List invalid samples in README for testing

[will-moore]

Maybe move these somewhere more permanent first?

• Invalid version (fails to load schema) https://deploy-preview-16--ome-ngff-validator.netlify.app/?source=https://minio-dev.openmicroscopy.org/idr/v0.3/invalid/9836842_version.zarr
• Various dataset/.zattrs errors (see Check datasets shape and dtype #16) https://minio-dev.openmicroscopy.org/idr/v0.3/invalid/9836842_dimCount.zarr
• axes invalid (schema-validation): https://minio-dev.openmicroscopy.org/idr/v0.3/invalid/9836842_axes.zarr
• Image contains invalid labels: https://minio-dev.openmicroscopy.org/idr/v0.4/invalid/6001240.zarr

[joshmoore]

I'd suggest test-data:

mc ls uk1/idr/zarr/test-data
[2022-10-24 09:13:06 CEST]     0B 64x64-fake-v0.1/
[2022-10-24 09:13:06 CEST]     0B 64x64-fake-v0.2/
[2022-10-24 09:13:06 CEST]     0B bf2raw/
[2022-10-24 09:13:06 CEST]     0B idr0044/
[2022-10-24 09:13:06 CEST]     0B labels-0.1.3-dev4/
[2022-10-24 09:13:06 CEST]     0B test-v0.1.zarr/
[2022-10-24 09:13:06 CEST]     0B test-v0.2.zarr/

though we may want to add some more structure to it.

[will-moore]

Another sample I created for testing various NGFF viewers at ome/ngff#139: has downsample factor of 4: created by generating with bioformats2raw (v0.2), then removing alternating levels from the pyramid and updating the .zattrs to make a v0.4 sample:
https://minio-dev.openmicroscopy.org/idr/v0.4/idr0082/9846318.zarr/0

[joshmoore]

maybe an explicit uk1/idr/zarr/invalid-data would be safer.