Not parsing "multiscales" metadata if "omero" metadata is missing

[yuriyzubov]

Hello!
The official OME-NGFF spec states that omero metadata is optional.

When I try to open a zarr group that contains multi-scale image pyramid with iohub.open_ome_zarr() method, with the following .zattrs metadata:

 "multiscales": [
       {
         "name": "ds_1",
         "version": "0.4",
         "axes": [
           { "name": "z", "type": "space", "unit": "micrometer" },
           { "name": "y", "type": "space", "unit": "micrometer" },
           { "name": "x", "type": "space", "unit": "micrometer" }
         ],
         "datasets": [
           {
             "path": "s0",
             "coordinateTransformations": [
               {
                 "type": "scale",
                 "scale": [0.5, 0.5, 0.5]
               }
             ]
           }
         ]
       }
     ]

It invokes the warning method, since the omero metadata is missing, but self.metadata.multiscales remains empty as well.

Perhaps it is better to make the omero field optional in the definition of ImagesMeta?

class ImagesMeta(MetaBase):
    """Metadata needed for 'Images' (or positions/FOVs) in an OME-NGFF dataset.
    https://ngff.openmicroscopy.org/0.4/index.html#image-layout"""

    multiscales: list[MultiScaleMeta]
    omero: Optional[OMEROMeta] = None

And rewrite _parse_meta() method?

def _parse_meta(self):
    multiscales = self.zattrs.get("multiscales")
    omero = self.zattrs.get("omero")
    if multiscales:
        try:
            self.metadata = ImagesMeta(
                multiscales=multiscales, omero=omero
            )
            self.axes = self.metadata.multiscales[0].axes
            if omero:
                self._channel_names = [
                    c.label for c in self.metadata.omero.channels
                ]
        except ValidationError:
            self._warn_invalid_meta()
    else:
        self._warn_invalid_meta()

so the multiscales metadata is parsed properly, even when the omero attribute is missing.

[ziw-liu]

@yuriyzubov Thanks for reporting this bug! Would you want to turn your code snippets into a PR?