It depends on lovelace card you use.
By default Home Assistant uses “km” for length in the metric system and the speed is just length/time.
I modified the animated weather card to use m/s (GitHub - tdragon/weather-card: Weather Card with animated icons for Home Assistant Lovelace), but I didn’t have time to create a proper pull request for this change.
You can try it, it works in my setup:
2 Likes
it works, thanks!
1 Like
Hi,
Thanks for the great integration. I have been using this now couple of months.
However, I have noticed one strange issue. The only version that is working for me is 0.4.0.
If I try to update any of the newer version (0.4.1 - 0.4.5), I will get this error:
This error originated from a custom integration.
Logger: custom_components.fmi
Source: helpers/update_coordinator.py:193
Integration: Finnish Meteorological Institute
First occurred: 17.06.36 (5 occurrences)
Last logged: 17.10.56Timeout fetching fmi data
Any ideas what might cause this?
When I downgrade back to 0.4.0, component start to get all the data again.
Most likely it is caused by lightning update. On my setup it takes up to 10 minutes sometimes:(
For now I just commented it out in the code.
Here as well: after the thunderstorm last night FMI integration crashed and does not start any more. What part of the code should I comment out? I am not so interested in ligthnings.
In init.py, after “# Update lightning strikes”
comment lines:
await self._hass.async_add_executor_job(
update_lightning_strikes
)
_LOGGER.debug("FMI: Lightning Conditions updated!")
This will disable lightning functionality.
I had an idea to make it optional, but don’t have time right now to implement it.
I also noticed that for my location the forecast is totally wrong. For example, the official FMI app shows +30 today at 14:00, and the current temperature is +26.
But my Home Assistant shows 25.4 at the same time and current is 17.9 
@anand-p-r Do you have any idea what could be wrong with the forecast? I tried to query fmi_client manually and it gives me same wrong values
Thanks! Now it works.
Do you get a correct data from the FMI?
On official FMI site I see following:
but from integration it looks like:
And if I query FMI API, I get same readings :(:
2021-06-23 11:00:00+03:00 23.38
2021-06-23 12:00:00+03:00 23.5
2021-06-23 13:00:00+03:00 24.61
2021-06-23 14:00:00+03:00 25.22
2021-06-23 15:00:00+03:00 24.06
2021-06-23 16:00:00+03:00 23.41
2021-06-23 17:00:00+03:00 22.9
2021-06-23 18:00:00+03:00 23.21
Got an answer from FMI:
In Open data we hava two weather models. In our application and in our webpages we have meteorologist forecast. They are different thing. Meteorologist use many forecast models and also her/his experience. In this situation meteorologist use much her/his experience especially near coast.
Hirlam is not good in temperatures and Harmonie is better, so it is better to take Harmonie from Open data than Hirlam.
Quickly checked, the integration uses Hirlam and there is no easy way to change it. Moreover, the Harmony doesn’t provide weather symbol…
@anand-p-r Could we update the client to get data from both models and merge them?
@RedViper Thanks for clarifying this with FMI! As you also observed, there is no easy way to merge both models. Need to study this and see.
@konna Can you try and see if version 0.4.6 works for you? Lightning data pulls were timing out and I have configured this optional now (default disabled) in addition to some other improvements.
Hi,
I uninstalled version 0.4.0 and installed 0.4.6 via HACS. This version is now working.
Thanks a lot!
1 Like
This is really quite bad model with temperatures.
Here’s 48h with met.no data, actual Ruuvi temperature, and this addon.
1 Like
FMI states now that they discontinue support for hirlam in September. And I also got frustrated with the hirlam forecast accuracy so I tried to create harmony fetch. It was pretty easy actually even for a novice like me. They have python library fmiopendata with which you can download that forecast. My version just got unnecessarily complex as I wanted to have both models available at the same time. And as you stated above there’s no condition data. So, the condition data just needs to be generated based on precipitation (rainy conditions), visibility (which can be added for foggy confitions), and cloudiness. Total radiation and dew point can be added too (as those are returned by the api). Could develop that further by letting used define what’s the boundaries for different icons, what are the icons used and in which conditions …
And as stated in github would be nice to load sensor data from closest observation and not use forecast data. Or show both if one wants to check how close their forecast matches with observation (if you happen to have observation station close to you and in same kind of environment).
Hi !
Anyvody else having same problem? FMI stopped working after 2022.11.0? 2022.11.1 did not help either.
Logger: custom_components.fmi
Source: helpers/update_coordinator.py:182
Integration: Finnish Meteorological Institute
First occurred: 15:34:39 (9 occurrences)
Last logged: 15:41:20
Error fetching fmi data:
1 Like
Got new version from GitHub - anand-p-r/fmi-hass-custom: Home Assistant Custom Component Supporting FMI Based Weather and Sensor Platforms
Found also that they’ve changed the Weather component so the newest FMI only works with new version of Weather component, but my HA is 2022.3… Oh, I love the update hell.
Is working now with newest version of FMI and old version of fmi/weather.py.
You got it. Salute the updates. Gotta try to play with versions then. (Or live without until fixed 
)
Update: Don’t know why I was waiting this to appear in HACS. It’s manually installed 
My bad, now everything is working again after this latest version. Now HA 2022.11.2.
The different sensors created by this integration don’t have a state_class. Therefore no statistics are recorded. Does anyone else have this problem?
Is this integration still up to date? I get wrong temperature at my location when comparing to FMIs web site with the same location.