When running mriqc I haven’t been able to get it to recognize the folder paths/files in my .bidsignore file. This file works when confirming my data is in bids format, but not with mriqc. Is mriqc supposed to ignore the paths in the .bidsignore file? And, if not is there anyway this can be added as a functionality?
Can someone please provide input on my previous post? I’m not sure if I am missing something obvious and need to troubleshoot more, or if the .bidsignore file is simply not set to work with mriqc.
Thank you for your input and time.
.bidsignore is specifically a tool for the BIDS validator. The purpose is to enable dataset curators to include useful files that are not yet standardized under BIDS. It is not a general-purpose concept that BIDS apps are expected (or even encouraged) to use.
Could you give an example of an ignored file, and what MRIQC does with it that you don’t want?
Thank you for the input.
The purpose that I was hoping to use it for was to include useful files in my bids subject folder that are not yet standardized for use in MRIQC.
For example, this is one line from my .bidsignore file:
And below is the actual path with a file with in it that I don’t want to be run through MRIQC or fMRIPREP. It is a restricted FOV, high resolution axial anatomical scan that only cuts through the hippocampus.
I can use this .bidsignore file with fMRIPREP and the bids validator to ignore this non-standard hippocampal slice no problem, and thus ignore this non-standard file. This is advantageous because I can keep all of my data together. I would also like to be able to do this with MRIQC. But the only way to run MRIQC using this approach is to remove this misc folder completely or zip it (I don’t think it is a good practice to move subject data from their bids folder, and so I’ve been zipping these data, but this is not an ideal solution).
All this being said, maybe I am missing something obvious, but why would users not be encouraged to take advantage of the .bidsivore file? It seems like a highly useful, intuitive way to keep standard and non-standard data for each subject within their respective bids folders while also taking advantage of the bidsvalidator, FMRIPREP and MRIQC tools.
Thanks for any input!
Thanks. That makes it more concrete.
Just to be clear, fMRIPrep does not respect the
.bidsignore file. The difference here is that fMRIPrep looks for
T1w images with the query
get(suffix='T1w', dataype='anat'), while MRIQC appears to use a different mechanism.
I think a bug fix will be needed, either to be more precise about the query or to enable validation. Could you open an issue for this?
The basic problem is that “ignoring” can mean two things: 1) Pretend the file doesn’t exist (ignore the file); 2) Pretend the file is valid (ignore the error).
These are identical for the validator, since it only exists to report errors.
For an app, following (1) means leaving the file alone, following (2) means using it if it matches some expected criterion. In many cases, the reason somebody includes an invalid file in a dataset is precisely because they have some tool that expects it. Making it mandatory for all tools to ignore the files rather than ignore the errors would work against that.
Thanks for the input!
Thanks for clarifying the .bidsignore file. I mistakenly thought it was doing something in fMRIprep. I just opened an issue suggestion to make the query for fMRIPrep and MRIQC the same.