main.py

from __future__ import print_function
from object_watcher.object_watcher import ObjectWatcher
from object_watcher.object_watcher import create_dlib_frontal_face_detector
from simple_audio_recorder.simple_audio_recorder import SimpleAudioRecorder
from stt.simple_stt import SimpleSTT
from websocket import create_connection
import json
import time
import traceback

transcriber = SimpleSTT()

URL_TEMPLATE = "{scheme}://{host}:{port}{path}"


def send_message(message, host="localhost", port=8181, path="/core", scheme="ws"):
    payload = json.dumps({
        "type": "recognizer_loop:utterance",
        "context": "",
        "data": {
            "utterances": [message]
        }
    })
    url = URL_TEMPLATE.format(scheme=scheme, host=host, port=str(port), path=path)
    ws = create_connection(url)
    ws.send(payload)
    ws.close()


def transcribe_and_send(path_to_source):
    try:
        text = transcriber.transcribe(path_to_source)
        print(text)
        send_message(text)
    except Exception:
        traceback.print_exc()

if __name__ == "__main__":
    recorder = SimpleAudioRecorder()

    with SimpleAudioRecorder() as recorder:
        fw = ObjectWatcher(detector=create_dlib_frontal_face_detector())
        fw.register_object_entered_callback(recorder.start_recording, "output.wav")
        fw.register_object_entered_callback(print, "recording!")

        fw.register_object_left_callback(recorder.stop_recording)
        fw.register_object_left_callback(print, "done recording!")
        fw.register_object_left_callback(transcribe_and_send, "output.wav")

        fw.start()

        try:
            while True:
                time.sleep(2)
        except KeyboardInterrupt:
            fw.terminate()
            fw.join()