MIndboggle (latest) crashing on processing freesurfer 6 data

[cecs1]

Hello,
I'm trying to process some data through mindboggle on data that was ran successfully through Freesurfer 6. It doesn't matter which scan I use, the same error occurs.

I'm getting errors below (there are many after this of similar nature) - they appear to relate to the Surface_to_vtk routine.

Any advice appreciated.
Thanks,
Chris

191202-20:16:59,146 nipype.workflow INFO:
Workflow Mindboggle settings: ['check', 'execution', 'logging', 'monitoring']
191202-20:16:59,197 nipype.workflow INFO:
Running serially.
191202-20:16:59,198 nipype.workflow INFO:
[Node] Setting-up "Mindboggle.Freesurfer_annot" in "/home/jovyan/work/mindboggled/working/0315/Mindboggle/_hemi_rh/Free
surfer_annot".
191202-20:16:59,201 nipype.workflow INFO:
[Node] Running "Freesurfer_annot" ("nipype.interfaces.io.DataGrabber")
191202-20:16:59,205 nipype.workflow INFO:
[Node] Finished "Mindboggle.Freesurfer_annot".
191202-20:16:59,206 nipype.workflow INFO:
[Node] Setting-up "Mindboggle.Surfaces" in "/home/jovyan/work/mindboggled/working/0315/Mindboggle/_hemi_rh/Surfaces".
191202-20:16:59,209 nipype.workflow INFO:
[Node] Running "Surfaces" ("nipype.interfaces.io.DataGrabber")
191202-20:16:59,211 nipype.workflow INFO:
[Node] Finished "Mindboggle.Surfaces".
191202-20:16:59,212 nipype.workflow INFO:
[Node] Setting-up "Mindboggle.Surface_to_vtk" in "/home/jovyan/work/mindboggled/working/0315/Mindboggle/_hemi_rh/Surfac
e_to_vtk".
191202-20:16:59,274 nipype.workflow INFO:
[Node] Running "Surface_to_vtk" ("nipype.interfaces.utility.wrappers.Function")
191202-20:16:59,656 nipype.workflow WARNING:
Storing result file without outputs
191202-20:16:59,656 nipype.workflow WARNING:
[Node] Error on "Mindboggle.Surface_to_vtk" (/home/jovyan/work/mindboggled/working/0315/Mindboggle/hemi_rh/Surface_to
vtk)
191202-20:16:59,659 nipype.workflow ERROR:
Node Surface_to_vtk.a1 failed to run on host a460efb46421.
191202-20:16:59,660 nipype.workflow ERROR:
Saving crash info to /root/crash-20191202-201659-root-Surface_to_vtk.a1-12bb0046-09dc-4661-bf46-c45e734ce5aa.txt
Traceback (most recent call last):
File "/opt/miniconda-latest/envs/mb/lib/python3.6/site-packages/nipype/pipeline/plugins/linear.py", line 48, in run
node.run(updatehash=updatehash)
File "/opt/miniconda-latest/envs/mb/lib/python3.6/site-packages/nipype/pipeline/engine/nodes.py", line 479, in run
result = self._run_interface(execute=True)
File "/opt/miniconda-latest/envs/mb/lib/python3.6/site-packages/nipype/pipeline/engine/nodes.py", line 569, in _run_interface
return self._run_command(execute)
File "/opt/miniconda-latest/envs/mb/lib/python3.6/site-packages/nipype/pipeline/engine/nodes.py", line 662, in _run_command
result = self._interface.run(cwd=outdir)
File "/opt/miniconda-latest/envs/mb/lib/python3.6/site-packages/nipype/interfaces/base/core.py", line 382, in run
runtime = self._run_interface(runtime)
File "/opt/miniconda-latest/envs/mb/lib/python3.6/site-packages/nipype/interfaces/utility/wrappers.py", line 144, in _run_inte
rface
out = function_handle(**args)
File "", line 54, in freesurfer_surface_to_vtk
File "/opt/miniconda-latest/envs/mb/lib/python3.6/site-packages/nibabel/freesurfer/io.py", line 179, in read_geometry
coords = np.fromfile(fobj, ">f4", vnum * 3).reshape(vnum, 3)
ValueError: cannot reshape array of size 1287294 into shape (389005987,3)

[PeerHerholz]

Hi @cecs1,

thank you very much for posting this problem.

Before we go any further: are you using Mindboggle within a container (docker/singularity) or through the bare-metal installation?
That function wasn't changed in a while I think, hence I'm wondering where this is coming from...

Could you maybe also post an example of how you run freesurfer and mindboggle?

Best, Peer

[cecs1]

Thanks Peer. I run the docker container version docker run --rm -ti -v $HOST:$DOCK -p 5000:5000 nipy/mindboggle Then, I set variables: export HOST=/home/csteward export DOCK=/home/jovyan/work export ID=0315 export FREESURFER_OUT=$DOCK/freesurfer_subjects export FREESURFER_SUBJECT=$FREESURFER_OUT/$ID export OUT=$DOCK/mindboggled Then I run Mindboggle on the freesurfer 6 output (which was run separately outside the docker container) mindboggle $FREESURFER_SUBJECT --out $OUT I can tell you if I run everything in the docker container ie freesurfer and mindboggle, there are no problems. I can also tell you that the output from running freesurfer 6 separately compared to the freesurfer 6 output that mindboggle generates are different ie at least in terms of number of files generated eg the surf directory has 80 folders using mindboggle freesurfer, but 96 using external freesurfer. I run freesurfer with recon-all and no special options - I assume mindboggle is doing the same? regards, Chris ************************************************** Dr Chris Steward Dept of Medicine and Radiology, Imaging Manager, MBC@RMH University of Melbourne/Royal Melbourne Hospital Grattan St, Parkville, 3050 Tel: (03) 903 56508 E-mail: [email protected] ****************************************************

…

________________________________ From: Peer Herholz <[email protected]> Sent: 04 December 2019 01:25 To: nipy/mindboggle <[email protected]> Cc: Chris Steward <[email protected]>; Mention <[email protected]> Subject: Re: [nipy/mindboggle] MIndboggle (latest) crashing on processing freesurfer 6 data (#196) Hi @cecs1<https://github.com/cecs1>, thank you very much for posting this problem. Before we go any further: are you using Mindboggle within a container (docker/singularity) or through the bare-metal installation? That function wasn't changed in a while I think, hence I'm wondering where this is coming from... Could you maybe also post an example of how you run freesurfer and mindboggle? Best, Peer — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#196?email_source=notifications&email_token=AL246FOHUGDMHOVT2HPBMZDQWZT6PA5CNFSM4JUI6C32YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEFZRNAI#issuecomment-561190529>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AL246FITLDM3XBOKS6J2ISDQWZT6PANCNFSM4JUI6C3Q>.

[PeerHerholz]

Hi @cecs1,

thank you very much for the additional information.

Hm, interesting. I never saw that/heard of it before. How about you @binarybottle, @satra?

@cecs1 did you already try an older version of the mindboggle container ?
Regarding freesurfer: mindboggle installs it through neurodocker which grabs a min version as you can see e.g., here. Unfortunately, I can't remember the difference to the classic version right now. AFAIK, it excludes some example data, etc. which should not affect the number of output folders. @kaczmarj could you maybe provide any pointers here (sorry for always tagging you!)?

[binarybottle]

I am afraid I haven't run into this problem before.
Based on the error, I would have guessed that there is an issue with how nibabel is dealing with the FS surface data, but it doesn't make sense to me that the FS data would be different when FS is run inside vs. outside the container.

[PeerHerholz]

The inside vs. outside container FS data thingy indeed seems weird...

@cecs1 could you maybe try the following for both, the recon-all output from freesurfer within and outside a container:

import nibabel as nb
surf = nb.freesurfer.read_geometry(surface_file)

where surface_file is a freesurfer surface file.

[kaczmarj]

@PeerHerholz you can inspect the contents of https://dl.dropbox.com/s/nnzcfttc41qvt31/recon-all-freesurfer6-3.min.tgz. it was made using the instructions in freesurfer/freesurfer#70 (comment). the binaries are exactly the same as in freesurfer version 6.0.0, there are just fewer of them (the package was trimmed to only include things necessary for recon-all).

[PeerHerholz]

Thx @kaczmarj for the information and pointers! @cecs1 do you perhaps have any updates?