Continuous speech to text using watson in python with websocket and record from microphone
- Python 2.7
- Pip
- portaudio, can be installed with
brew install portaudio (mac) orapt-get install portaudio19-dev(linux)
Install with pip:pip install stt-watson
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
Bootstrap example:
fromstt_watson.SttWatsonimportSttWatsonfromstt_watson.SttWatsonAbstractListenerimportSttWatsonAbstractListener"""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'}"""classMyListener(SttWatsonAbstractListener):def__init__(self):pass""" This give hypothesis from watson when your sentence is finished """deflistenHypothesis(self,hypothesis):print"Hypothesis: {0}".format(hypothesis)""" This give the json received from watson """deflistenPayload(self,payload):print(u"Text message received: {0}".format(payload))""" This give hypothesis from watson when your sentence is not finished """deflistenInterimHypothesis(self,interimHypothesis):print"Interim hypothesis: {0}".format(interimHypothesis)myListener=MyListener()sttWatson=SttWatson('watson_user','watson_password')sttWatson.addListener(myListener)sttWatson.run()