Sendurance

A centralized fitbit data-gathering tool to allow aggregating data from a fleet of fitbits, for scientific study purposes.

Instead of using python-fitbit, this connects directly with requests-oauthlib. It caches tokens in ./tokens/ and attempts to make the refresh-token dance happen automatically using geckodriver.

Requirements

python3 -m virtualenv .venv
source .venv/bin/activate
pip3 install -r requirements.txt

Install a recent Firefox and put geckodriver somewhere in your path, or put it in the sendurance directory.

Mass downloading

Using the Fleet Downloader

Create a client.json containing your Fitbit API id, secret, and the shared password on your fitbit accounts:

{"id": "XYZ", "secret": "something", "password": "shared"}

Create a csv file with columns ID, Email log-in, First Day, and Last Day. Include rows for every Fitbit.

Then create a work event Excel file. TODO: Document as used by MP2 study.

Create an input.json file for each fleet, to link these two things:

{
    "fleet_file": "/path/to/fleet_data.csv",
    "work_event_file": "/path/to/work_events.xlsx",
    "output_path": "/path/to/output"
}

You can edit fleet_downloader.py to download the data you want. Change the lines at the bottom:

Downloader.save_steps()
Downloader.save_hrv()
Downloader.save_sleep()

Then do

python3 fleet_downloader.py input.json

Other download styles

See afchron_downloader for an example of a study-specific downloader that works from a single wacky excel tracking file.

Run with -v or -vv for more details, including screenshots of geckobrowser during the login dance.

You will need a client.json with id (your client ID) and secret for the Fitbit API, and also a password field with what to use to log into the fitbits.

You can add custom per-fitbit passwords for accounts that did not get the main password in client.json in a hash password-overrides.

Developer details

For development/console stuff, you'll want IPython:

pip3 install ipython

Getting started with code

Store your client id and secret somewhere secure (like in a client.json file, see console.py for an implementation)

Then you can just:

from fitbit_api import FitbitApi
fitbit = FitbitApi('[email protected]', client_id, client_secret)

Console mode

Create a client.json file that looks something like

{"id": "77DFRL", "secret": "deadbeef001010101001111" }

Then run a console with a fitbit account:

python3 console.py [email protected] password

Now there's a "fitbit" object you can play with.

fitbit.activities()
fitbit.get(...)