No error/warning when IDs in participant.tsv do not match directory names

[JuliusWelzel]

For a workshop, I wanted to showcase the BIDS-Validator. One thing the participants noticed, is that if the IDs in the participants.tsv file do not match any of the subject's directories IDs, no warning or error is thrown. Is this behavior intended?

[mateuszpawlik]

We had a similar case last week, also at a workshop 😉

When subject directories are there, and there are no records in the participants.tsv file, BIDS Validator (1.14.6) doesn't raise any error.

I though this would be caught by issue 49 (PARTICIPANT_ID_MISMATCH), but it seems that it's implemented only the other way.

[effigies]

Unclear if this issue was ever implemented in the legacy. It certainly hasn't been for some time, although it is now implemented in the schema validator.

PRs to the legacy validator are still accepted, but our priority at the moment is completing the schema validator.

[mateuszpawlik]

I don't understand. Are there two validators? Could you explain the difference between legacy and schema validator? How can I use the schema validator if it's more robust?

Something is implemented:

bids-validator/bids-validator/validators/bids/subjects.js

Lines 16 to 22 in 2e66a5a

	issues.push(
	new Issue({
	code: 49,
	evidence: evidence,
	file: participants.file,
	}),
	)

[effigies]

Apologies, we have done a bad job of communicating this on the repository. The BIDS validator is a large amount of custom code that gets more baroque over time. A few years ago the steering group approved the idea of expanding the then-nascent BIDS schema into something that could encode most validatable rules, and we would rewrite the validator as an interpreter of the schema.

That turned out not to be possible piece-meal, and so the new interpreter is being built in the shell of the old. You can find it (with a README) here: https://github.com/bids-standard/bids-validator/tree/master/bids-validator/src

For some more context: https://docs.google.com/presentation/d/1ldEbElaFm__jtkLoEcn2PQ-LGj1dfmdjWxDvE11eiNk/edit#slide=id.p

Anyway, we are now working to determine any remaining gaps in the schema validator and hope to switch to it by default very soon.

[mateuszpawlik]

I'm happy to try out the new validator if that's helpful.