Hey! We met at OHBM in Brisbane and wanted to follow up on that idea we discussed.
We'd like our different neuro tools (Styx, niwrap, boutiques, etc.) to have a standardized way of describe/identify file formats.
Since you've already got many of these format definitions here, what do you think about extracting them into a simple JSON registry that other tools could use? Just the identifiers, extensions, maybe magic numbers and some other metadata.
Vaguely something like:
{
"nifti-1": {
"extensions": [".nii", ".nii.gz"],
"magic": "6E2B3100"
},
"dicom": {
"extensions": [".dcm", ".ima"],
"magic": "4449434D"
}
}
(Simplified here of course)
Would make it way easier for different frameworks to stay consistent. Happy to help put it together if you're interested!
Edit: Could vaguely follow/align with the IANA media types
Hey! We met at OHBM in Brisbane and wanted to follow up on that idea we discussed.
We'd like our different neuro tools (Styx, niwrap, boutiques, etc.) to have a standardized way of describe/identify file formats.
Since you've already got many of these format definitions here, what do you think about extracting them into a simple JSON registry that other tools could use? Just the identifiers, extensions, maybe magic numbers and some other metadata.
Vaguely something like:
{ "nifti-1": { "extensions": [".nii", ".nii.gz"], "magic": "6E2B3100" }, "dicom": { "extensions": [".dcm", ".ima"], "magic": "4449434D" } }(Simplified here of course)
Would make it way easier for different frameworks to stay consistent. Happy to help put it together if you're interested!
Edit: Could vaguely follow/align with the IANA media types