Question: given a .zarr what is the best way to say that it is .ome.zarr / ngff?

[yarikoptic]

Question: given a .zarr what is the best way to say that it is .ome.zarr / ngff?

[d-v-b]

Unfortunately, .omero is optional; The multiscales attribute might be better for this kind of thing. You might be interested in this proposal to fix the lack of a proper namespace in the metadata (among other changes).

[joshmoore]

It's being discussed to some degree in zarr-developers/zarr-specs#262 (ZEP0004)

[yarikoptic]

We may want to encourage, i.e. SHOULD, the use of .ome.zarr in the spec to improve awareness of OME-Zarr and be able to identify an OME-Zarr directly from a path / URL.

...

the use of .ome.zarr in the spec

Discussed by @normanrz at the 2024-04-03 OME Community Meeting: https://hackmd.io/So61knrQR0iLftd2LGAVjA

@normanrz :

and I would wholeheartedly support that -- only for UX!

[will-moore]

Yes, as @d-v-b says, check for multiscales.

[yarikoptic]

You might be interested in this proposal to fix the lack of a proper namespace in the metadata (among other changes).

and there it would become "explicit" through use of a URL pointing to a versioned ngff?

{
  "zarr_format": 3,
  "node_type": "group",
  "attributes": {
    "https://ngff.openmicroscopy.org/0.6": {
      "multiscales": [

or something else?

[joshmoore]

Yes, the single JSON file (zarr.json) would either have or not have the appropriate key.

[yarikoptic]

do you know if some more explicit formalization was discussed in zarr, e.g. to have something like "namespaces" or "extensions" on the top level which would have contained smth like

 [
  {
    "name": "ome",
    "version": "0.6",
    "schema": "https://ngff.openmicroscopy.org/jsonschemas/0.6" # 
  }
 ]

so, while dealing with a .zarr a developer of a tool could know what additional schemas implemented within that .zarr and potentially even load them to validate.
Just making it some "unique versioned URL" requires knowledge of its correspondence to specific effort/project, and since URL doesn't point to schema (it is 404 ATM) - not immediately usable for validation etc.

[yarikoptic]

any idea on how would it look for OME/NGFF if that ZEP0004 accepted in current form?

[thewtex]

We may want to encourage, i.e. SHOULD, the use of .ome.zarr in the spec to improve awareness of OME-Zarr and be able to identify an OME-Zarr directly from a path / URL.

[thewtex]

the use of .ome.zarr in the spec

Discussed by @normanrz at the 2024-04-03 OME Community Meeting: https://hackmd.io/So61knrQR0iLftd2LGAVjA