[google_cloud_texttospeech] Add cache option #267
Conversation
| '--'.join([md5, language_code, name, str(speaking_rate)]) | ||
| + '.mp3') | ||
| if os.path.exists(cache_filename): | ||
| shutil.copy(cache_filename, args.output) |
There was a problem hiding this comment.
how about rospy.loginfo something like "Using caching sound file ('/home/..../....wav') for 'Hello'"
There was a problem hiding this comment.
In current implementation, text2wave is not intended to be used in ROS because not called rospy.init_node('nodename') at first.
I agree with your idea of adding logging.
Is it okay to add rospy.init_node at the beginning of text2wave's main function?
There was a problem hiding this comment.
Oh, we need to use print here...
There was a problem hiding this comment.
| VoiceSelectionParams = google.cloud.texttospeech.types.VoiceSelectionParams | ||
|
|
||
| cache_enabled = os.environ.get( | ||
| 'GOOGLE_CLOUD_TEXTTOSPEECH_CACHE_ENABLED', False) |
There was a problem hiding this comment.
This could be param.
PATH, PYTHONENCODING, GOOGLE_APPLICATION_CREDENTIALS must be env, because that is used in google.cloud.texttospeech modules, but GOOGLE_CLOUD_TEXTTOSPEECH_CACHE_ENABLED is only used in this file.
I also feel set cache enable could be True, by default.
There was a problem hiding this comment.
I found the solution has a problem.
text2wave will be called at sound_play node using os.system https://github.com/ros-drivers/audio_common/blob/master/sound_play/scripts/soundplay_node.py#L267 .
I want to specify namespace to pass as ROS param, but I can't specify it correctly just by calling it in os.system.
Is there a way to correctly give the ROS namespace to the text2wave called in os.system?
There was a problem hiding this comment.
I see, then, using env variable make sense. I think using env is ok.
There was a problem hiding this comment.
OK.
Also, I modified the code for caching to default behavior.
|
@iory can you solve the conflict? |
|
I solved the conflicts. Please check the commit when you have time. |
What is this
google cloud texttospeechwaits for a reply of API communication.It takes a little time and money for the API.
This PR reducess them by adding cache functionality.
The cache is enabled by default.