Skip to content

Instantly share code, notes, and snippets.

Created December 4, 2017 00:40
Show Gist options
  • Save anonymous/e3bc3f72cb9a9b8de0dae6ff3b7e3ba6 to your computer and use it in GitHub Desktop.
Save anonymous/e3bc3f72cb9a9b8de0dae6ff3b7e3ba6 to your computer and use it in GitHub Desktop.
Use the google assistant SDK to trigger custom actions in Home Assistant using JSON over MQTT, the adapter intent parser python library, and picoTTS for audio responses
#!/usr/bin/env python
# Copyright (C) 2017 Google Inc.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from __future__ import print_function
import argparse
import os.path
import json
import subprocess
import paho.mqtt.client as mqtt
import google.oauth2.credentials
from google.assistant.library import Assistant
from google.assistant.library.event import EventType
from google.assistant.library.file_helpers import existing_file
import intent_parser
BROADCAST_TOPIC = "assistant/broadcast"
assistant_room = None
mqtt_client = None
def on_connect(client, userdata, flags, rc):
mqtt_client.subscribe('assistant/{}/tts'.format(assistant_room), qos=2)
mqtt_client.subscribe('assistant/broadcast', qos=2)
def on_message(client, userdata, msg):
message = msg.payload.strip().decode('utf-8')
if msg.topic == BROADCAST_TOPIC:
payload_split = message.split(":")
message = payload_split[0].strip()
source = payload_split[1].strip()
if source == assistant_room:
message = "Your message has been shared"
pico_command = ['pico2wave', '-l', 'en-GB', '-w', 'tmp.wav', message]
subprocess.call(pico_command)
subprocess.call(['aplay', 'tmp.wav'])
def process_event(event, assistant):
if event.type == EventType.ON_CONVERSATION_TURN_STARTED:
subprocess.call(['aplay', 'hotword.wav'])
if event.type == EventType.ON_RECOGNIZING_SPEECH_FINISHED:
text = event.args.get('text')
print("Raw: {}".format(text))
intent, massaged_text = intent_parser.parse_intent(text)
if intent is not None:
assistant.stop_conversation()
intent['raw'] = massaged_text
intent['source'] = assistant_room
mqtt_client.publish("assistant/{}/intent".format(assistant_room), payload=json.dumps(intent), qos=2)
def main():
global assistant_room, mqtt_client
parser = argparse.ArgumentParser(
formatter_class=argparse.RawTextHelpFormatter)
parser.add_argument('--credentials', type=existing_file,
metavar='OAUTH2_CREDENTIALS_FILE',
default=os.path.join(
os.path.expanduser('~/.config'),
'google-oauthlib-tool',
'credentials.json'
),
help='Path to store and read OAuth2 credentials')
parser.add_argument("broker_ip", type=str)
parser.add_argument("broker_port", type=int)
parser.add_argument("room_name", type=str)
args = parser.parse_args()
with open(args.credentials, 'r') as f:
credentials = google.oauth2.credentials.Credentials(token=None,
**json.load(f))
broker_ip = args.broker_ip
broker_port = args.broker_port
assistant_room = args.room_name
mqtt_name = "_".join([assistant_room, "assistant"])
mqtt_client = mqtt.Client(client_id=mqtt_name, protocol=mqtt.MQTTv31)
mqtt_client.on_connect = on_connect
mqtt_client.on_message = on_message
mqtt_client.username_pw_set(mqtt_name)
mqtt_client.connect(broker_ip, broker_port)
mqtt_client.loop_start()
with Assistant(credentials) as assistant:
for event in assistant.start():
process_event(event, assistant)
if __name__ == '__main__':
main()
__author__ = 'seanfitz'
"""
A sample program that uses multiple intents and disambiguates by
intent confidence
try with the following:
PYTHONPATH=. python examples/multi_intent_parser.py "what's the weather like in tokyo"
PYTHONPATH=. python examples/multi_intent_parser.py "play some music by the clash"
"""
import json
import sys
from adapt.entity_tagger import EntityTagger
from adapt.tools.text.tokenizer import EnglishTokenizer
from adapt.tools.text.trie import Trie
from adapt.intent import IntentBuilder
from adapt.parser import Parser
from adapt.engine import IntentDeterminationEngine
tokenizer = EnglishTokenizer()
trie = Trie()
tagger = EntityTagger(trie, tokenizer)
parser = Parser(tokenizer, tagger)
engine = IntentDeterminationEngine()
entity_json = {
"Normal":{
"Room": [
"living room",
"bathroom",
"kitchen",
"garage",
"bedroom",
"office",
"house"
],
"LightObject": [
"lights",
"light"
],
"LampObject": [
"lamp",
"lamps",
"lampes"
],
"MediaObject": [
"TV"
],
"LevelVerb": [
"dim",
"turn up",
"turn down"
],
"PowerVerb": [
"turn on",
"turn off"
],
"MediaVerb": [
"mute",
"pause",
"resume"
],
"BroadcastVerb": [
"broadcast",
"announce"
],
"TalkVerb": [
"say",
"tell"
],
"ListVerb": [
"add",
"remove",
"read"
],
"ListType": [
"to do",
"to dos",
"grocery",
"groceries",
"shopping"
],
"AllModifier": [
"all"
],
"GoodnightVerb": [
"good night",
"nighty night",
"commence bed time sequence"
]
},
"Regex": [
"(?P<Percentage>[0-9]+%)",
"add (?P<ListItemAdd>.*) to",
"remove (?P<ListItemRemove>.*) from",
"its (?P<SceneEvent>.*) time",
"the (?P<InputName>.*) input"
]
}
for entity, values in entity_json['Normal'].items():
for value in values:
engine.register_entity(value, entity)
for regex_entity in entity_json['Regex']:
engine.register_regex_entity(regex_entity)
intents = []
intents.append(\
IntentBuilder("LevelIntent")\
.require("LevelVerb")\
.optionally("AllModifier")\
.optionally("Room")\
.one_of("LightObject", "LampObject", "MediaObject")\
.require("Percentage")\
.build())
intents.append(\
IntentBuilder("PowerIntent")\
.require("PowerVerb")\
.optionally("AllModifier")\
.optionally("Room")\
.one_of("LightObject", "LampObject", "MediaObject", "InputName")\
.optionally("Percentage")\
.build())
intents.append(\
IntentBuilder("MediaIntent")\
.require("MediaVerb")\
.optionally("AllModifier")\
.optionally("Room")\
.optionally("MediaObject")\
.build())
intents.append(\
IntentBuilder("ListIntent")\
.require("ListVerb")\
.optionally("ListItemAdd")\
.optionally("ListItemRemove")\
.require("ListType")\
.build())
intents.append(\
IntentBuilder("TalkIntent")\
.require("TalkVerb")\
.require("Room")\
.build())
intents.append(\
IntentBuilder("SceneIntent")\
.require("SceneEvent")\
.build())
intents.append(\
IntentBuilder("GoodnightCommand")\
.require("GoodnightVerb")\
.build())
for intent in intents:
engine.register_intent_parser(intent)
def massage_text(val):
if isinstance(val, str):
val = val.replace("'", "")
return val
else:
return val
def massage_json(json_val):
massaged_json = {}
for key, val in json_val.items():
massaged_json[massage_text(key)] = massage_text(val)
return massaged_json
def parse_intent(val):
val = massage_text(val)
results = engine.determine_intent(val)
for obj in results:
if obj is not None and obj.get('confidence') > 0:
return (massage_json(obj), val)
return (None, val)
if __name__ == "__main__":
raw = ' '.join(sys.argv[1:])
intent, text = parse_intent(raw)
if intent is not None:
intent['raw'] = text
print(json.dumps(intent, indent=4))
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment