOF
+The <info>%command.name%</info> command transcribes audio from a file using the
+Google Cloud Speech API and prints word time offsets.
+
+<info>php %command.full_name% audio_file.wav</info>
+
+EOF
     )
-    ->addOption(
-        'stream',
-        null,
-        InputOption::VALUE_NONE,
-        'Stream the audio file.'
+    ->setCode(function (InputInterface $input, OutputInterface $output) {
+        $audioFile = $input->getArgument('audio-file');
+        $languageCode = $input->getOption('language-code');
+        transcribe_sync_words($audioFile, $languageCode, [
+            'encoding' => $input->getOption('encoding'),
+            'sampleRateHertz' => $input->getOption('sample-rate'),
+        ]);
+    });
+
+$application->add(new Command('transcribe-async'))
+    ->setDefinition($inputDefinition)
+    ->setDescription('Transcribe an audio file asynchronously using Google Cloud Speech API')
+    ->setHelp(<<<EOF
+The <info>%command.name%</info> command transcribes audio from a file using the
+Google Cloud Speech API asynchronously.
+
+<info>php %command.full_name% audio_file.wav</info>
+
+EOF
     )
     ->setCode(function (InputInterface $input, OutputInterface $output) {
-        $encoding = $input->getOption('encoding');
+        $audioFile = $input->getArgument('audio-file');
         $languageCode = $input->getOption('language-code');
-        $sampleRate = $input->getOption('sample-rate');
+        transcribe_async($audioFile, $languageCode, [
+            'encoding' => $input->getOption('encoding'),
+            'sampleRateHertz' => $input->getOption('sample-rate'),
+        ]);
+    });
+
+$application->add(new Command('transcribe-async-gcs'))
+    ->setDefinition($inputDefinition)
+    ->setDescription('Transcribe audio asynchronously from a Storage Object using Google Cloud Speech API')
+    ->setHelp(<<<EOF
+The <info>%command.name%</info> command transcribes audio from a Cloud Storage
+object asynchronously using the Google Cloud Speech API.
+
+<info>php %command.full_name% gs://my-bucket/audio_file.wav</info>
+
+EOF
+    )
+    ->setCode(function (InputInterface $input, OutputInterface $output) {
         $audioFile = $input->getArgument('audio-file');
-        $options = [
-            'encoding' => $encoding,
-            'languageCode' => $languageCode,
-            'sampleRateHertz' => $sampleRate,
-        ];
-        if ($isGcs = preg_match('/^gs:\/\/([a-z0-9\._\-]+)\/(\S+)$/', $audioFile, $matches)) {
-            list($bucketName, $objectName) = array_slice($matches, 1);
-        }
-        if ($isGcs) {
-            if ($input->getOption('stream')) {
-                throw new LogicException('Cannot stream from a bucket!');
-            }
-            if ($input->getOption('async')) {
-                transcribe_async_gcs($bucketName, $objectName, $languageCode, $options);
-            } else {
-                transcribe_sync_gcs($bucketName, $objectName, $languageCode, $options);
-            }
-        } else {
-            if ($input->getOption('async')) {
-                transcribe_async($audioFile, $languageCode, $options);
-            } elseif ($input->getOption('stream')) {
-                $encodingInt = constant("Google\Cloud\Speech\V1\RecognitionConfig_AudioEncoding::$encoding");
-                streaming_recognize($audioFile, $languageCode, $encodingInt, $sampleRate);
-            } else {
-                transcribe_sync($audioFile, $languageCode, $options);
-            }
+        $languageCode = $input->getOption('language-code');
+        if (!preg_match('/^gs:\/\/([a-z0-9\._\-]+)\/(\S+)$/', $audioFile, $matches)) {
+            throw new \Exception('Invalid file name. Must be gs://[bucket]/[audiofile]');
         }
+        list($bucketName, $objectName) = array_slice($matches, 1);
+        transcribe_async_gcs($bucketName, $objectName, $languageCode, [
+            'encoding' => $input->getOption('encoding'),
+            'sampleRateHertz' => $input->getOption('sample-rate'),
+        ]);
+    });
+
+$application->add(new Command('transcribe-async-words'))
+    ->setDefinition($inputDefinition)
+    ->setDescription('Transcribe an audio file asynchronously and print word time offsets using Google Cloud Speech API')
+    ->setHelp(<<<EOF
+The <info>%command.name%</info> command transcribes audio from a file using the
+Google Cloud Speech API asynchronously and prints word time offsets.
+
+<info>php %command.full_name% audio_file.wav</info>
+
+EOF
+    )
+    ->setCode(function (InputInterface $input, OutputInterface $output) {
+        $audioFile = $input->getArgument('audio-file');
+        $languageCode = $input->getOption('language-code');
+        transcribe_async_words($audioFile, $languageCode, [
+            'encoding' => $input->getOption('encoding'),
+            'sampleRateHertz' => $input->getOption('sample-rate'),
+        ]);
+    });
+
+$application->add(new Command('transcribe-stream'))
+    ->setDefinition($inputDefinition)
+    ->setDescription('Transcribe a stream of audio using Google Cloud Speech API')
+    ->setHelp(<<<EOF
+The <info>%command.name%</info> command transcribes audio from a stream using
+the Google Cloud Speech API.
+
+<info>php %command.full_name% audio_file.wav</info>
+
+EOF
+    )
+    ->setCode(function (InputInterface $input, OutputInterface $output) {
+        streaming_recognize(
+            $input->getArgument('audio-file'),
+            $input->getOption('language-code'),
+            $input->getOption('encoding'),
+            $input->getOption('sample-rate')
+        );
     });
 
 // for testing

speech/src/streaming_recognize.php

@@ -28,6 +28,7 @@
 use Google\Cloud\Speech\V1\RecognitionConfig;
 use Google\Cloud\Speech\V1\StreamingRecognitionConfig;
 use Google\Cloud\Speech\V1\StreamingRecognizeRequest;
+use Google\Cloud\Speech\V1\RecognitionConfig_AudioEncoding;
 
 /**
  * Transcribe an audio file using Google Cloud Speech API
@@ -64,8 +65,10 @@ function streaming_recognize($audioFile, $languageCode, $encoding, $sampleRateHe
     try {
         $config = new RecognitionConfig();
         $config->setLanguageCode($languageCode);
-        $config->setEncoding($encoding);
         $config->setSampleRateHertz($sampleRateHertz);
+        // encoding must be an enum, convert from string
+        $encodingEnum = constant(RecognitionConfig_AudioEncoding::class . '::' . $encoding);
+        $config->setEncoding($encodingEnum);
 
         $strmConfig = new StreamingRecognitionConfig();
         $strmConfig->setConfig($config);

speech/src/transcribe_async.php

@@ -49,9 +49,6 @@ function transcribe_async($audioFile, $languageCode = 'en-US', $options = [])
         'languageCode' => $languageCode,
     ]);
 
-    // When true, time offsets for every word will be included in the response.
-    $options['enableWordTimeOffsets'] = true;
-
     // Create the asyncronous recognize operation
     $operation = $speech->beginRecognizeOperation(
         fopen($audioFile, 'r'),
@@ -74,12 +71,6 @@ function transcribe_async($audioFile, $languageCode = 'en-US', $options = [])
         foreach ($alternatives as $alternative) {
             printf('Transcript: %s' . PHP_EOL, $alternative['transcript']);
             printf('Confidence: %s' . PHP_EOL, $alternative['confidence']);
-            foreach ($alternative['words'] as $wordInfo) {
-                printf('  Word: %s (start: %s, end: %s)' . PHP_EOL,
-                    $wordInfo['word'],
-                    $wordInfo['startTime'],
-                    $wordInfo['endTime']);
-            }
         }
     }
 }

speech/src/transcribe_async_gcs.php

@@ -55,9 +55,6 @@ function transcribe_async_gcs($bucketName, $objectName, $languageCode = 'en-US',
     $storage = new StorageClient();
     $object = $storage->bucket($bucketName)->object($objectName);
 
-    // When true, time offsets for every word will be included in the response.
-    $options['enableWordTimeOffsets'] = true;
-
     // Create the asyncronous recognize operation
     $operation = $speech->beginRecognizeOperation(
         $object,
@@ -80,12 +77,6 @@ function transcribe_async_gcs($bucketName, $objectName, $languageCode = 'en-US',
         foreach ($alternatives as $alternative) {
             printf('Transcript: %s' . PHP_EOL, $alternative['transcript']);
             printf('Confidence: %s' . PHP_EOL, $alternative['confidence']);
-            foreach ($alternative['words'] as $wordInfo) {
-                printf('  Word: %s (start: %s, end: %s)' . PHP_EOL,
-                    $wordInfo['word'],
-                    $wordInfo['startTime'],
-                    $wordInfo['endTime']);
-            }
         }
     }
 }