Using the AT&T Speech to Text API With Python

Question

I am trying to use the AT&T speech to text API. So far, I can get the access token

def get_access_token(client_id, client_secret):
headers = {'Content-Type': 'application/x-www-form-urlencoded', 'Accept': 'application/json'}

data = {'client_id': client_id, 'client_secret': client_secret, 'scope': 'SPEECH',
        'grant_type': 'client_credentials'}

response = requests.post(oauth_url, data=data, headers=headers)
return response.text

So far, this is what I have for sending the audio file to get the json response:

def get_text_from_file(file, access_token):
headers = {'Authorization': 'Bearer ' + access_token, 'Accept': 'application/json', 'Content-Type': 'audio/wav',
           'X-SpeechContext': 'Generic', 'Connection': 'Keep-Alive'}

But I am not sure how to send the file. Can anyone help?

Answer 1

This is what I just got working, using the requests library, and a couple of other resources I'll link below

import json
import requests

class ATTSpeech:
    CLIENT_ID = "SOME"
    CLIENT_SECRET = "ID"
    TOKEN = None

    def __init__(self, *args, **kwargs):
        self.get_token()


    def get_token(self):
        # Get Access Token via OAuth.
        # https://matrix.bf.sl.attcompute.com/apps/constellation-sandbox
        response = requests.post("https://api.att.com/oauth/token", {
            "client_id": self.CLIENT_ID,
            "client_secret": self.CLIENT_SECRET,
            "grant_type": "client_credentials",
            "scope": "SPEECH,STTC"
        })
        content = json.loads(response.content)
        self.TOKEN = content["access_token"]


    def text_from_file(self, path):

        with open(path, 'rb') as f:
            response = requests.post("https://api.att.com/speech/v3/speechToText",
                headers = {
                    "Authorization": "Bearer %s" % self.TOKEN,
                    "Accept": "application/json",
                    "Content-Type": "audio/wav",
                    "X-SpeechContext": "Generic",
            }, data=f)
        content = json.loads(response.content)
        return content

https://sites.google.com/site/brssbrss/attspeechapi

http://changingjasper.blogspot.com/2014/06/making-jasper-use-at-speech-api.html

Usage is something like the below, assuming you saved this file as ATTEngine

from ATTEngine import ATTSpeech
a = ATTSpeech()
a.text_from_file('/Users/issackelly/Desktop/here.wav')