Pushing NIfTIs to XNAT

October 22, 2018

The primary data format with which XNAT works has been and is still DICOM. XNAT can work with NIfTI but it generally has to be using secondary resources attached to subjects, experiments or scans.

This example shows how to get a dataset made of NIfTIs uploaded to XNAT without their corresponding DICOM versions.

Note: We will be using the pyxnat library to make REST API calls from Python code.

Step 1: create subjects

from pyxnat import Interface

config_file = '/home/grg/.xnat_bsc.cfg'
subjects = ['subject01', 'subject02', 'subject03']
project = 'MY_PROJECT'

# setup an XNAT connection
c = Interface(config=config_file)

# create subjects
for subject in subjects:

   uri =  '/data/projects/%s/subjects/%s'%(project, subject)

   response = c.put(uri)
   subject_uid = response.content
   print('New subject %s created!' %subject_uid)

Step 2: create experiments

experiments = {}
for subject in c.select.project(project).subjects():
     s = c.select.project(project).subject(subject.id())
     e = s.experiment('%s_MR1'%subject.label())

     options = {'xsiType':'xnat:mrSessionData'}
     e.create(**options)
     print('New experiment %s created!' %e.id())
     experiments[subject] = e

Step 3: create scans

scans = {}
for subject, experiment in experiments.items():
     options = {'xsiType':'xnat:mrScanData',
                'type':'DWI',
                'series_description':'DWI',
                'quality':'usable'}
     sc = experiment.scan('1')

     sc.create(**options)     
     print('New scan %s created!' %scan.id())
     scans[subject] = sc

Step 4: create resources

resources = {}
for subject, scan in scans.items():
    res = scan.resource('NIFTI')
    res.create()
    print('New resource %s created!' %res.id())
    resources[subject.label()] = res

Step 5: add files to just created resources

# NIfTIs files given in order matching with subjects
files = ['subject01.nii', 'subject02.nii', 'subject03.nii']

for subject, filename in zip(subjects, files):    
    resource = resources[subject]

     f = resource.file(filename)
     response = f.put(src=filename, format='NIFTI', content='NIFTI',
       extract=False, overwrite=True)

     file_uid = response.content
     print('File %s added!' %file_uid)

pyxnat has a function to remove a resource from XNAT:

uri = '/data/project/%s/subjects/%s/experiments/%s/'\
  %(project, subject_id, experiment_uid)
response = c.select(uri).delete(uri, delete_files=True)

The delete_files option deletes the files that come associated to the removed resources.

Note that for security reasons, the URI should contain both the subject's ID and the experiment's ID in order to proceed. Same goes for the upload step.

Many thanks to Jordi Huguet from BarcelonaBeta for such valuable and continuous help with this matter.

Please let me know if you liked this post by clicking the button below.