Confusing locations - categories don't correspond to the location settings

[yshish]

I switched the location from Med to Cal but the image which loaded was from the Med, displaying a (dark) Radiolarian - the physical data below the image tells Mediterranean (despite I've switched for the California) but the menu offer me the California current categories. So I saw an image from wrong location but couldn't classify it by the correct categories.

Here is a screenshot.. The following image was also from the Med (with Cal categories). They both display among the Recent classifications.

[yshish]

I finished the classification, clicked to discuss the image (opened it in a new tab) and then clicked on the blue button to return to the classification page. The image which loaded was from the Med again but the menu with categories as well! I clicked on the main page to check which location has been set. The answer is Mediterranean. Odd!

So I switched to California again and the image which loaded was from Cal as well as the menu categories.

[yshish]

It could be the same issue which was reported by Kirsten here: http://talk.planktonportal.org/#/boards/BPK0000002/discussions/DPK0000a7n?page=1&comment_id=555a88e15135d04cdb0056eb

[srallen]

No, the issue Kirsten reported was resolved in #71 and was not related to sign outs like I surmised. It was totally just a display issue.

When this issue happens, do you by any chance notice whether or not it logged you in correctly when clicking to return to classification?

[srallen]

I think I fixed this just from doing some code refactoring in ad45557cba3d1226ecf8d7c33f9eabb1b6114538

I'm going to close for now, but do reopen if you encounter another instance of the tool controls being out of sync with the location label.

[yshish]

It has just happened again but there were some differences*. I kept the Classification page opened in one browser tab and opened the Home page in a new one. Then I switched the location from MED to CAL in the new tab. The images which loaded in the new tab were definitely from the MED however the list of categories in the pop up menu were from the CAL, and the inscription was telling me I'm classifying the CAL locality. When I clicked on the Discuss button to check what location it displays on the Talk page (I opened it in a new tab), it was correctly showing the MED location.

I wasn't logged out when it happened. It hasn't changed after classifying next images.

Screenshots:

• from the Classification
• from the Talk

Compare the location!

*(the list of categories was showing the same location as the the Classification page but the image was actually from the other location as proved on the Talk page)

What do you think?

[srallen]

@yshish, I think I finally tracked down the issue. This was a difficult one to replicate, but I believe the problem was the asynchronicity of user log in and subject selection and the fact that 10 subjects are queued typically. I needed to clear out the subject queue after user authentication or log out since the group may change at log in/log out depending on user preference or random selection.

[yshish]

I thought it shouldn't make problems any more..

I opened the home page in a new tab, switched the location from MED to CAL and all the images I got to classify were from the MED however it said they're from CAL (and the category list was offering the CAL animals only) but the Talk page confirmed they're from MED. After at least 10 wrongly signed frames I got a real CAL one to classify.

Shall I figure out the exact number of images which are shown wrongly?

[srallen]

Ok, it's loading up the 10 subject queue somewhere still where I need to clear it out. I'll check on it today, but I should be able to fix it quickly now that I know what the problem is.

[yshish]

It seems to be problem only when switching from MED to CAL! Is it possible?

[srallen]

As noted in #52, @yshish, could you double check this using the production site? Please let me know either way. I haven't been able to replicate this issue on the production site. (FYI I manually deploy, so the beta site doesn't always stay in sync with production, also shouldn't be used unless testing a new feature. It doesn't use production data!)

Anyway, if you aren't seeing it anymore in production, I'll re-close the ticket.

[yshish]

It works perfectly now, when I switched the beta version for the production one. You can re-close the ticket.

[srallen]

Thanks!!

[yshish]

@srallen User suzeroo reported getting images for the other location than she chose:

Much earlier today I've classified in both the Cal and Med data sets. A few minutes ago I was in the Med set, and saw a small solmaris - unusual for the Med set, so I clicked on discuss. The correct image showed up, but the label is the Cal set, and yshish had commented on that image 2 years ago. Then a few minutes later it happened again - I'm in the Med set and get another small solmaris, and when I clicked on discuss - the correct image showed up, but labeled as from the Cal set. The first image is APK00074sd . The second one is image APK0006x7q .

The other thing that happened today is that when I'm in the Med set, and click on discuss, make my comment, then click on return to classifying - it took me to the Cal set.

At least I've been able to classify and haven't yet gotten an out of data message!

Here is the report: http://talk.planktonportal.org/#/boards/BPK0000002/discussions/DPK0000erq?page=1&comment_id=56d2ae36cc453d4cb00001d1 Could you look at it please? Thanks!

[yshish]

The problems she's reported occur on five different computers she uses: two use Windows 7, one has Windows 10, all use Chrome.