BUG: Rendering wrong in Insight and Web

[pwalczysko]

Login to Gretzky, user-4, group read-only-1.
Dataset Petr Big....

Concerns any big image in the folder (.vms, trestle, svs, leica-scn...).

Workflow

• Open images in full viewer.
• Zoom in and out
• pan

Results Insight:

• birds eye view inset okay but
• main viewer is wrongly rendered colorise
  • looks blue entirely (all the .vms, .ndpi images)
  • some tiles are blue, some are okay (leica-scn)
  • minority of tiles are blue, the rest is okay (.svs)
• when you try to zoom in and out then:
  • in some cases you get the bug away (svs)
  • in some cases this does not help at all ((vms, ndpi)

Results in Web:

• birds eye insert looks wrongly rendered colorise (blue-ish)
• the main viewer is okay
• if you zoom in and out, no help at all - the same situation at every zoom - main viewer okay, the insert bluish

Please note that the bluish color cannot be get rid of by changing the rendering settings - the colors are wrong from the onset, from the moment when you open the images - no way to readjust the colors back to normal.
All other features and functions of big images viewing seem to work fine.

The plates and small images view fine.

[jburel]

Probably related to one of the new PR. The only I can think of is the "caching PR"

[pwalczysko]

@jburel: @jamoore has just turned the caching off. The images look exactly the same as before.

[jburel]

The difference in bird eye view is due to the fact that insight uses the thumbnail, the web the lowest resolution.

[mlinkert]

I can reproduce this locally with just ​https://github.com/openmicroscopy/openmicroscopy/pull/1058 and ​https://github.com/openmicroscopy/bioformats/pull/465. Investigating now.

[mlinkert]

I think this is fixed with ​https://github.com/joshmoore/openmicroscopy/pull/19 which will update ​https://github.com/openmicroscopy/openmicroscopy/pull/1058 if/when merged.

Leaving this open until Josh can confirm that those are sane changes, and Petr can confirm that it really does work.

[jamoore]

Issue was caused by a refactoring of the old importer code (cF. OMEROWrapper, etc.) Melissa rolled this back in ​https://github.com/joshmoore/openmicroscopy/pull/19 , but this will ned to be reviewed again during cleanup.

[Melissa Linkert <melissa@…>]

(In [fa95c3949877d30aed7560cbd78eb2da96b9f027/ome.git] on branch develop) Minimal refactoring; limits where Memoizers are created

There is almost certainly more cleanup that can be done in terms of
where the reader stack is constructed and how it is stored. This is at
least a start, and is hopefully safe and minimally invasive.

See #10750.