DataBiosphere / azul

Metadata indexer and query service used for AnVIL, HCA, LungMAP, and CGP
Apache License 2.0
7 stars 2 forks source link

HCA schemas for descriptor lacks `drs_uri` property #6516

Open nadove-ucsc opened 2 months ago

nadove-ucsc commented 2 months ago

This property is only present in LungMap snapshots, which use a version of the schema from an un-merged feature branch.

https://raw.githubusercontent.com/HumanCellAtlas/metadata-schema/lungmap-phantom-files/json_schema/system/file_descriptor.json

nadove-ucsc commented 2 months ago

Assignee to consider next steps.

hannes-ucsc commented 1 month ago

I dusted off the old PR against the schema repo: https://github.com/HumanCellAtlas/metadata-schema/pull/1575 and asked for EBI's help in finishing it.

hannes-ucsc commented 1 month ago

I will continue tending to that new schema PR. Once it has been merged, we'll need to ask LungMAP to stop referring to the schema on the stale feature branch and instead point to the HCA schema repository. They'll likely need to update all of their current projects and TDR needs to re-import them.

hannes-ucsc commented 1 month ago

Enrique from EBI helped with the PR and I requested review on it.

hannes-ucsc commented 1 week ago

Schema PR has been merged and deployed.

hannes-ucsc commented 1 week ago

Additionally, LungMAP states that they updated all schema references to use the deployed schema but we shouldn't trust their word on it.

Assignee to create PR that requires all HCA schema URLs to use netloc schema.humancellatlas.org.

We may have cans that refer to other schema locations but those should only be accepted during tests.

dsotirho-ucsc commented 1 week ago

@hannes-ucsc: "Let's initially assume schema.humancellatlas.org is referenced by metadata JSON files, canned or live. Fix any test failures, then file a PR."