Check datasets shape and dtype

will-moore commented 1 year ago

This PR validates a few things that aren't checked by the schema validation:

1 - All dataset.shapes are the same length (number of dimensions is the same for all)
2 - All dataset.shapes are the same length as the length of multiscale.axes.
3 - Checks dimension_separator is / when version is 0.2-0.4
4 - All dtypes are the same (see https://github.com/ome/ngff/pull/154)
5 - Fixes handling of invalid version (previously got stuck with 'loading...'

To test:

NB: This currently asserts that the dtypes are all equal for v0.4 (so this PR can be tested). Need to revert to testing v0.5 and above BEFORE merging!

For other tests above, we really need some invalid sample data available... Might try to make some...

cc @constantinpape

netlify[bot] commented 1 year ago

Deploy Preview for ome-ngff-validator ready!

Name	Link
Latest commit	58094dd2c62388bd5fb8f2e375dbf8ecdcb8f632
Latest deploy log	https://app.netlify.com/sites/ome-ngff-validator/deploys/6356cd12c609250008c85bf6
Deploy Preview	https://deploy-preview-16--ome-ngff-validator.netlify.app/
Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

To edit notification comments on pull requests, go to your Netlify site settings.

joshmoore commented 1 year ago

Nice, @will-moore! I got some "Error loading" messages from the URLs above.

Also: it might be worth considering how we maintain multiple validators for these non-json-schema constraints. Would defining "rule numbers" for each of them be feasible/useful?

will-moore commented 1 year ago

I got some "Error loading" messages from the URLs above.

As expected! The "Error loading" is what you get if you use an invalid version, since we use the version to construct the schema URL. We don't try to maintain a list of all valid versions in this tool. I guess we could try to handle that 404 with a more custom message.

Other errors for the other test URL are also expected (violations of rules 1, 2, 3 & 4.

joshmoore commented 1 year ago

For me, "error loading" meant something like "the data is not accessible". :+1: for "incompatible version: ${version}", etc.

will-moore commented 1 year ago

@joshmoore: better?

Sorry, wrong image - this is the right one...

joshmoore commented 1 year ago

I like it! Thanks.

One last thing: did the file for "4: dtype mismatch" get deleted?

will-moore commented 1 year ago

Hmmm - not sure why that got deleted. But here's another with the same dtype issue: https://deploy-preview-16--ome-ngff-validator.netlify.app/?source=https://minio-dev.openmicroscopy.org/idr/v0.4/idr0077/9836832_z.zarr

ome / ome-ngff-validator