Talking Cub

A Smart Speaker built for Eastern Nazarene College's fair.

Installation and Initial Set Up

This software makes use of GCloud text-to-speech and speech-to-text APIs. Make sure you create an account and generate a private key to be used.

Raspberry Pi - Speaker

1. Copy a valid GCloud private key as gc_private_key.json in this directory.

2. Install Google Cloud SDK.

3. Install google-cloud-speech and google-cloud-texttospeech packages.

sudo apt-get install portaudio19-dev python-pyaudio python3-pyaudio alsa-tools alsa-utils jackd2
# Start Service
jack_control start

4. Install PyAudio

pip3 install pyaudio

Web App

The web server can either be run locally in the Raspberry Pi or in an external host using docker-compose for both cases. Take into account that in either case, you will have to edit ANSWER_SVC_URL in device/constants.py to match your endpoint's IP. That is ANSWER_SVC_URL = "http://localhost:8000/api/v1/answer/" for the Raspberry Pi, or ANSWER_SVC_URL = "http://{{YOUR HOST'S IP}}/api/v1/answer/".

If running locally in the Pi:

$ docker-compose -f web/docker-compose.pi.yml

If installing in an external host:

$ docker-compose -f web/docker-compose.yml

Run

$ python3 device/app.py

TODO

Webapp

1. Add last_edited to the webapp's inquiry DB and make inquiries in dashboard ordered by date.
2. Make the PIN evaluation (authorization) better. Possibly taking it from an env_variable or docker secret.

Speaker - Device