While trying out metalad using nifti1 extractor as the one which needs "sparse"
access to actual file content, it was crashing due to
TypeError: Object of type bytes is not JSON serializable
which were due to the fact that some fields, e.g.
*(Pdb) p str(header['descrip'])
"b'FreeSurfer Aug 11 2009'"
are byte numpy arrays... I decided to just to the final tune up for all values
regardless where they came from - I do not envision a use case where they should
be a byte string.
If it would be the case -- metalad would need to be the one to handle that
somehow I guess.
While trying out metalad using nifti1 extractor as the one which needs "sparse" access to actual file content, it was crashing due to
which were due to the fact that some fields, e.g.
are byte numpy arrays... I decided to just to the final tune up for all values regardless where they came from - I do not envision a use case where they should be a byte string.
If it would be the case -- metalad would need to be the one to handle that somehow I guess.