Closed Najmahan closed 9 years ago
@Najmahan do you also want that we test that all the files have the same datatype, data_type and pixdims? We could also raise a warning in case different file_types are used. See "definitions" below, obtained from fslinfo.
data_type FLOAT32 dim1 170 dim2 256 dim3 256 dim4 1 datatype 16 pixdim1 1.200000 pixdim2 1.000000 pixdim3 1.000000 pixdim4 1.000000 cal_max 0.0000 cal_min 0.0000 file_type NIFTI-1+
Warning yes, but aborting the process no (unless the violation is in dim4, time). The preprocessing will be normalizing data; and I suspect inconstancies in data type and dimension may be common. I can see it being important for 'precision' studies though, but not sure if the typical users are bothered by this.
OK, to keep it simple for now, I focused on dim* (error if different) and pixdim4 (warning if different as mentioned in #26). The other params are ignored.
I wouldn't worry about dim{1,2,3}. only dim 4 should be the same. One caveat of this is that If the data is scrubbed, it may not have the same dim4 from one dataset to next. This should be documented as a 'limitation'. [we did previously discuss setting the frames in the configuration file, but the idea was abandoned for simplicity.]
in TICA & GICA, make sure all input files have the same dimension. If not, stop the task, create a report of filenames and their dimensions and the user will be instructed to go to the parameter file and de-select the violating files.