openhumans / oh-fitbit-integration Goto Github PK

There were discussions of getting walking speed data into OH and after looking at the Fitbit API it looks like there are 2 different API endpoints that aren't being utilized that contain activity speed information:

1. "Activities list" which contains a single speed datapoint for an entire activity
2. "Activity type" which returns more detailed information about an activity including min & max speed per "level" https://dev.fitbit.com/build/reference/web-api/activity/#get-activity-type

We should add in the latter to our fetching to see how much data accumulates for someone who has years worth of activity data.

Looking at this:
https://github.com/OpenHumans/oh-fitbit-integration/blob/master/datauploader/tasks.py

... it seems that the endpoint is going to https://api.fitbit.com/1/user

That prefix is still used for the following:

• Activity endpoints which have the a format https://api.fitbit.com/1/user/[user-id]/activities/tracker/...
• Heart endpoints which have the a format https://api.fitbit.com/1/user/[user-id]/activities/heart/...
• Body & weight endpoints which have the a format https://api.fitbit.com/1/user/[user-id]/activities/body/...

However from the sleep documentation: https://dev.fitbit.com/build/reference/web-api/sleep/

... I see the endpoint seems to be updated to be like https://api.fitbit.com/1.2/user/-/sleep/. (Note the "1.2".)

I was looking into this because I'd like to be collecting the sleep stages data (minute-by-minute information about which "stage" of sleep is occurring), which appears to be available in this new endpoint. (I think the old one is working but deprecated: https://dev.fitbit.com/build/reference/web-api/sleep-v1/ )

Currently our data import task (https://github.com/OpenHumans/oh-fitbit-integration/blob/master/datauploader/tasks.py) uses only the presence of a given key to determine whether we already have the corresponding data or not.

This is problematic as a given year is the top-level key in our JSON. So e.g. once we have fetched any data for 2018 we will ignore all new data and not import further data. This needs to be fixed by checking whether the data for a given key is complete and re-running the import for that key if its not.

See my public fitbit data at https://www.openhumans.org/member/gedankenstuecke for an example of this behaviour.

Right now you can push the update data button as many times as you want and I think it'll also trigger that update as often as you push the button?

Would be good if we could fully implement the cool down (through last_submitted as for moves)