search

petergardfjall / garminexport

Garmin Connect activity exporter and backup tool
Apache License 2.0
507 stars 83 forks source link

export currently not working #23

Closed frechina closed 6 years ago

frechina commented 6 years ago

It seems garmin removed https://connect.garmin.com/legacy/session

503 Service Unavailable No server is available to handle this request.

Anyone fixing this?

jedie commented 6 years ago

Very sad :(

dlenski commented 6 years ago

I have no trouble logging in, but then the activitySummary endpoint fails. As usual, Garmin has monkeyed with their available API endpoints in various ways.

Also as usual, Tapiriik has updated to make itself work with the latest API endpoints, and cribbing off of its source code would be a good way to get this working again:

https://github.com/cpfair/tapiriik/blob/master/tapiriik/services/GarminConnect/garminconnect.py

(It looks like the activitySummary endpoint has moved to https://connect.garmin.com/modern/proxy/activity-service/activity/{activityId}.)

ajaiantilal commented 6 years ago

I did run into this issue in May, but when i updated it to the latest Apr commit the issue went away. This code still seems to be working for me and i can download all my activities

jedie commented 6 years ago

I successfully used https://github.com/pe-st/garmin-connect-export yesterday to download all my gpx files from garmin connect...

I think it makes sense to work together.

petergardfjall commented 6 years ago

I believe that this should be fixed with 01a08e1. Apparently attempting to download a .fit file for an activity without a .fit file now results in a 500 response (used to be 404, which makes a lot more sense).

A lot of different issues seem to be implied by comments in this issue, but I'm unsure of which are actual issues with the code and which are caused by shaky Garmin services. Since I've just run a full backup of my own data without issues I'm assuming things are working again. Feel free to re-open (or open a new issue) if your problem persists.

@dlenski The code should already have been updated to make use of the new endpoint, so that cannot be the reason things are going awry.