fixed #42, fixed #43, got sample rate programmatically

tatt/config.py

@@ -8,14 +8,6 @@ BUCKET_NAME_FMTR_TRANSCRIPT = 'tatt-transcript-{}'
 BUCKET_NAME_FMTR_TRANSCRIPT_GOOGLE = 'tatt_transcript_{}'
 
 
-def GOOGLE_SPEECH_USE_ENHANCED():
-    enhanced = os.getenv('GOOGLE_SPEECH_USE_ENHANCED')
-    if enhanced and enhanced == 'true':
-        return True
-    else:
-        return False
-
-
 if os.getenv('AWS_CONFIG_FILEPATH'):
     AWS_CONFIG_FILEPATH = Path(os.getenv('AWS_CONFIG_FILEPATH'))
 else:

tatt/helpers.py

@@ -124,10 +124,18 @@ def get_transcription_jobs_dict():
 
 
 def get_num_audio_channels(filepath):
+    return get_media_info(filepath).channels
+
+
+def get_sample_rate(filepath):
+    return get_media_info(filepath).samplerate
+
+
+def get_media_info(filepath):
     if isinstance(filepath, pathlib.PurePosixPath):
         filepath = str(filepath)
     with audioread.audio_open(filepath) as f:
-        return f.channels
+        return f
 
 
 def shell_call(command):

tatt/transcribe.py

@@ -88,15 +88,33 @@ def status(job_name):
 
 
 @cli.command()
+@click.option('--punctuation', is_flag=True, default=True, 
+              help='only for Google Speech, defaults to True')
+@click.option('--speaker-id', is_flag=True, default=True, 
+              help='only for Google Speech, defaults to True')
+@click.option('--model', default='phone_call', 
+              help='only for Google Speech, defaults to "phone_call"')
+@click.option('--use-enhanced', is_flag=True, default=True,
+              help='only for Google Speech, defaults to True')
 @click.argument('media_filepath', type=str)
 @click.argument('service_name', type=str)
-def this(media_filepath, service_name):
+def this(media_filepath, service_name, punctuation, speaker_id, model,
+         use_enhanced):
     """Sends a media file to be transcribed."""
+    if service_name == 'google':
+        transcribe_kwargs = dict(
+            enable_automatic_punctuation=punctuation,
+            enable_speaker_diarization=speaker_id,
+            model=model,
+            use_enhanced=use_enhanced,
+            )
+    else:
+        transcribe_kwargs = {}
     try:
         service = get_service(service_name)
     except KeyError as e:
         raise click.ClickException(
-            f'No such service! {print_all_services(print_=False)}')
+            f'No such service! {helpers.make_string_all_services()}')
 
     try:
         s = service(media_filepath)
@@ -107,7 +125,7 @@ def this(media_filepath, service_name):
       f'Okay, transcribing {media_filepath} using {service_name}...')
 
     try:
-        job_num = s.transcribe()
+        job_num = s.transcribe(**transcribe_kwargs)
     except exceptions.AlreadyExistsError as e:
         raise click.ClickException(str(e))
     click.echo(f'Okay, job {job_num} is being transcribed.  Use "get" '

tatt/vendors/google.py

@@ -30,7 +30,7 @@ def _check_for_config():
 class Transcriber(TranscriberBaseClass):
 
     SUPPORTED_FORMATS = ['flac']
-    cost_per_15_seconds = .009
+    cost_per_15_seconds = [.004, .006, .009]
     no_config_error_message = (
             'Please sign up for the Google Speech-to-Text API '
             'and put the path to your credentials in an '
@@ -80,9 +80,14 @@ class Transcriber(TranscriberBaseClass):
     def check_for_config() -> bool:
         return _check_for_config()
 
-    def transcribe(self) -> str:
+    def upload_file_if_too_big(self):
+        """10MB limit as of Mar 7, 2019"""
+        pass
+
+    def transcribe(self, **kwargs) -> str:
         self.convert_file_format_if_needed()
-        self._request_transcription()
+        self.upload_file_if_too_big()
+        self._request_transcription(**kwargs)
 
     def _check_if_transcript_exists(self, transcript_name=None):
         return storage.Blob(
@@ -93,16 +98,17 @@ class Transcriber(TranscriberBaseClass):
     def _request_transcription(
             self, 
             language_code='en-US',
-            # model='video',
+            enable_automatic_punctuation=True,
+            enable_speaker_diarization=True,
+            model='phone_call',
+            use_enhanced=True,
             ) -> str:
         """Returns the job_name"""
         if self._check_if_transcript_exists():
             raise exceptions.AlreadyExistsError(
                 f'{self.basename} already exists on {NAME}')
         num_audio_channels = helpers.get_num_audio_channels(self.filepath)
-
+        sample_rate = helpers.get_sample_rate(self.filepath)
-        use_enhanced = config_mod.GOOGLE_SPEECH_USE_ENHANCED()
-        print(use_enhanced)
 
         with io.open(self.filepath, 'rb') as audio_file:
             content = audio_file.read()
@@ -110,19 +116,16 @@ class Transcriber(TranscriberBaseClass):
 
         config = speech.types.RecognitionConfig(
             encoding=speech.enums.RecognitionConfig.AudioEncoding.FLAC,
-            sample_rate_hertz=44100,
+            sample_rate_hertz=sample_rate,
             audio_channel_count=num_audio_channels,
             enable_separate_recognition_per_channel=True,
             enable_word_confidence=True,
             enable_word_time_offsets=True,
             language_code=language_code,
-            enable_automatic_punctuation=True,
+            enable_automatic_punctuation=enable_automatic_punctuation,
-            enable_speaker_diarization=True,
+            enable_speaker_diarization=enable_speaker_diarization,
-            # not clear whether this has to be 'phone_call' in order to
+            model=model,
-            # use_enhanced
-            model='video',
             use_enhanced=use_enhanced,
-            # model=model,
             )
 
         self.operation = self.speech_client.long_running_recognize(config,