stt-watson
Continuous speech to text using watson in python with websocket and record from microphone
Requirements
- Python 2.7
- Pip
- portaudio, can be installed with
brew install portaudio(mac) orapt-get install portaudio19-dev(linux)
Installation
Install with pip: pip install stt-watson
Run the playground
Simply run in command line: stt-watson
At the first launch it will create a config file located to ~/.config-stt-watson.yml and ask you your watson credentials
Usage for developers
Bootstrap example:
from stt_watson.SttWatson import SttWatson
from stt_watson.SttWatsonAbstractListener import SttWatsonAbstractListener
"""
Example of listener to use data given by stt-watson (stt-watson notify hypothesis to his listeners when he receive it)
Hypothesis format:
{
'confidence': '0.1' // confidence of the sentence or words if exist
'transcript': 'the transcription of your voice'
}
"""
class MyListener(SttWatsonAbstractListener):
def __init__(self):
pass
"""
This give hypothesis from watson when your sentence is finished
"""
def listenHypothesis(self, hypothesis):
print "Hypothesis: {0}".format(hypothesis)
"""
This give the json received from watson
"""
def listenPayload(self, payload):
print(u"Text message received: {0}".format(payload))
"""
This give hypothesis from watson when your sentence is not finished
"""
def listenInterimHypothesis(self, interimHypothesis):
print "Interim hypothesis: {0}".format(interimHypothesis)
myListener = MyListener()
sttWatson = SttWatson('watson_user', 'watson_password')
sttWatson.addListener(myListener)
sttWatson.run()