search

catalystneuro / datta-lab-to-nwb

MIT License
1 stars 2 forks source link

HEK DataInterface #46

Closed pauladkisson closed 1 year ago

pauladkisson commented 1 year ago

Fixes #45.

pauladkisson commented 1 year ago

I'm getting this critical issue from nwbinspector bc the OnePhotonSeries isn't long enough (I double checked that time is the first dimension):

0.0  /Volumes/T7/CatalystNeuro/NWB/Datta/conversion_nwb/HEK_exp3.nwb: check_data_orientation - 'OnePhotonSeries' object at location '/acquisition/Signal1PSeries'
       Message: Data may be in the wrong orientation. Time should be in the first dimension, and is usually the longest dimension. Here, another dimension is longer.
CodyCBakerPhD commented 1 year ago

It's a static image, right? That would be expected in this case - that is the one critical importance check that is 'heuristic' and can be wrong in a situation just like this

bendichter commented 1 year ago

Do you guys think a static image would be better as an Image object?

CodyCBakerPhD commented 1 year ago

There's a bunch of metadata more common to ophys, e.g., grid_spacing that applies here

So really, the best solution would be to have a static image type avaialble in ophys

Can maybe kick that up as a part of the ndx-micropscopy extension during the dev days

pauladkisson commented 1 year ago

It is not a static image, just very low sampling frequency (0.25Hz) and relatively short recording durations (~1min). Ex. the shape of the signal in experiment 3 is (235, 520, 688). I suppose I can just suppress this issue if there's no better fix.

CodyCBakerPhD commented 1 year ago

I suppose I can just suppress this issue if there's no better fix.

Yep, there's nothing to even fix in that case. It's just that that check is a heuristic that fails in cases like that

CodyCBakerPhD commented 1 year ago

@pauladkisson LGTM except for the requirements file