I don't think you've posted your weewx.conf. It could prove useful. Make
sure you obfuscate any passwords first. Or, use the utility weectl debug.
Yes, if this is a bug in WeeWX, it should be possible to reproduce it using
the simulator. This is what I used (diff below). It emits a bad value 5% of
the time, zero the rest of the time.
===================================================================
diff --git a/src/weewx/drivers/simulator.py b/src/weewx/drivers/simulator.py
--- a/src/weewx/drivers/simulator.py (revision
1566a768af0963a43bd1c7b174293c46226a7fb5)
+++ b/src/weewx/drivers/simulator.py (date 1736899735080)
@@ -249,6 +249,11 @@
else:
self.packet_number = 0
amt = 0
+ import random
+ if random.randint(0, 100) > 95:
+ amt = 10
+ else:
+ amt = 0
return amt
For StdQC, I used
[[MinMax]]
rain = 0, 10, mm
rainRate = 0, 3, mm_per_hour
-tk
On Wed, Jan 15, 2025 at 2:14 PM Francis Perea <[email protected]>
wrote:
> Hi Tom,
>
> I really appreciate all your efforts.
>
> I don't really know how I could send you that simulator configuration.
> What would you need? My weewx.conf?
>
> I have discovered that it is not really a fault of my sensor but some
> interferences from a neighbour ham radio station. Talking with him he made
> me a comment and later we have confirmed that every time he transmit in a
> concrete frequency my station get wrong observations. At least I know now
> where the faults come from. And more than before I think that a correct
> software QC is the better way to keep wrong values to get into de DB
> because it seems more complicate to me to avoid the interferences.
>
> I would really need to know how to prevent those values from getting into
> the database and still think that a correct StdQC service configuration
> would do the trick but I don't know how.
>
> Would you please suggest any other configuration or tests I could develop
> to try to trim my config?
>
> Thanks once again for all your time and interest.
>
> Regards,
>
> Francis
>
> El miércoles, 15 de enero de 2025 a las 1:52:52 UTC+1, Tom Keffer escribió:
>
>> I am sorry, Francis, but I am unable to reproduce this.
>>
>> I modified a copy of the simulator to emit an out-of-range value for
>> "rain" randomly about 5% of the time. After letting it run for 40 minutes,
>> I looked at the database. There was nothing but zeros in both the archive
>> table and in the daily summaries:
>>
>> sqlite> select datetime(dateTime, 'unixepoch', 'localtime'), rain,
>> rainRate from archive;
>>
>> 2025-01-14 16:10:00|0.0|0.0
>>
>> 2025-01-14 16:15:00|0.0|0.0
>>
>> 2025-01-14 16:20:00|0.0|0.0
>>
>> 2025-01-14 16:25:00|0.0|0.0
>>
>> 2025-01-14 16:30:00|0.0|0.0
>>
>> 2025-01-14 16:35:00|0.0|0.0
>>
>> 2025-01-14 16:40:00|0.0|0.0
>>
>> 2025-01-14 16:45:00|0.0|0.0
>>
>> sqlite> select datetime(dateTime, 'unixepoch', 'localtime'), sum, max
>> from archive_day_rain;
>>
>> 2025-01-14 00:00:00|0.0|0.0
>>
>> sqlite> select datetime(dateTime, 'unixepoch', 'localtime'), sum, max
>> from archive_day_rainRate;
>>
>> 2025-01-14 00:00:00|0.0|0.0
>>
>> There is something peculiar about your setup, but I don't know what it
>> is. If you can send me a configuration that reproduces the behavior by
>> using the simulator, I'm happy to take another look.
>>
>> -tk
>>
>> On Mon, Jan 13, 2025 at 8:17 AM Francis Perea <[email protected]>
>> wrote:
>>
>>> Hi again Tom,
>>>
>>> Definitely not enough with my new StdQC configuration. I still get wrong
>>> observations into my DB.
>>>
>>> Just now I have this in my weewx.conf:
>>>
>>> [StdQC]
>>>
>>> [[MinMax]]
>>> rain = 0, 10, mm
>>> rainRate = 0, 3, mm_per_hour
>>>
>>> And I still get messages like these:
>>> ene 13 16:10:17 meteopi weewxd[193488]: WARNING weewx.qc: 2025-01-13
>>> 16:10:00 CET (1736781000) Archive value 'rainRate' 4.046533333333339
>>> outside limits (0.0, 0.11811023609999999)
>>>
>>> ene 13 17:00:17 meteopi weewxd[193488]: WARNING weewx.qc: 2025-01-13
>>> 17:00:00 CET (1736784000) Archive value 'rain' 0.539 outside limits (0.0,
>>> 0.3937007874015748)
>>> ene 13 17:00:17 meteopi weewxd[193488]: WARNING weewx.qc: Discarding
>>> observation
>>> ene 13 17:00:17 meteopi weewxd[193488]: WARNING weewx.qc: 2025-01-13
>>> 17:00:00 CET (1736784000) Archive value 'rainRate' 4.399999999999997
>>> outside limits (0.0, 0.11811023609999999)
>>> ene 13 17:00:17 meteopi weewxd[193488]: WARNING weewx.qc: Discarding
>>> observation
>>>
>>> Which generate these values into my DB:
>>>
>>> dateTime rain
>>> rainRate
>>> 1736781000 0.09899999999999998 3.079999999999999
>>> 1736784000 NULL 4.3119999999999985
>>>
>>> It seems that the rainRate filter is not correct or not filtering.
>>>
>>> Any clue?
>>>
>>> Thanks once again for your help.
>>>
>>>
>>> El domingo, 12 de enero de 2025 a las 18:51:15 UTC+1, Francis Perea
>>> escribió:
>>>
>>>> Hi Tom,
>>>>
>>>> I get what you mean. I have been making all kind of tests to avoid
>>>> wrong observations to get into the data base and I even think I began
>>>> setting min-max values for rainRate, but I'll try again.
>>>>
>>>> What I get from your excellent explanation is that I should try to
>>>> filter rainRate, not rain, and see if that keeps "pollution" out of the DB,
>>>> isn't it?
>>>>
>>>> I think I'll use this:
>>>>
>>>> rainRate = 0, 3
>>>>
>>>> Do you think it would be enough? Or would you suggest me any other way
>>>> of filtering? Do you think I should also try to modify the rain_period
>>>> parameter from its default 1800 seconds?
>>>>
>>>> Regarding the [[RainRate]] section you see in my config, I've to be
>>>> honest and say that because of my desperation I even asked ChatGPT how to
>>>> proceed and this is what he propossed. I have already removed it.
>>>>
>>>> Thanks again for your support and patience.
>>>>
>>>>
>>>> El sábado, 11 de enero de 2025 a las 23:56:00 UTC+1, Tom Keffer
>>>> escribió:
>>>>
>>>>> It looks like the field "rain" was nulled out (otherwise, it would be
>>>>> 0.0, not NULL), so no bug.
>>>>>
>>>>> The field rainRate is not null, *however* the fields "rain" and
>>>>> "rainRate" are completely different. The former represents the amount of
>>>>> rain that fell, the latter how fast it fell. The field rainRate may be
>>>>> emitted from your device, in which case, by default, WeeWX uses that
>>>>> value.
>>>>> I suspect your byows hardware does not do this, so WeeWX will calculate it
>>>>> using a *running average* over 15 minutes. So, if you get a bad rain
>>>>> value, it will continue to "pollute" the rainRate value for up to 15
>>>>> minutes afterwards. You need to filter rainRate as well, or set its
>>>>> running
>>>>> average window, given by option rain_period
>>>>> <https://www.weewx.com/docs/5.1/reference/weewx-options/stdwxcalculate/#rainrater>,
>>>>> to something shorter.
>>>>>
>>>>> [StdQC]
>>>>>>
>>>>>
>>>>>>
>>>>>> [[RainRate]]
>>>>>> min = 0
>>>>>> max = 3
>>>>>> min_delta = 0
>>>>>> max_delta = 3
>>>>>> reject = True
>>>>>>
>>>>>
>>>>> I don't know what any of this is. Did you write an addition to StdQC?
>>>>> If so, you're going to have to debug it!
>>>>>
>>>>> -tk
>>>>>
>>>> --
>>>
>> You received this message because you are subscribed to the Google Groups
>>> "weewx-development" 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-development/91ec157e-5f06-406b-8c16-f55fb40cd694n%40googlegroups.com
>>> <https://groups.google.com/d/msgid/weewx-development/91ec157e-5f06-406b-8c16-f55fb40cd694n%40googlegroups.com?utm_medium=email&utm_source=footer>
>>> .
>>>
>> --
> You received this message because you are subscribed to the Google Groups
> "weewx-development" 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-development/19706d14-3e4d-466a-9455-f438d2d4b7d2n%40googlegroups.com
> <https://groups.google.com/d/msgid/weewx-development/19706d14-3e4d-466a-9455-f438d2d4b7d2n%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
>
--
You received this message because you are subscribed to the Google Groups
"weewx-development" 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-development/CAPq0zEAhHAFZXQANwh%2BXo7Hv%3DwF1qYEN8JUaRphCWvT9DnnVFg%40mail.gmail.com.
