briis / smartweather

WeatherFlow Smart Weather Component for Home Assistant
MIT License
108 stars 11 forks source link

Pressure readings bug #52

Closed Notrial closed 3 years ago

Notrial commented 3 years ago

Hey m8, It's Tempest guy again :)

I think I discovered a bug of some sort. I'm checking my pressure readings (both Sea level and station) and I notice it drops to 0 a few times a day, which naturally messes up with the graphs, ranges etc.

Please take a look at the screenshot below: image

I don't see those "0" values in Weatherflow app / cloud, so it must be some sensor or API issue. Any clues?

Cheers!

briis commented 3 years ago

Do you get any errors in the Log file?

Notrial commented 3 years ago

Nothing related to the Smartweather component as far as I can see...let me inspect it further.

briis commented 3 years ago

I rechecked the code, if I get bad data from the API, I return a 0 value on all numeric sensors. There can be numerous reasons why we get bad data, especially when the are getting data from a source on the Internet.

Would it be possible for you to ignore all events with a 0 value in your chart?

Notrial commented 3 years ago

What you see as a flat line is actually correct readings btw... (small pressure differences)...but those 0 values mess up the graph range,

Notrial commented 3 years ago

hmm that makes sense. I'm using https://github.com/kalkih/mini-graph-card ....I'll look into it. Is there any way you can omit those 0 readings on your end?

dz4news commented 3 years ago

I’m seeing the same issue with the pressure readings from my Tempest. I am using the historical graph template from within HA.

briis commented 3 years ago

II have released v2.2.2 that should fix this issue. It will simply not update the value, if it is 0, so it will just continue with the old value, until the next valid number is received. I have not had the chance to test this properly, as I have not gotten any fault data. So let me know if this solution fixes or do not fix the problem. I will close this issue now, but feel free to re-open.