Fix version number in v05 multiscales

[dstansby]

Hi @dstansby - I was just testing a tool to generate TypeScript types from Pydantic - see https://github.com/marketplace/actions/pydantic-to-typescript

When I ran it on the _v05 classes...

$ pydantic2ts --module src/ome_zarr_models/_v05/image.py --output image05.ts

It gave me a Multiscale with version?: "0.4" | null; and I also don't see the "version": "0.5" which is required on ImageAttrs I think?

Apologies if this is already on your radar but I didn't recognise it in the issues above. Thanks.

/**
 * Model for the metadata of OME-Zarr data.
 *
 * See https://ngff.openmicroscopy.org/0.4/#image-layout.
 */
export interface ImageAttrs {
  /**
   * The multiscale datasets for this image
   *
   * @minItems 1
   */
  multiscales: [Multiscale, ...Multiscale[]];
  omero?: Omero | null;
  [k: string]: unknown;
}
/**
 * An element of multiscales metadata.
 */
export interface Multiscale {
  axes: Axis[];
  /**
   * @minItems 1
   */
  datasets: [Dataset, ...Dataset[]];
  version?: "0.4" | null;
  coordinateTransformations?: [unknown] | [unknown, unknown] | null;
  metadata?: {
    [k: string]: unknown;
  };
  name?: JsonValue | null;
  type?: {
    [k: string]: unknown;
  };
  [k: string]: unknown;
}

Originally posted by @will-moore in #88

[dstansby]

I've removed the version number in v05 over in #146 (thanks for reporting, this was not on our list and should have been!). For

and I also don't see the "version": "0.5" which is required on ImageAttrs I think?

ImageAttrs models the metdata that sits under the 'ome' key, so shouldn't contain the ome version, as that also sits directly under the 'ome' key. Does that make sense?

[will-moore]

From https://ngff.openmicroscopy.org/0.5/index.html#multiscale-md we get:

"attributes": {
    "ome": {
      "version": "0.5",
      "multiscales": [
        {
...

so we have "version" and "multiscales" as siblings.

If ImageAttrs models the metadata that sits under the 'ome' key, then I would expect it to include version?

[dstansby]

Woops, thanks for pointing that out... fix should be at #149, could you take a look and see if that looks right now?