WeeWX Beta 4.0.0b10 + vantage + LOOP1 = missing altimeter value

[Paul Anderson]

Hi All,

First off huge fan of WeeWx it is amazing software written by an unbelievable group of developers ! Thank you all for your tireless effects and dedication. 

Running Beta 4  version  4.0.0b10

Hardware Vantage Pro 2

Set to use LOOP1 so WeeWx needs to calculate altimeter as it's not in LOOP1.

Unfortunately it seems to silently fail to calculate altimeter, if used in a tag it returns as na.

And a sqlite query returns:

Note the missing altimeter values.

sqlite> select datetime, barometer,altimeter,pressure from archive order by datetime desc limit 5;

1580259900|29.792||29.6113663568201

I know this is going to sound crazy, the issue seems to have began at this commit:

I'm sure that Halloween Goblins were somehow involved :) 

First Know Bad commit

Document parameters of WXCalculate.__init__

https://github.com/weewx/weewx/commit/bb130bb55c3c941c9160388161198ada199b3032

@tkeffer

tkeffer committed on Oct 31, 2019

Last Know Good commit

Don't bail on a type until you have to.

https://github.com/weewx/weewx/commit/2286aece0abdaa2d1fc67e7da0627a620b9673ff

@tkeffer

tkeffer committed on Oct 30, 2019

Ran each commit with the same weewx.conf

Logged the run and put a sqlite query at the bottom.

Because I had such a difficult time believing it has gone unnoticed so long looked for some evidence beyond my own setup. TK your DW3693 station stopped sending altimeter data to CWOP on 2019-11-09T23:10:00Z

CWOP hasn't seen a Pressure reading from DW3693 since then.

I'm guessing that you updated to First Know Bad commit at that time or Switched from LOOP2 to LOOP1 at that time?

LAST Barometric Pressure to CWOP Nov 9 23:00 Z

Station     Date_Time      altimeter   air_temp

_ID

D3693,2019-11-09T22:20:00Z,30.16,55.99,71.0

D3693,2019-11-09T22:30:00Z,30.16,55.99,70.0

D3693,2019-11-09T22:40:00Z,30.17,55.99,70.0

D3693,2019-11-09T22:50:00Z,30.16,55.99,70.0

D3693,2019-11-09T23:00:00Z,30.16,55.99,70.0

D3693,2019-11-09T23:10:00Z,,55.99,70.0

D3693,2019-11-09T23:15:00Z,,55.99,71.0

D3693,2019-11-09T23:25:00Z,,55.0,71.0

D3693,2019-11-09T23:35:00Z,,55.0,71.0

D3693,2019-11-09T23:45:00Z,,55.0,73.0     

Thanks,

Paul

Attached

4.0.0b10.log.txt

Last_Know_Good_Commit_2286a.log.txt

First_Know_Bad_commit_bb130.log.txt

weewx.conf.txt

[gjr80]

I just updated to the latest 4.0 code and ran it under python3 with the simulator providing only barometer out of the three pressures (for the purposes of this exercise this is closely resembles the Davis loop packet). altimeter was not calculated for each loop packet; however, it was calculated for and appeared in each archive record. Coupled with the changes in the 'First_Know_Bad_commit' being of a cosmetic nature only I am not convinced yet there is an issue in the code base. Nothing seems out of place in your log or config, can you run WeeWX directly and provide screen captures of the console output making sure we see at least one archive record (lines starting with REC:) and the loop packets (lines starting with LOOP:) leading up to the archive record.

Gary

[gjr80]

OK, had another look and I see the problem; the changes in the 'Bad commit' inadvertently caused WeeWX to try to calculate altimeter before pressure and since altimeter is dependent on pressure that causes a problem if pressure does not exist (as is the case on a Davis system using other than LOOP2). I have applied a fix to the 4.0 code base, guess we will see if that survives Tom's critical eye, but either way 4.0 will have the problem fixed.

Well spotted and thanks.

Gary

[Thomas Keffer]

Good sleuthing!

The solution is a lot simpler: svc_dict is actually an ordered dictionary (specifically, a ConfigObj.section). So, all we have to do is make sure they are listed in the correct order in DEFAULTS_INI. Right now, they are in alphabetical order.

-tk

--
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 weewx-user+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/weewx-user/688dcca9-89e1-4c13-a3ea-51e0a728c079%40googlegroups.com.

[Paul R Anderson]

Verified Gary's commit fixed altimeter issue under Python 2.7.16, and  Python 3.7.3 values now being saved to database. Ever if the method changes i'am sure it will be fine.

Sidenote seemed like the perfect time to try the new  wee_database --calc-missing utility It of course worked perfectly and calculated all the missing altimeter fields in my database. Great job!!!!! 

Paul

To view this discussion on the web visit https://groups.google.com/d/msgid/weewx-user/CAPq0zEDteVux8OOA9BK9dVT2n%2BEcKqWccnyN1kM51WxdhufgtA%40mail.gmail.com.