--cifti-output issues for data processed in fsLR-32k space in 20.0.1rc1

[helloTC]

Hi, I installed the latest version of fmriprep (v20.0.1rc) in my servers by using pip command. I would like to get processed data in fsLR-den32k space. It looks fsLR no long need to be included in the --output-spaces in this lastest version because after running
fmriprep test_fmriprep test_fmriprep_output participant -w work --output-spaces fsLR --cifti-output
I got an error as

ValueError: space identifier "fsLR" is invalid.
Valid identifiers are: MNI152Lin, MNI152NLin2009cAsym, MNI152NLin6Asym, MNI152NLin6Sym, MNIInfant, MNIPediatricAsym, NKI, OASIS30ANTs, PNC, fsaverage, T1w, T2w, anat, fsnative, func, run, sbref, session.

While I noticed in documentation of the v20.0.1rc still implied that the users need to include fsLR in the --output-spaces, it seems a bug here.

Issue #1887 allows CIFTI output when using --cifti-output and will compliant with HCP grayordinate space (cortex in fsLR, subcortical in MNI152NLin6Asym).
However, after running command only with --cifti-output: fmriprep test_fmriprep test_fmriprep_output participant -w work -cifti-output, I got one another error as

Process Process-2:
Traceback (most recent call last):
File "/usr/local/neurosoft/miniconda3/lib/python3.6/multiprocessing/process.py", line 258, in _bootstrap
self.run()
File "/usr/local/neurosoft/miniconda3/lib/python3.6/multiprocessing/process.py", line 93, in run
self._target(*self._args, **self._kwargs)
File "/usr/local/neurosoft/miniconda3/lib/python3.6/site-packages/fmriprep/cli/run.py", line 676, in build_workflow
work_dir=str(work_dir),
File "/usr/local/neurosoft/miniconda3/lib/python3.6/site-packages/fmriprep/workflows/base.py", line 267, in init_fmriprep_wf
use_syn=use_syn,
File "/usr/local/neurosoft/miniconda3/lib/python3.6/site-packages/fmriprep/workflows/base.py", line 623, in init_single_subject_wf
use_syn=use_syn,
File "/usr/local/neurosoft/miniconda3/lib/python3.6/site-packages/fmriprep/workflows/bold/base.py", line 945, in init_func_preproc_wf
repetition_time=metadata['RepetitionTime'])
File "/usr/local/neurosoft/miniconda3/lib/python3.6/site-packages/fmriprep/workflows/bold/resampling.py", line 744, in init_bold_grayords_wf
for hemi in 'LR'
File "/nfs/s2/userhome/huangtaicheng/.local/lib/python3.6/site-packages/nipype/interfaces/base/traits_extension.py", line 425, in validate
value = super(MultiObject, self).validate(objekt, name, newvalue)
File "/usr/local/neurosoft/miniconda3/lib/python3.6/site-packages/traits/trait_types.py", line 2573, in validate
return TraitListObject(self, object, name, value)
File "/usr/local/neurosoft/miniconda3/lib/python3.6/site-packages/traits/trait_handlers.py", line 2474, in init
raise excp
File "/usr/local/neurosoft/miniconda3/lib/python3.6/site-packages/traits/trait_handlers.py", line 2466, in init
value = [validate(object, name, val) for val in value]
File "/usr/local/neurosoft/miniconda3/lib/python3.6/site-packages/traits/trait_handlers.py", line 2466, in
value = [validate(object, name, val) for val in value]
File "/nfs/s2/userhome/huangtaicheng/.local/lib/python3.6/site-packages/nipype/interfaces/base/traits_extension.py", line 329, in validate
value = super(File, self).validate(objekt, name, value, return_pathlike=True)
File "/nfs/s2/userhome/huangtaicheng/.local/lib/python3.6/site-packages/nipype/interfaces/base/traits_extension.py", line 134, in validate
self.error(objekt, name, str(value))
File "/usr/local/neurosoft/miniconda3/lib/python3.6/site-packages/traits/trait_handlers.py", line 236, in error
object, name, self.full_info(object, name, value), value
traits.trait_errors.TraitError: Each element of the 'current_sphere' trait of a DynamicTraitedSpec instance must be a pathlike object or string representing an existing file, but a value of '[]' <class 'str'> was specified.

I guess this problem is due to configurations of fsLR template in my servers. I have set TEMPLATEFLOW_HOME in my servers but it looks useless.

I searched issues for that but haven't found any solutions, perhaps I missed something important.
Could someone give me some suggestions on this problem? Or if any lower versions of fmriprep could generate processed data in fsLR-32k space as cifti output, it would also be a great help to me.

Best,
Taicheng

[helloTC]

I submitted this question in Neurostar page, closed here.

[mgxd]

Hi @helloTC

It looks like one of the fsLR templates is not found - you may need to clear your templateflow cache (~/.cache/templateflow) and then reinstall templateflow. Then you could ensure the templates are present by running this line in your terminal prior to running fmriprep.

python -c "import templateflow.api as tf; tf.get('fsLR')"

If possible, I'd recommend using a container service such as Docker or Singularity, which allows for running versions in a fully encapsulated environment, also avoiding the complex dependency graph that fmriprep has.

[mgxd]

Version 20.0.3 is out with corrected CIFTI outputs, it's highly recommended that you regenerate your files with that.

Feel free to reopen if you have any issues.