Closed monique2208 closed 1 week ago
The place to fix this will be in the schema:
I would add the following selector:
- (datatype != "meg" || entities.subject != "emptyroom")
Rather than matching on task-noise
, catching the meg/sub-emptyroom
seems better targeted to this case.
Would you be willing to open a PR over on the spec?
Rather than matching on
task-noise
, catching themeg/sub-emptyroom
seems better targeted to this case.
This would not work for me. task-noise recordings are allowed under individual participant sessions instead of sub-emptyroom.
Can you update this section of the schema (note that it's in the specification repository) with your proposal?
I'm not sure about the syntax, so someone should check, but something like this (replacing line 18):
- !(datatype == "meg" && (entities.subject == "emptyroom" || entities.task == "noise"))
It could be simpler since emptyroom
should have task-noise
, but I'd keep the subject check to make it a bit more robust.
Yes, that looks right though I would probably just distribute the not
:
$$\neg(A \land (B \lor C))$$ $$\neg A \lor \neg(B \lor C)$$ $$\neg A \lor \neg B \land \neg C$$
datatype != "meg" || entities.subject != "emptyroom" && entities.task != "noise"
I am getting a warning EVENTS_TSV_MISSING about the noise task that is part of empty room data for an meg dataset. Since this is explicitly how this task should be labelled according to the specification, should there not be an exception for
noise
as there is forrest
?we are actually upgrading this warning to an error for our system, is there a way to avoid this error for the noise task specifically through configuration?