> another question: what do you mean by " top of the interval" ?
In my understanding the top of the interval is the end.

I didn't address this behavior as an issue, I just asked (myself) the 
question, if this behavior is desired and how does this behavior fit into 
WeeWX? You're right that a 5-min average with an the interval of 07:22 to 
07:27, imported for the WeeWX interval from 07:25 to 07:30 wouldn't be 
correct, but if your hardware only ever sent data in a 5-minute-interval, 
WeeWX would have stamped the same data from 07:27 with 07:30 when actively 
running and receiving these values as a loop package (setting any 
highs/lows to 07:27), but when incoming from backfill, an archive entry 
stamped with 07:27 is created. I consider this inconsistent behavior, 
without proposing either way "correct" or not.

> For me there is no desired behaviour here - what was archived is stored. 
Accumulation only where accumulation is possible.
Im a bit confused: you state "there is no desired behavior" and in the same 
line you describe it: "what was archived is stored. Accumulation only where 
accumulation is possible." Accumulation would be possible, if there is more 
than one value in the device storage for an WeeWX archive_interval.
I've set the GW3000's interval to 1 min, to see what will happen: when 
backfilling, WeeWX now produces an archive entry every minute, that's in my 
understanding in contradiction to accumulate, when possible.

Anyway, I'm with you that this not a real issue and I also don't care if 
the backfilled data is aligned to the usual interval, but from a technical 
point of view worth discussing about.

Rainer Lang schrieb am Sonntag, 13. Juli 2025 um 11:34:03 UTC+2:

>
> I don't think that this "deviated" timestamp is a real issue.
>
> Assuming your weewx srchiving interval is 5 minutes (300 seconds) and also
> assuming the GW3000 archive interval is also 5 minutes, its archiving time 
> stamp will be whenever it starts archiving - any time between minute 00 and 
> minute 05 and multiples to fill the hour. If the timestamp is not 00, 05, 
> 15, ...,55, what is weewx supposed to do ? If the timestamp is 03 instead 
> of 05, there is nothing to fill the loop with. So it will store with 
> whatever the GW3000 timestamp was.
>
> And where is the problem ? For the backfill period you will get the data 
> that was stored in the GW3000 and then weewx will continue with 00, 05, 10 
> etc. whatever is the time once the backfill is completed. 
> For this situation to be aligned and corrected with a full 5 minute 
> archiving at the weewx end, you would have to recreate the timestamp - and 
> create wrong data by the way as they didn't occur at that new "aligned" 
> time.
> For me there is no desired behaviour here - what was archived is stored. 
> Accumulation only where accumulation is possible.
> In my situation the GW3000 interval is two minutes, the weewx interval 
> too. I don't care if the timestamp for the backfill period is 1, 3, 5, 7, 9 
> or 0, 2, 4, 6, 8.
> I don't think that making weewx "correct" or better say "align" the 
> timestamp (because this alignment will provide an incorrect time) is worth 
> the effort.
>
> Either, leave it as it is, or make sure your GW3000 archive at a full 5 
> minute interval timestamp or whatever is your weewx interval.
> Anyhow, I don't see any issue with some records having a timestamp not 
> matching the start minute of the weewx archiving during normal operation.
>
> It may be different if e.g. the GW3000 interval is smaller/shorter than 5 
> minutes, at least so short, that more than one record will fit into the gap 
> (assuming the weewx interval is still 5 minutes) - then the loop could be 
> used for accumulation, but I don't say it needs to be done that way. 
>
> It had such a situation before where my weewx instance crashed over a 
> weekend and 72 hours of data were missing. But I had the data archived in 
> my Meteobridge in a one minute interval and had one minute interval 
> exports. Before editing the export file by removing 4 records for every 
> five minutes, I simply imported them all. Less work, no harm done. 
>
> another question: what do you mean by " top of the interval" ? The 
> beginning or the end ? Weewx uses the timestamp of the end of loop 
> interval. Only this makes sense for a correct accumulation.
> The data from the GW3000 are the data at the time of the GW3000 archiving 
> - no accumulation there.
> On 13.07.2025 10:32, '[email protected]' via weewx-user wrote:
>
> I've played around a little bit with the installation of your fork and 
> Werners ecowitt_http.py and got many things working. 
>
> Calculation radiation from luminosity isn't required anymore, since the 
> device is calculation it and it is mapped, but luminosity now has to be 
> calculated :)
> When backfilling data, the timestamp of the record of the device seems to 
> be used, so with the default archive interval, when backfilling, you get 
> values for 07:27(or whatever it would be in the particular case) instead of 
> 07:30. Since WeeWX is accumulating values in an interval and storing these 
> accumulations with the timestamp from the top of the interval, I don't 
> think this is the desired behavior.
>
> So, for now, I consider the following as relevant settings in my 
> weewx.conf:
>
> [Station]
>     # Set to type of station hardware. There must be a corresponding 
> stanza
>     # in this file, which includes a value for the 'driver' option.
>     station_type = EcowittHttp
>
> ##############################################################################
> [StdCalibrate]
>    
>     [[Corrections]]
>         # For each type, an arbitrary calibration expression can be given.
>         # It should be in the units defined in the StdConvert section.
>         # Example:
>         #foo = foo + 0.2
>         luminosity = radiation*126.7 if radiation is not None else None
>         lightning_distance = lightning_distance if lightning_strike_count 
> > 0 else None
>
> ##############################################################################
> #   This section is for quality control checks. If units are not specified,
> #   values must be in the units defined in the StdConvert section.
> [StdQC]
>    
>     [[MinMax]]
>         co2 = 0, 65534, ppm # max value is one below the theoretical max, 
> because sometimes after a battery change the sensor delivers the max value.
>         pm1_0 = 0, 6553, microgram_per_meter_cubed
>         pm2_5 = 0, 6553, microgram_per_meter_cubed
>         pm4_0 = 0, 6553, microgram_per_meter_cubed
>         pm10_0 = 0, 6553, microgram_per_meter_cubed
>
> ##############################################################################
> [EcowittHttp]
>     # This section is for the Ecowitt Local HTTP API driver.
>    
>     # The device IP address:
>     ip_address = 10.0.1.84
>    
>     # How often to poll the API, default is every 20 seconds:
>     poll_interval = 10
>    
>     # The driver to use:
>     driver = user.ecowitt_http
>    
>     # Is a WN32P used for indoor temperature, humidity and pressure
>     wn32_indoor = True
>    
>     # Is a WN32 used for outdoor temperature and humidity
>     wn32_outdoor = True
>     # how many attempts to contact the device before giving up
>     max_tries = 3
>     # wait time in seconds between retries to contact the device
>     retry_wait = 5
>     # max wait for device to respond to a HTTP request
>     url_timeout = 10
>     [[catchup]]
>         source = device # source = either, both, net, device  ## not set 
> = None, default is then either or both
>
>
> Ian Millard schrieb am Samstag, 12. Juli 2025 um 22:53:50 UTC+2:
>
>> @Michael, 
>>
>> I am doing exactly the same thing. Can we share the pertinent parts of 
>> weewx.conf please? Obviously excluding sensitive information. If we can 
>> align our thoughts on this, I will update the repro to reflect. You can 
>> email me directly [email protected]
>>
>> Thanks,
>> Ian
>>
>> On 12 Jul 2025, at 19:05, '[email protected]' via weewx-user <
>> [email protected]> wrote:
>>
>> @Ian see: https://kainzbauer.net/weather/Test/Rif/ws68/index.html this 
>> is running the ecowitt http driver from your repo with the ecowitt_http.py 
>> from Werners repo.
>>
>> Ian Millard schrieb am Samstag, 12. Juli 2025 um 13:26:59 UTC+2:
>>
>>> @Michael, 
>>>
>>> I cannot test air quality data as I do have a device. However I am able 
>>> to calculate: -
>>>
>>> [StdCalibrate]
>>>     
>>>     [[Corrections]]
>>>         radiation = luminosity/126.7 if luminosity is not None else None
>>>         luminosity = 0 if luminosity == None else luminosity
>>>
>>> On 12 Jul 2025, at 10:25, '[email protected]' via weewx-user <
>>> [email protected]> wrote:
>>>
>>> Ian's fork doesn't support luminosity/radiation and pm/co2 values, 
>>> probably other things. I've copied the contents of Werners version of 
>>> https://github.com/WernerKr/Ecowitt-or-DAVIS-stations-and-Season-skin/blob/main/ecowitt_http/ecowitt_http.py
>>>  over the one I've installed from Ian's fork, getting 
>>> luminosity/radiation and pm/co2 working. Werner has posted about his work 
>>> here: https://groups.google.com/g/weewx-user/c/-rbfcQsNo_s but I'm not 
>>> sure how this really fits into Gary's work and their forks.
>>>
>>> Auchtermuchty Weather schrieb am Freitag, 4. Juli 2025 um 17:45:37 UTC+2:
>>>
>>>> Thanks for this Ian. When I have some spare time I will give it a 
>>>> whirl. 
>>>>
>>>> I have found the following in this thread:
>>>> https://github.com/hoetzgit/weewx-gw1000
>>>> *https://github.com/Millardiang/weewx-ecowitt_local_http/tree/development 
>>>> <https://github.com/Millardiang/weewx-ecowitt_local_http/tree/development> 
>>>> *- 
>>>> this is Ian's fork with working catch-up
>>>>
>>>> https://github.com/WernerKr/Ecowitt-or-DAVIS-stations-and-Season-skin/tree/main/ecowitt_http
>>>>
>>>> If I search github for weewx-ecowitt there are more but sorted by 
>>>> recently updated
>>>>
>>>> https://github.com/bidord/weewx-gw1000
>>>>
>>>> However, as above, I will try Ian's out. My last contact with Gary was 
>>>> on the 23rd April. :(
>>>>
>>>> On Thursday, 3 July 2025 at 12:20:39 UTC+1 steepleian wrote:
>>>>
>>>>> Yes it is very worrying. 
>>>>>
>>>>> This is the link to my fork: -
>>>>>
>>>>> [image: weewx-ecowitt_local_http.png] 
>>>>>
>>>>> Millardiang/weewx-ecowitt_local_http at development 
>>>>> <https://github.com/Millardiang/weewx-ecowitt_local_http/tree/development>
>>>>> github.com 
>>>>> <https://github.com/Millardiang/weewx-ecowitt_local_http/tree/development>
>>>>>
>>>>> <https://github.com/Millardiang/weewx-ecowitt_local_http/tree/development>
>>>>>
>>>>> https://claydonsweather.org.uk
>>>>>
>>>>> On 3 Jul 2025, at 10:48, Auchtermuchty Weather <[email protected]> 
>>>>> wrote:
>>>>>
>>>>> 
>>>>>
>>>>>
>>>>>
>>>>> On Sunday, 29 June 2025 at 11:44:21 UTC+1 Tom -KQ5S wrote:
>>>>>
>>>>> This is interesting.
>>>>>
>>>>> <sip>
>>>>>
>>>>> I call it deeply worrying. I was in touch with him around Easter when 
>>>>> he said he was on the verge of having the catch-up ready for me to test. 
>>>>> Then I was busy in various ways and his repository had vanished when I 
>>>>> came 
>>>>> back to looking for it. I have done some searching (probably not very 
>>>>> effectively) and can't find a trace of him. 
>>>>>
>>>>> Should point out it wasn't the only useful utility he had there - 
>>>>> there was PolarWindPlot, and I think a couple more.
>>>>>
>>>>> Worrying about Gary aside, where is the best place to download a copy 
>>>>> of the driver with backfil?
>>>>>
>>>>>
>>>>> -- 
>>>>> You received this message because you are subscribed to the Google 
>>>>> Groups "weewx-user" group.
>>>>> To unsubscribe from this group and stop receiving emails from it, send 
>>>>> an email to [email protected].
>>>>>
>>>>> To view this discussion visit 
>>>>> https://groups.google.com/d/msgid/weewx-user/332e8804-ad61-41c5-8dd4-3d5bbb3dbd3cn%40googlegroups.com
>>>>>  
>>>>> <https://groups.google.com/d/msgid/weewx-user/332e8804-ad61-41c5-8dd4-3d5bbb3dbd3cn%40googlegroups.com?utm_medium=email&utm_source=footer>
>>>>> .
>>>>>
>>>>>
>>> -- 
>>> You received this message because you are subscribed to the Google 
>>> Groups "weewx-user" group.
>>> To unsubscribe from this group and stop receiving emails from it, send 
>>> an email to [email protected].
>>>
>>> To view this discussion visit 
>>> https://groups.google.com/d/msgid/weewx-user/fa1c26a8-b1cd-4244-b500-5f3e3356da22n%40googlegroups.com
>>>  
>>> <https://groups.google.com/d/msgid/weewx-user/fa1c26a8-b1cd-4244-b500-5f3e3356da22n%40googlegroups.com?utm_medium=email&utm_source=footer>
>>> .
>>>
>>>
>>>
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "weewx-user" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to [email protected].
>>
>> To view this discussion visit 
>> https://groups.google.com/d/msgid/weewx-user/7e5b4660-87db-439e-9f49-f5b95043d881n%40googlegroups.com
>>  
>> <https://groups.google.com/d/msgid/weewx-user/7e5b4660-87db-439e-9f49-f5b95043d881n%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>>
>>
>> -- 
> You received this message because you are subscribed to the Google Groups 
> "weewx-user" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to [email protected].
>
> To view this discussion visit 
> https://groups.google.com/d/msgid/weewx-user/1c3e9553-2b30-450a-8bfa-b949c90a58ecn%40googlegroups.com
>  
> <https://groups.google.com/d/msgid/weewx-user/1c3e9553-2b30-450a-8bfa-b949c90a58ecn%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/weewx-user/cc618911-07a0-4380-b6f9-1e8c0e66ab81n%40googlegroups.com.

Reply via email to