Mentioned at https://forum.image.sc/t/fast-online-stack-browser/74691/23

E.g. with the link from post above: https://allen-cell-animated.github.io/website-3d-cell-viewer/?url=https://animatedcell-test-data.s3.us-west-2.amazonaws.com/variance/1.zarr

or with IDR example https://allen-cell-animated.github.io/website-3d-cell-viewer/?url=https://uk1s3.embassy.ebi.ac.uk/idr/zarr/v0.4/idr0101A/13457537.zarr

NB: It seems that the viewer only works with 5D data - E.g. https://allen-cell-animated.github.io/website-3d-cell-viewer/?url=https://uk1s3.embassy.ebi.ac.uk/idr/zarr/v0.4/idr0062A/6001240.zarr fails with

Uncaught (in promise) RangeError: too many indices for array; expected 4, got 5
    at checkSelectionLength (core.mjs?3f36:250:1)
    at replaceEllipsis (core.mjs?3f36:334:1)
    at normalizeArraySelection (core.mjs?3f36:277:1)
    at new BasicIndexer (core.mjs?3f36:412:1)
    at ZarrArray.getBasicSelectionND (core.mjs?3f36:2715:1)
    at ZarrArray.getBasicSelection (core.mjs?3f36:2711:1)
    at ZarrArray.get (core.mjs?3f36:2696:1)
    at eval (FetchZarrWorker.js?ed40:65:1)

Afaik the images in the idr catalog are all in C-order with tczyx axes ordering. I think it would be useful to also check for other configurations such as F-order with cxyz axes. Here is a sample dataset: https://data-humerus.webknossos.org/data/zarr/scalable_minds/l4dense_motta_et_al_demo/color/ (uses . dim sep, tho).

Replace the plate sample in https://ome.github.io/ome-ngff-tools/ - the sample present there (https://uk1s3.embassy.ebi.ac.uk/idr/zarr/v0.4/idr0001A/2551.zarr) does not import to OMERO.

See also ome/ZarrReader#43

cc @will-moore

We could list other tools, e.g.

https://github.com/thewtex/ngff-zarr

The Z downsample example uses this dataset:

https://minio-dev.openmicroscopy.org/idr/v0.4/idr0077/9836832_z_dtype_fix.zarr

It appears that the base resolution contains data, but the subsequent resolutions are all 0:

import tensorstore as ts
x = ts.open({'driver': 'zarr', 'kvstore': 'https://minio-dev.openmicroscopy.org/idr/v0.4/idr0077/9836832_z_dtype_fix.zarr/1/'}).result()
data = x.read().result()
assert data.min() == 0
assert data.max() == 0

Using this dataset for testing viewer support naturally leads to confusion...

Collection of urls pointing to public data
We can then decide where that collection should be presented

• https://idr.github.io/ome-ngff-samples/
• https://uk1s3.embassy.ebi.ac.uk/bia-zarr-test/bia_examples.html
• https://s3.embl.de/i2k-2020/platy-raw.ome.zarr (https://s3.embl.de/i2k-2020)
• https://zenodo.org/record/7120354#.Y2ziK-zP1Oc
• https://webknossos.org/datasets/scalable_minds/l4dense_motta_et_al_demo#2924,4474,1770,0,3.4
• Links to consider in https://hackmd.io/@ome/HJq5cSNV5

We need to add a table or a page dedicated to conversion tools

• https://github.com/FrancisCrickInstitute/OmeSliCC
• https://www.glencoesoftware.com/products/ngff-converter/
• bf2raw cli usage

E.g: https://avivator.gehlenborglab.org/?image_url=https://uk1s3.embassy.ebi.ac.uk/idr/zarr/v0.3/idr0079A/9836998.zarr
code at https://github.com/hms-dbmi/viv

Import OME-NGFF images into a public group on e.g. https://omero.lifesci.dundee.ac.uk/ server as examples of how OMERO handles these images.
Then add urls to the table (similar to webknossos)

I think it would be useful to include rows that indicate support for storage backends, because most tools only support a subset of the common backends.

Imo the most popular backends are:

• S3 (public and private buckets) and compatible object storage
• HTTP(S)
• Local file system
• GCS
• Azure blob storage

15% of male humans, to my best knowledge, have a red-green weakness.
Would be nice to find another color scheme.
Have to look for one.
We need 4 colors, right?

See https://github.com/bigdataviewer/bigdataviewer-omezarr

A user coming to the page from a particular background (e.g. "vEM") might expect a few summary rows in order to better understand which tool is recommended. Admittedly, the definition of these rows is likely trickier than the ones to date.

cc: @normanrz