warn/error on bidsignore entries matching "legit" files

[yarikoptic]

trafficstars

I initially thought that it might be more of interest to OpenNeuro itself, but then decided that it better be addressed at the validator level.

Look through .bidsignore files among OpenNeuro datasets:

$> grep sub */.bidsignore
ds000117/.bidsignore:**/sub-*_ses-mri_run-*_echo-*_FLASH.nii.gz
ds000224/.bidsignore:sub-*/ses-struct*/anat/*veno*
ds000234/.bidsignore:sub-01/func/
ds000234/.bidsignore:sub-02/func/
ds000234/.bidsignore:sub-03/func/
ds000234/.bidsignore:sub-04/func/
ds000234/.bidsignore:sub-05/func/
ds000235/.bidsignore:sub-01/func/
ds000235/.bidsignore:sub-02/func/
ds000235/.bidsignore:sub-03/func/
ds000235/.bidsignore:sub-04/func/
ds000236/.bidsignore:sub-01/func/
ds000236/.bidsignore:sub-02/func/
ds000236/.bidsignore:sub-03/func/
ds000236/.bidsignore:sub-04/func/
ds000236/.bidsignore:sub-05/func/
...
ds000240/.bidsignore:sub-61/func/sub-61_task-restEyesOpen_asl.nii.gz
ds000240/.bidsignore:sub-62/func/sub-62_task-restEyesOpen_asl.nii.gz
ds000240/.bidsignore:sub-63/func/sub-63_task-restEyesOpen_asl.nii.gz
ds000248/.bidsignore:**/sub-*_acq-flipangle*-*_MEFLASH.nii.gz
ds001219/.bidsignore:/sub-*/EEG/**
ds001420/.bidsignore:sub-01/ses-01/pet/
ds001420/.bidsignore:sub-01/ses-02/pet/
ds001420/.bidsignore:sub-02/ses-01/pet/
ds001420/.bidsignore:sub-02/ses-02/pet/
ds001421/.bidsignore:sub-01/ses-01/pet/
ds001421/.bidsignore:sub-01/ses-02/pet/
ds001454/.bidsignore:sub-01/ses-*/beh
ds001454/.bidsignore:sub-02/ses-*/beh
ds001454/.bidsignore:sub-03/ses-*/beh
ds001454/.bidsignore:sub-04/ses-*/beh
ds001454/.bidsignore:sub-05/ses-*/beh
ds001454/.bidsignore:sub-06/ses-*/beh
ds001454/.bidsignore:sub-07/ses-*/beh
ds001454/.bidsignore:sub-08/ses-*/beh
ds001454/.bidsignore:sub-09/ses-*/beh
ds001454/.bidsignore:sub-10/ses-*/beh
ds001454/.bidsignore:sub-11/ses-*/beh
ds001454/.bidsignore:sub-12/ses-*/beh
ds001454/.bidsignore:sub-13/ses-*/beh
ds001454/.bidsignore:sub-14/ses-*/beh
ds001454/.bidsignore:sub-15/ses-*/beh
ds001454/.bidsignore:sub-16/ses-*/beh
ds001454/.bidsignore:sub-17/ses-*/beh
ds001454/.bidsignore:sub-18/ses-*/beh
ds001454/.bidsignore:sub-19/ses-*/beh
ds001454/.bidsignore:sub-20/ses-*/beh
ds001454/.bidsignore:sub-21/ses-*/beh
ds001454/.bidsignore:sub-22/ses-*/beh
ds001454/.bidsignore:sub-23/ses-*/beh
ds001454/.bidsignore:sub-24/ses-*/beh
ds001512/.bidsignore:/sub-*/EEG/**
ds001547/.bidsignore:sub*/stat/*.nii
ds001750/.bidsignore:**/sub-*/ses-*/meg/sub-*_ses-*_task-*_meg.ds/sub-*_ses-*_task-*_meg.eeg
ds001750/.bidsignore:**/sub-*/ses-*/meg/sub-*_ses-*_task-*_meg.ds/bad.segments
ds001750/.bidsignore:**/sub-*/ses-*/meg/sub-*_ses-*_task-*_meg.ds/BadChannels
ds001750/.bidsignore:**/sub-*/ses-*/meg/sub-*_ses-*_task-*_meg.ds/hz*.ds/BadChannels
ds001750/.bidsignore:**/sub-*/ses-*/behavdata/*
ds001750/.bidsignore:**/sub-*/ses-*/hpi/*
ds001769/.bidsignore:**/sub-*/ses-movie/func/*_desc-denoisedSm5_bold.nii.gz
ds001769/.bidsignore:**/sub-*/ses-movie/func/*_desc-denoisedUnsm_bold.nii.gz
ds001769/.bidsignore:**/sub-*/ses-movie/func/*_desc-denoisedSm5_bold.json
ds001769/.bidsignore:**/sub-*/ses-movie/func/*_desc-denoisedUnsm_bold.json
ds001769/.bidsignore:**/sub-*/ses-movie/func/*_components.nii.gz
ds001769/.bidsignore:**/sub-*/ses-movie/func/*_mixing.tsv
ds001769/.bidsignore:**/sub-*/ses-movie/func/*_decomposition.json
ds001769/.bidsignore:**/sub-*/ses-movie/func/*_componentLabels.txt
ds001840/.bidsignore:/sub-*/eyetrack
ds001980/.bidsignore:/sub-*/fmap/sub-*_acq-multiband_dir-PA_dwi.nii.gz
ds002015/.bidsignore:**/anat/sub*.nii.gz
ds002015/.bidsignore:**/anat/sub*.json
ds002016/.bidsignore:**/anat/sub*.nii.gz
ds002016/.bidsignore:**/anat/sub*.json

and you can see, that some smart scientists decided to just "workaround" bids-validator by pretty much matching many if not all files, which otherwise are, or should be, BIDS compliant. (e.g. patterns such as **/anat/sub*.nii.gz). So I wondered if bids-validator may be somehow could outsmart them? E.g. whenever bidsignore entry matches a file which is legit BIDS file, it might as well worth raising an error to prevent such "swallow them all" entries to bidsignore