Refactor PlateLabels

[joshmoore]

https://github.com/ome/ome-zarr-py/pull/62 introduced the PlateLabels class which requires some special casing in order to provide 2 nodes from the same URL. One idea for working around this was to use a #labels suffix or similar to differentiate.

see: https://github.com/BioSchemas/specifications/issues/475 however for potential RDF issues with #suffixes.

[AlasdairGray]

@joshmoore I can't see enough detail in #62 to understand your context here, however, if your #suffixes are to define something within the page then they will be fine. The issue in Bioschemas was that the URL of the page was being used for something within the page rather than the page itself.

[joshmoore]

Thanks for the context, @AlasdairGray. I'll look into using BMUSE to scrape the suggestion I have here to see if it's a problem or not.

[AlasdairGray]

Happy to help review the output if you need it.

[joshmoore]

:+1: Ta

[sbesson]

After some investigation I am pretty sure this is the source of the ome_zarr download issue reported in https://forum.image.sc/t/hcs-ome-zarr-data-unable-to-download/57114.

$ ome_zarr info https://uk1s3.embassy.ebi.ac.uk/idr/zarr/v0.1/plates/2551.zarr/
WARNING:ome_zarr.io:version mismatch: detected:FormatV01, requested:FormatV03
WARNING:ome_zarr.io:version mismatch: detected:FormatV03, requested:FormatV01
WARNING:ome_zarr.io:version mismatch: detected:FormatV01, requested:FormatV03
WARNING:ome_zarr.io:version mismatch: detected:FormatV03, requested:FormatV01
WARNING:ome_zarr.io:version mismatch: detected:FormatV03, requested:FormatV01
WARNING:ome_zarr.io:version mismatch: detected:FormatV01, requested:FormatV03
WARNING:ome_zarr.io:version mismatch: detected:FormatV03, requested:FormatV01
https://uk1s3.embassy.ebi.ac.uk/idr/zarr/v0.1/plates/2551.zarr/ [zgroup]
 - metadata
   - Plate
 - data
   - (1, 2, 16, 520, 1032)
https://uk1s3.embassy.ebi.ac.uk/idr/zarr/v0.1/plates/2551.zarr/ [zgroup]
 - metadata
   - PlateLabels
 - data
   - (1, 1, 16, 520, 1032)

leads to the top-level path being duplicated in https://github.com/ome/ome-zarr-py/blob/35d56e81a3bbc11cfb355770d29e0c1e5d8c041c/ome_zarr/utils.py#L61-L63 and the sorted operation below in https://github.com/ome/ome-zarr-py/blob/35d56e81a3bbc11cfb355770d29e0c1e5d8c041c/ome_zarr/utils.py#L75 failing as all the elements are duplicated and of equal weight

[jluethi]

Is the refactor of PlateLabels still ongoing? I've been trying to get some Label data into our test cases and it works well to display the label data if I look at a single field of view (dragging zarr_plate.zarr/B/03/0) in napari:

But if I drag the whole plate (zarr_plate.zarr) into napari, it only loads the intensity channels and seems to ignore the label channel.

(Same if I drag in a specific well, e.g. zarr_plate.zarr/B/03

I saw that some PlateLabels functionality was disabled earlier this year here to make plate loading work: https://github.com/sbesson/ome-zarr-py/pull/1 Is the fact that labels aren't loaded anymore for plates an intended outcome of this?

[will-moore]

Need to generate some new test data to have a look at this again...

[will-moore]

Ah - I found https://gist.github.com/will-moore/0f4cb6b1fdd60a255fcbb956a54a645e to create some data... With a lot of hacks to the code, I got some labels showing...

But even this is not working properly yet (only seeing a single image channel) So I'll try fixing and cleaning this up...