API Documentation and a few questions

So a couple of notes,

I recently had a ticket logged for my weewx driver that made me go on quite the long search for clarity. As part of that I opened a ticket with weatherflow since their documentation linked on this forum for the API (hosted on github I think) doesn’t match Quick Start . In that ticket they referenced the apiddocs site as the correct source of truth.

Not sure how to request the administrators of the forum to update those links? Thoughts???

Second, part of what I’m working through is a pretty big discrepancy between rain readings and nearby stations for one of the users. He’s seeing 2x - 5x the amount of rain reported from his as other nearby non-tempest stations. I’ve read about the need to calibrate and avoid vibrations, he’s going to look into that but I’m curious about a few other things… To be clear, I use the websocket API and my driver is at GitHub - livysdad27/tempestWS: Weatherflow Tempest WebSocket Driver if you want to take a gander.

  1. What is the real difference between obs_st[12] (rain accumulation in mm over last 60 seconds) and obs_st[19] (nearcast rain). Is the second just an adjusted version of 12 (accumulation over last 60 seconds) or is it a daily counter or???

  2. Has anyone ever tested the rest api vs websocket to see if they have the same obs definitions???

Thoughts?

Many thanks!

Billy

Nearcast IS the WF ‘corrected’ rain totals, based on their ‘voodoo magic’ observations of other rain source ‘near’ the area where the station in question is located.

A bit more about the rain items:

1 Like

Thanks btst,

That’s helpful. I’m trying however to find the definition of the websocket obs_st[19] definition for “Nearcast Rain Accumulation”. I want to validate that it’s returning “since last observation” in the same way that field 12 is. I may just open another ticket to double check.

Many thanks…

Billy

1 Like

I can not help with your detail specific question. Nearcast rain does not work for my area so I never used it. This post is similar and links to some information regarding Nearcast rain that might help understand slightly

1 Like

Wow it is hard to navigate the newly reorganized pages…

GettingStarted => Quick Start => Observation and Event Definitions

This gets you to the Tempest Observations collected from midnight to midnight in the station’s local timezone. Denoted in API responses with type obs_st_ext

12 = rain during reporting interval
19 = nearcast rain during reporting interval
17 = the reporting interval

18 = rain since midnight
20 = nearcast rain since midnight
(and I'd expect daylight savings transition days to be a mess here)

The next block of examples are Tempest Daily Observations collected from midnight to midnight in the station’s local timezone. Denoted in API responses with type obs_st_ext

28 = rain
29 = nearcast rain

So I’d expect (hope) that:

28 = 18
28 = sum_of_the_12

29 = 20
29 = sum_of_the_19

I have no idea which way(s) into their REST vs. Websockets gets you to obs_st and obs_st_ext. given how cryptically their docs are written, but they seem to use the same arrays for each way.

Also note the UDP API array content for tempest observations is of course different in what’s where since it has no nearcast anything.

I must say I’m pretty lost as well :slight_smile:

I hadn’t seen:

I was working off:

https://weatherflow.github.io/Tempest/api/

1 Like

Which is more up to date?

The QuickStart gives a few options to register lightning messages:

{
  "type":"geo_strike_listen_start",
  "lat_min": 37.28,
  "lat_max": 41.32,
  "lon_min": -101.76,
  "lon_max": -91.00 ,
  "id": 12345
}

with strike_type = xxx

And the other docs say if I send:

{
  "type":"listen_start",
  "device_id":1110,
  "id":"2098388936"
}

I’ll get evt_strike messages.

Does one work - do both?