Musiquity

DSCI 551 Project: Generates a dynamic, personalized workout mix based on streamed heartrate. Currently in Proof-of-Concept stage.

Prerequisites

Please ensure the following software is available:

• Docker and docker-compose

• npm

Please ensure the following python packages are available:

• numpy

• statsmodels

• django

• django-cors-headers

• django-sslserver

• influxdb

• django-mysql

• spotipy

Setting up

• Create a docker volume called influx_volume and mysql_volume for persistent data storage.

• In the musiquity_frontend folder run npm install.

• Run python recursively_dcu.py in the root directory. This brings up the redis, Influx and MySQL dockers.

• Create a file called config.json in web_api/web_api structured as follows:

{
    "mysql_data": {
        "username": "mysql_username",
        "password": "mysql_password"
    }
}

• In web_api/daemons, create a file called api.json that is structured as follows:

{
    "api": {
        "key": "your_spotify_application_key",
        "secret": "your_spotify_application_secret"
    }
}

• In web_api/auth_service create a file called client_id.json that contains the oauth authentication details for Google authentication. (Note: Ensure that your oauth2 client can read activity and fitness data at the very least)

• Go to the web_api folder and run python manage.py createsuperuser and create a superuser for your instance.

• Then run python manage.py runsslserver. The SSL server is required for Google's OAuth2.

• In the musiquity_frontend folder, start the UI server with npm start.

Initializing data

• Open https://localhost:8080/admin/ to open the admin console and login as the superuseryou just created.

• Create a user under the 'Users' group.

• Under the Backend Service group, create the appropriate models for your cutoffs (You can poll data and use spark for this later)

• Create your "seed lists". The tracks in your seed lists must be spotify URIs for tracks separated by a comma.

• Copy your username and replace the default with it in helpers/api_querier/api_querier.py and helpers/spoofers/influxdb_spoofer_delay.py.

• Now go to https://localhost:8080/auth and authenticate. This should return an "authorized" message with the location of the token.txt file

Adding data to influxDB

• When you have heart rate data recorded from a workout, run api_querier.py <<token.txt absolute path>> in helpers/api_querier. This will poll data and store it in a points.txt file.

• Due to issues with Google Fit Streaming, for demo purposes we shall simulate the same workout. Run post_simulator.py. This shall simulate data streaming.

Getting predictions

• In web_api/daemons run daemons.py.

• If no errors come up there should be data in redis.

• Open localhost:3000 and input your username (whatever you used in api_querier.py). You should see your current heart rate (with a little bit of trend) and suggestions.