@sstucker and @jayd1860 , we need to update our data that is linked here. I guess we should do this once the first official release is made... hopefully very soon... and then we put up the snirf files that meet that official release.

@fangq has discussed this with me offline and raised the issue that example files should be 1) small, 2) address specific test cases, 3) be many examples that cover different scenarios. I agree with all this and in a perfect world someone would fix this today.

But my suggestion in this PR is the lesser of two evils. Currently the linked data is simply wrong. I know this because users have wasted days using it and complained to me. I agree with his points, and ideally the current examples would be fixed. But no one has done that, and every week the current files are linked it wastes peoples times and makes this project look unprofessional.

While my linked data isn't build for this purpose (they are example data for MNE and BIDS), they are correct according to the current specification. We can simply swap the link back to the ideal dataset when its updated.

with the above 2363e1c commit, can you let me know what else should be fixed in fNIRS-samples?

Are the sample files now compatible with the spec? If so, then just close this.

I just downloaded data from the readme link and tried snirf-samples-master/test/minimum_example.snirf and snirf-samples-master/basic/Simple_Probe.snirf and both threw errors in my reader. So I guess this isn't fixed (or my reader is broken).

I don't have time at the moment to specifically check what these issues were. I will leave this to you as owner of that data if you wish to fix the issue rather than merge this PR.

Edit: I also used the python snirf validator being developed by @dboas lab and the sample files dont even read in to it, just crashes. What software are you reading these files with to test them? The validator reads both my SNIRF files and the Homer SNIRF files, so yours must be something different. This just highlights the confusion/issue here and the time being wasted comparing files from different places.

go ahead and replace the sample submodule with yours.

Like this? Sorry I haven't used submodules before, might have made an error here.

done. see my above commit. submodules are pointers to a specific commit in a specific repo. after switching the URL in .gitmodules, you have to check out the updated module and run git add samples to update the commit hash.

nonetheless, my understand is that your sample repo serves as examples for BIDS conversion. We still have a need for a sample repo that provides small but wide variety of SNIRF data types (simple probe, complex probe, CW/TD/FD data, with/wo aux, ...) in order to perform efficient testing of parsers/validators/software.

if a validator crashes on the old sample data, something should be fixed, more likely it is the validator's bug. anyways, will look into when I have time.