API
37 watchers
Aug 2014
11:04am, 25 Aug 2014
14,793 posts
|
fetcheveryone
Two updates to report: 1) I've added GPX support to training/import 2) There's some documentation here: http://www.fetcheveryone.com/api-doc.php - at the moment it just lists the accepted / required parameters, but I'll add to it as we go. It's auto-generated, so as I add new methods, they'll show up automatically. |
Aug 2014
10:18pm, 25 Aug 2014
68 posts
|
jonbev
That would be sweet. I spoke to garmin about their api recently for use with GoldenCheetah and they wanted 5k! WTF!
|
Aug 2014
8:22pm, 26 Aug 2014
99 posts
|
nuttyficket
Small bit of API feedback: if I upload iffy run data (I didn't url encode it correctly), I still get a response outcome of SUCCESS even though no run has been added to my log - I'd expect a FAIL. The returned run id is still an empty field in this case, though.
|
|
Aug 2014
10:35pm, 26 Aug 2014
14,799 posts
|
fetcheveryone
Ok, I'll see what I can do about that. To all watchers of this thread: join the developer group as well please - link on the right. Ta. I also have some Python scripts to upload tomoz (thanks Badger). Would be great to have some Java and Objective C classes to demo too. Anyone? |
Aug 2014
8:19am, 27 Aug 2014
8 posts
|
BFC
I hit the same issue nuttyficket did with incorrectly encoded and compressed data returning success. Another request/suggestion: can uid be an optional argument? You seen to already know who we are so it seems unnecessary to specify our own uid in requests. Regarding other implementations: I've got a ruby implementation which needs a little tidying and then I'll publish the GitHub (you should bung your examples up on GitHub too;-)) repo here. |
|
Aug 2014
9:28am, 27 Aug 2014
14,800 posts
|
fetcheveryone
Morning  I've made some adjustments. If the import code can't process the file i.e. can't create a valid XML DOM from the string you supply, it will return a fail message now. BFC - you're right in that technically we could do without the uid when you're importing to your own account - but I validate it against the list of users who have approved an 'app' to work on their behalf. |
|
Aug 2014
9:40am, 27 Aug 2014
14,801 posts
|
fetcheveryone
Badger's Python examples are now live - see the link in the thread description over on the right.
|
Aug 2014
11:28am, 27 Aug 2014
4,384 posts
|
Badger
Just run some code that pulls the last few activities & identifies the newest, then runs through tcx files in a directory, identifies newer ones than are in the log and uploads them as the appropriate type. With a UI and some improvements to detecting a fresh upload, this might actually be useful. Putting them on GitHub is a good call, I reckon. OK by you, Fetch? |
|
Aug 2014
11:30am, 27 Aug 2014
14,807 posts
|
fetcheveryone
My only concern at the moment is what will happen if someone imports a trillion files. But hey, go for it
|
|
Aug 2014
12:08pm, 27 Aug 2014
4,385 posts
|
Badger
I imported two, seeing as that was all the un-uploaded files I had Anything you can do to choke things on your end? Where my (current) code would run into trouble is if somebody made a new account and tried to upload all their existing data; as it stands, it won't upload anything older than the newest, even if that isn't in the database already. This is probably quite a useful way for it to be dumb. |
Related Threads
-
Coders
Oct 2024
-
Pi Club
Jan 2022
-
Android App
Apr 2021
-
Coders
Aug 2020
-
Wireframe tools
May 2019
-
Tapiriik?
Mar 2019
-
Oracle club
Sep 2018
-
Raspberry Pi
Jan 2018
-
Calling All Coders
Jan 2014
-
Upload from apple watch
Jul 2024