Show chgrp target tree if Projects OR Datasets exist. See #10917

[will-moore]

This is in dev_4_4 as #1241, now for develop...

[mtbc]

@gusferguson, any comments on this?

[gusferguson]

@will-moore

Testing with Web client on gretzky 5.0.0-alpha4-473-c8bc509-dirty-ice33-b309 user-2 private-1 and read-only-1 Mac Safari

Moved images from private-1 to read-only-1 group which had only datasets in it.
Datasets appeared as expected in dialog.
Dialogue appeared correctly - able to select and move images to the 2 datasets there successfully.
Behaves as expected.

I think this is good to have.

Still getting issue with Stack Error when not selecting a dataset as reported in:
https://trac.openmicroscopy.org.uk/ome/ticket/11118

[joshmoore]

@gusferguson : the exception should be fixed with gh-1288. Can you confirm?

[gusferguson]

@joshmoore

The exception has been fixed, but now the images simply disappear from the one group and never appear in the other - orphaned images remains empty.

[pwalczysko]

There is some strange memorizing of the previous action going on here:

• login as user-2, import Beta catenin.lif from test_images_good/leica-lif twice, once into group private-1, once into group read-only-1
• go to Web client and being in read-only-1 try to move the 5 images of just imported MIF in following manner
  • select one and go Move to private-1
  • warning popup appears as expected, click Move All
  • Move window appears, select leica-liff dataset as target for the images
  • move proceeds as expected, now you have two 5-image MIFs in the leica-lif dataset in private-1 (the one you have imported into private-1 using Insight, the other which you moved from read-only-1 to it)
  • nov select one of the just moved images of the MIF and try to move it back to read-only-1, this time do not target the move - expecting the images to end up in the Orphaned
  • move proceeds successfully, but the images are not in Orphaned - instead, they are in the original dataset leica-lif (the empty dataset from which they were moved in the first place)
  • in order to see what will happen when I would delete the dataset which is somehow in a memory of the Move machinery, I moved the MIF back to private-1 and
  • tried to delete the empty leica-lif dataset in read-only - go an error - see screenshot below
    

[joshmoore]

The bug from 11109 is certainly that the links (DatasetImageLink, ScreenPlateLink, etc) were being hidden away, after the Image had moved groups. Does that sound like the "memorizing" that you're speaking of? If so, the next build should correct it.

[pwalczysko]

Possibly.

[pwalczysko]

No - the new build did not fully solve the problem described above, just partially. The first back-move, meaning the non-targeted move from the private-1 back to read-only group of the just moved MIF, although non-targeted, ends up in the original folder, where these data first came from. When I move the MIF again to private-1, then they really end up in orphaned images, as expected (this has improved). Looks like a first-time-only bug. The error is gone though, I can delete the empty dataset now.

[pwalczysko]

Sorry, when I went more carefully through the workflow above, I realized that I still get the warning and error as described abvoe when trying to delete the empty dataset. This is not fixed at all.

[pwalczysko]

Aditionally, when I try to delete the dataset with the 10 images from the 2 MIFs in private-1, I cannot delete it either - got the same error.

[pwalczysko]

I cannot delete the empty dataset (leica-lif, user-2, read-only-1) even from Insight. Nor can I delete the other one, the one with 10 images inside from Insight. (user-2, leica-lif, private-1).

[pwalczysko]

In Insight, when you do the first step of the above workflow, then actually just one image of the 5 is moved, so the dataset IS SPLIT ! (after using Move All)

[joshmoore]

Was this not a purge build? i.e. data was left over. If that's the case, then part of the issues will be caused by the fact that there were remaining hidden links. I didn't think to check the PURGE_DATA setting for the build. There may be other issues; certainly the SPLIT'ing is vital to track down.

[will-moore]

@joshmoore @pwalczysko Sorry to interrupt the conversation, but I don't think any of these issues are related to this PR. I think is good to merge!

[joshmoore]

@will-moore: I agree that it's not caused by this PR but it's hard to test this PR when the move isn't working correctly. I'll chat with @pwalczysko this morning and try to get things settled.

[will-moore]

@joshmoore This PR was already tested with the target Datasets showing in the chgrp dialog (even before you try the chgrp).

[pwalczysko]

@will-moore : probably this is not the right PR to have this conversation - I was redirected here when replying / re-testing Gus's workflow. If this PR is all about whether the dataset is displayed in the Move window to be targeted to than yes, this works. All my problems are coming from the actual moving itself, I would think. On the other hand, as discussed with @joshmoore this morning, this was not caused by some old data - the bug is firmly there. I think that Josh is working on it right now.

[jburel]

The move issue should now been fixed. Final check tomorrow.

[pwalczysko]

This all works now, see #1288. Ready to merge.

[joshmoore]

Sorry for having blocked this, @will-moore. @pwalczysko, thanks!