fix #10399: correctly calculate row and column for wells

[mtbc]

Fixes http://trac.openmicroscopy.org.uk/ome/ticket/10399

In testing, try different plates and see that the wells in the "map of the plate" correspond to the thumbnails visible in the data manager, and that the appropriate thumbnails have well name titles.

[gusferguson]

@mtbc

Tested using OMERO.insight-4.4.6-402-48dc9a9-dirty-ice33-b174-mac on howe as user-7.

Had to test using already imported data as plate import crashing - so could not test with plate pwalczysko reported issue with in 10399 - which has discontiguous wells (looked for it under user-6 but has disappeared).

Behaving as expected:

• seeing all wells shown on map as thumbnails in centre pane
• select on map map well gives correct highlight of thumbnail in centre pane and vice versa

Issues:

• thumbnails in centre pane columns 2 or greater and rows 2 or greater not displaying well name titles (Column-Row) above them - see screenshot below - ? if this is intended or not.
• double click on plate in data tree to open map for first time - click on map off of a filled cell to get focus - then click on filled well and map disappears. This does not occur on subsequent viewings of map - only first.
• select thumbnail highlights thumbnail and corresponding well in map highlights. Click off thumbnail so no thumbnail is selected, and the map does not update to reflect deselection of well in map.

Had crashes and filed 2 bug tickets:

java.lang.Exception: Abnormal termination due to an uncaught exception.
java.lang.IllegalArgumentException: Object cannot null.
https://trac.openmicroscopy.org.uk/ome/ticket/10490

and crash at import:
http://trac.openmicroscopy.org.uk/ome/ticket/10491

[mtbc]

First issue: yes, it's a feature, not a bug. I had to ask too. (-:

I'll investigate to see if this PR could have caused the others.

[jburel]

Ideally we should remove the number/letter from the well sample itself. I had another solution which was leading to a possible crash on windows
we will need to review the layout later on (another PR)

The problems noticed by Gus are not related to the PR, one is due to an object not being loaded and one introduced in #784

[mtbc]

@jburel: thank you

[mtbc]

Yes, second two bullets seem to pre-exist this PR. (Confirmed on Mac with near-dev_4_4 system without these commits.)

[jburel]

FYI: http://trac.openmicroscopy.org.uk/ome/ticket/10491 has been fixed in #784 see cfc763b

[jburel]

@gusferguson: could you retest tomorrow? @scottlittlewood's fix will be in.

[gusferguson]

@mtbc
Tested using: OMERO.insight-4.4.6-430-4977636-dirty-ice33-b175 on Mac 10.8 and Windows 7 - howe user-7

Behaving as expected:

• plate with discontiguous wells displaying correctly.

Noticed discrepancy with naming on this test:

• wells in the first row and first column - the name in the bar in centre pane is A1 > Maik, B2 > Maik etc. whereas in the right hand pane the name is much longer - experimenter-descriptor.xml ... etc.
• in other wells not in the first row or column the name in the bar is the same as that in the right hand pane.
  See screenshot below.

No new issues just issues 2 and 3 listed above that predate this PR.

Crash on import resolved.
Still getting crash reported in https://trac.openmicroscopy.org.uk/ome/ticket/10490

[mtbc]

That last issue is interesting. I think it may arise from how BioFormats' ScanrReader.initFile constructs well row and column according to the well labels noted by the handler, but those metadata text strings you indicate instead seem to have well and field based on the maximum field count. I don't know if my summary is either correct or useful, but @melissalinkert can probably provide rationale should you wish.

[melissalinkert]

What you see in square brackets in the right-hand panel is what Bio-Formats (correctly) returns as the Image name; I can pretty much guarantee that the discrepancy between the middle and right panes has nothing to do with the reader.

I'd be surprised if the conditional call to formatWellSampleTitle() was not the issue; calling that on all well samples should yield the short names in the middle pane (e.g. "B 2"), while not calling it would yield the longer names based upon the Image name.

[jburel]

@mtbc, @melissalinkert: it is a display issue in insight.

[mtbc]

Ah, now I read it again, I completely misread @gusferguson's point, ignore me.

[jburel]

@gusferguson: Could you create a ticket (name discrepancy) and attach the screenshot to it?

[mtbc]

@gusferguson: Also please include Melissa's last paragraph, it might save time locating the cause.

[gusferguson]

@jburel https://trac.openmicroscopy.org.uk/ome/ticket/10500 BUG: Screen well name discrepancy