Some JSON_CT stuck on Connect

I’ll check them tomorrow

Hello @oadolph,

I was going to Windhoek, but unfortunately, I had to cancel due to other priorities. It would be nice to discuss the same matter with you there. I want to follow up on this discussion with @josip.kusak to find ways to prevent this error from happening with us here.

Thanks,
Albert

HI Albert,
I cannot help much, except to nudge programmers from time to time. :slightly_smiling_face:
Cheers,
Josip

1 Like

Well, me again going back to this topic.
After testing the latest SM (unreleased ver. 481), the number of unprocesable JSON packages grew dramatically. I recording some patrols and few observations while patrolling, and few independent observations. I made some photos associated with those observations. After import, I see those observations in SMART Desktop, including associated media files. I am not sure what unprocesable packages are. It seems (due to small file size) they contain some unimportant waypoints from patrol, but it would not be good if they would contain data about observations.
ZIP with those unprocesable JSONs is here.
Greetings,
Josip
JSONs_stuck_on_server_JK_20240303.zip (381.2 KB)

Hi Josip,
I quickly checked these unprocessable files. Beside one they are all waypoints where the location was not recorded properly:
“geometry”: {
“coordinates”: [
1e-10,
** 1e-10**
],
“type”: “Point”
},

Can you check, if the app has the permissions to access the location service of your phone and that the location service is working

Hello Oliver!
Yes, my device has the access to location service. Many other locations and track-log points were properly calculated. I kept my phone in my pocket while making those test patrols (testing ver. 481 of SM), and it may happen that some of track-log points were not calculated properly. I will just delete those invalid packages.
It would be good if there is some background control mechanism, which would filter-out erroneous locations, so that they do not bother us.
Thanks,
Josip

Hi Josip,

We often have packages that can’t be processed, sometimes a whole bunch of them but more recently I’ve managed to figure out why this has happened and been able to successfully unblock them, or marked them as complete knowing they contained no unprocessed data.

For most of these solutions I’ve been using the preview panel on the data queue page of the Connect webpage to match dates, times, deviceID, start and stop times, and observation counter, and comparing with patrols already in the desktop SMART.

The various reasons for blockage include:

  1. Packages with track points or observations not processing before the start patrol has arrived and been processed (we have phones in areas with poor connection so sometimes packages are missing or arrive out of sequence).

  2. Packages with duplicated data. Sometimes this can be substantive data such as a long series of track points, or an observation package that arrived twice. For stuck observations find the patrol on SMART Desktop with same start date and time and check the observation numbers. More commonly it is one or two track points that seem to get recorded at the time when the patrol is stopped. These latter type can be safely deleted. I think this latter duplication is a bit of a bug that could be ironed out. Long series of trackpoints can be harder to resolve but if you locate a likely patrol in desktop (same date/time period) you can check the details of track points by choosing “edit track” option on the patrol.

  3. Missing observation package so the observation counter is not sequential. My experience with solving this one hasn’t been consistent. Sometimes desktop seemed to accept a gap in the counter and other times I renumbered the counter in the stuck packages (download the package, edit the text file then import the package from local drive). In this latter case I knew I was accepting to have lost the missing observation.

  4. Missing start message. If the STOP message is available I downloaded it, changed the observation type from StopPatrol to StartPatrol, processed it so that it created the patrol and that allowed all related data to process. Then did what I could to revise the coordinates to something sensible. I was surprised to discover that the original Stop message also then processed successfully. I was quite sure it would fail due duplication of a key. Maybe that will come back to haunt me.

  5. Patrols spanning more than one day without the use of Pause. This is quite common for our project. Manually adding additional dates to the patrol in SMART desktop can unblock these packages.

Solving many of these issues would be made very much simpler if the track point packages included a patrolID key, and there was some way to cross reference the keys used in the packages (e.g. deviceID) to the data in Desktop. A list of deviceIDs for each user would be helpful. Or deviceIDs and Patrol number.

But generally in Smart 7.5 we’re having far fewer of these issues than previously.

All my best

Jeremy

1 Like

Thanks on the thorough explanation, Jeremy!
This is a long list of things which needs to be resolved, somehow.
I believe most of users would rather delete unprocesable packages, instead of trying to find the problem and resolve it.
Sincerely, Josip

1 Like

Here I am, back again with the apparently the same problem, and this time it is serious! I cannot get any data right from patrol which includes observations containing attached photos. I used the last official release of SM (ver. 478), spent the whole day driving around, and collected a number of observations and samples for which data is all in SMART. I know I did everything right, like “million” times before. Started patrol, collected observations with included photos as a rule, collected several samples, came back to field station, finished patrol, uploaded all data of today to Connect, got OK message. However, when processing new packages from Connect to Desktop, not a single package went OK, but I got again a list of processing errors. I cannot just delete those, since there is real data and info about collected samples. If I discard unprocesable packages, the whole day of work was for nothing.
I need those packages processed. What can be done? Please, anyone!
here they are, 85 MB: Dropbox - Error - Simplify your life

Greetings,
Josip