Well, the previous couple of pages discuss my guess about set_charge_limit and then brianman trying to find the charging current API. All guesses are welcome - it's just hard to do anything better than throwing a bunch of random words at the API and see what returns an error and what doesn't.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Model S REST API
- Thread starter timdorr
- Start date
-
- Tags
- Model S User Interface
Well, the previous couple of pages discuss my guess about set_charge_limit and then brianman trying to find the charging current API. All guesses are welcome - it's just hard to do anything better than throwing a bunch of random words at the API and see what returns an error and what doesn't.
dirkhh
Middle-aged Member
In case someone else trips over this... it appears that the streaming API sometimes sends older data sets again - i.e. I get data with a timestamp that is older than data I had already seen. This confused my algorithm. Now that I'm no longer making assumptions that timestamps are monotonous it appears that it reauthenticates as designed.
hans
P631
This is from memory from a month ago so verify it for accuracy ;-) If you complete the full 120 second HTTP long poll you get a new timestamp each time. If you terminate the HTTP session early from your client side app and then try polling again you will get the same old timestamp you got on the previous poll. I think you will get it over and over again for up to 5 minutes or so if I remember correctly.
P.S. I wish they would just use websockets or MQTT or something, anything better and more pubsub or connection oriented than COMET/Long Polling. Lowest common denominator wins again I guess.
hans
P631
Also, the timestamps are in Universal Time (aka GMT) so don't assume they are using the local timezone.
dirkhh
Middle-aged Member
That's strange. I get them all the time. And not only when canceling the http poll as hans suggests above. Also, I get missing segments (there will be something like 10-30 seconds with no data whatsoever...)
I guess I should run the streaming.js in parallel with my app to see if we get the same data... since I'm an absolute JS newbie I wouldn't be surprised at all if there were tons of bugs hiding in my code.
Last edited:
Vger
Active Member
Unix epoch time x1000. So milliseconds since Thu, 1 Jan 1970 00:00:01 UTC.
I use Epoch Converter - Unix Timestamp Converter a lot.
Vger
Active Member
Thanks. Figured it was something close to that. You saved me some trial and error.
Tesla's streaming service doesn't send the proper no-cache http headers in the response, so if using straight http gets, your client should cache the responses so you'll get duplicate data. One workaround for this is to add something unique to the querystring that tesla will ignore.
The other WindowsPhone issue with out of memory exceptions, the workaround is to set the connection timeout to less than 60 seconds. I don't have access to the source at the moment to remember what exactly I did for these 2 in my app
dirkhh
Middle-aged Member
I now have split my tool into two parts and things are working much better.
The backend is a small modification to streaming.js that stores all the data sets in a mongodb instead of in a flat file.
The frontend uses node.js to create a web server and Google maps to visualize the data in the database. It can replay past driving or show the life diving with automatic (AJAX) updates to the map.
Besides the data that the streaming API is giving us, what other information do people store about their cars? One could poll temperature data for example... So I wonder what others are doing.
Hans - are you interested in pull requests to add this to your project, or should I turn it into a separate project?
The backend is a small modification to streaming.js that stores all the data sets in a mongodb instead of in a flat file.
The frontend uses node.js to create a web server and Google maps to visualize the data in the database. It can replay past driving or show the life diving with automatic (AJAX) updates to the map.
Besides the data that the streaming API is giving us, what other information do people store about their cars? One could poll temperature data for example... So I wonder what others are doing.
Hans - are you interested in pull requests to add this to your project, or should I turn it into a separate project?
hans
P631
Sounds great dirkhh! I would be happy to incorporate your stuff into the existing project. I didn't start it to be "my" project anyway so the more contributors the better.
dirkhh
Middle-aged Member
Looking at integrating things into streaming.js I ran into a slight problem...
You use request(...).pipe(fs.createWriteStream( argv.file, {'flags': 'a'} )) to directly pipe things into the output file.
I need to use request(...).on('data', function(data){...}) to write things into the database.
I don't know JS well enough to know how to switch between those two without a ton of code duplication. My guess is that the .pipe() could be reimplemented as .on('data',...) - but I don't know exactly how and I don't know if you'd like this.
The alternative would of course be to simply add my file as a new JS file in the repository, but since most of it is identical to the existing file that seems like a waste as well...
Any advice?
superraz375
Member
I just received my car today and started digging around!
I found some extra option code values that are not documented as far as I can tell.
I took my best guess as to what they could mean.
I found some extra option code values that are not documented as far as I can tell.
I took my best guess as to what they could mean.
Code:
[FONT=Times]PK00, (parking sensors)[/FONT]
[FONT=Times]CW00, (cold weather package)[/FONT]
[FONT=Times]SP00, (security protection??)[/FONT]
[FONT=Times]YF00, (???)[/FONT]
[FONT=Times]X009, (???)[/FONT]
[FONT=Times]X027, (???)[/FONT]
[FONT=Times]X031, (???)[/FONT]
[FONT=Times]BS00 (???)
[/FONT]
araxara
X-LR M3AWD, MYLR
Strictly speaking this isn't "known" but based on the API surface area and behavior, it appears that it isn't streaming until requested and then it times out (stops streaming) every 15-30 minutes until a new request is sent.
Similar threads
