google_assistant.py

   1 #!/usr/bin/env python3
   2
   3 import logging
   4 from typing import NamedTuple
   5
   6 import requests
   7 import speech_recognition as sr  # type: ignore
   8
   9 import config
  10
  11 logger = logging.getLogger(__name__)
  12
  13 parser = config.add_commandline_args(
  14     f"Google Assistant ({__file__})",
  15     "Args related to contacting the Google Assistant",
  16 )
  17 parser.add_argument(
  18     "--google_assistant_bridge",
  19     type=str,
  20     default="http://kiosk.house:3000",
  21     metavar="URL",
  22     help="How to contact the Google Assistant bridge"
  23 )
  24 parser.add_argument(
  25     "--google_assistant_username",
  26     type=str,
  27     metavar="GOOGLE_ACCOUNT",
  28     default="scott.gasch",
  29     help="The user account for talking to Google Assistant"
  30 )
  31
  32
  33 class GoogleResponse(NamedTuple):
  34     success: bool
  35     response: str
  36     audio_url: str
  37     audio_transcription: str
  38
  39     def __repr__(self):
  40         return f"""
  41 success: {self.success}
  42 response: {self.response}
  43 audio_transcription: {self.audio_transcription}
  44 audio_url: {self.audio_url}"""
  45
  46
  47 def tell_google(cmd: str, *, recognize_speech=True) -> GoogleResponse:
  48     return ask_google(cmd, recognize_speech=recognize_speech)
  49
  50
  51 def ask_google(cmd: str, *, recognize_speech=True) -> GoogleResponse:
  52     payload = {
  53         "command": cmd,
  54         "user": config.config['google_assistant_username'],
  55     }
  56     url = f"{config.config['google_assistant_bridge']}/assistant"
  57     r = requests.post(url, json=payload)
  58     success = False
  59     response = ""
  60     audio = ""
  61     audio_transcription = ""
  62     if r.status_code == 200:
  63         j = r.json()
  64         success = bool(j["success"])
  65         response = j["response"] if success else j["error"]
  66         audio = f"{config.config['google_assistant_bridge']}{j['audio']}"
  67         if recognize_speech:
  68             recognizer = sr.Recognizer()
  69             r = requests.get(audio)
  70             if r.status_code == 200:
  71                 raw = r.content
  72                 speech = sr.AudioData(
  73                     frame_data=raw,
  74                     sample_rate=24000,
  75                     sample_width=2,
  76                 )
  77                 audio_transcription = recognizer.recognize_google(
  78                     speech,
  79                 )
  80     else:
  81         logger.error(
  82             f'HTTP request to {url} with {payload} failed; code {r.status_code}'
  83         )
  84     return GoogleResponse(
  85         success=success,
  86         response=response,
  87         audio_url=audio,
  88         audio_transcription=audio_transcription,
  89     )