To ensure that the data stored in a NIRSSeries meets the desired requirements, I recommend modifying the schema for NIRSSeries to define the "data" dataset explicitly as:
- name: data
dtype: numeric
dims:
- num_times
- num_channels
shape:
- null
- null
doc: Recorded NIRS data.
Thanks @rly, great point and thanks for the helpful suggestion on how to resolve it! We'll be sure to include this in the next version. We're planning to do more frequent updates and releases.
The class diagram for
NIRSSeries
says thatdata
should have a numeric dtype and 2 dimensions (time x channels), but as far as I can tell, the schema https://github.com/agencyenterprise/ndx-nirs/blob/main/spec/ndx-nirs.extensions.yaml#L179-L186 specifies no such requirements.NIRSSeries
extendsTimeSeries
and does not modify any properties of the "data" dataset.To ensure that the data stored in a
NIRSSeries
meets the desired requirements, I recommend modifying the schema forNIRSSeries
to define the "data" dataset explicitly as:See https://github.com/NeurodataWithoutBorders/nwb-schema/blob/dev/core/nwb.ecephys.yaml#L18-L42 and https://github.com/NeurodataWithoutBorders/nwb-schema/blob/dev/core/nwb.ecephys.yaml#L131-L141 as examples.