Error in affine transform during afni_proc.py

AFNI Message Board
1

I created a proc.subj script using afni_proc.py and did not see any errors during the creation of the script. However, when trying to run the script, I run into problems during the transformation. It seems like it is not recognizing base.nii but I cannot understand why not. Does the following give you any ideas why not?

++ 3dUnifize: AFNI version=AFNI_16.2.16 (Sep 4 2016) [64-bit]

  • Pre-processing: ADV…UWG
    ++ Output dataset ./anat.un.nii.gz
    ++ ===== Elapsed = 88.9 sec
    #Script is running (command trimmed):
    3dcopy /abin/MNI_avg152T1+tlrc ./base.nii
    ++ WARNING: file /home/jgowin/.afni.log is now 550997561 (551 million [mega]) bytes long!
    ++ 3dcopy: AFNI version=AFNI_16.2.16 (Sep 4 2016) [64-bit]
    #++ Aligning /mnt/nfs2/studies/relapsefmri/JoshScripts/A0322201401.results/awpy/base.nii data to /mnt/nfs2/studies/relapsefmri/JoshScripts/A0322201401.results/awpy/anat.un.nii data
    #Script is running (command trimmed):
    @auto_tlrc -base ./base.nii -input ./anat.un.nii -suffix .aff -no_ss -no_pre -init_xform CENTER
    Error: template (or -apar) dset not found: ./base.nii
    #**ERROR Failed in affine step
    ** ERROR - script failed
    3dbucket: No match.
2

From the results directory, what is the output of:

ls -l awpy/base.*

Maybe the .gz file is there, instead. Are you
using AFNI_AUTOGZIP?

  • rick
3

Oh wait, it is AFNI_COMPRESSOR that would do this
in the case of NIFTI. Do you have that set?

  • rick
4

Yes, it looks like the .gz file is there:

[jgowin@muso A0322201401.results]$ ls -l awpy/base.*
-rw-rw----+ 1 jgowin paulus-lab 315649 Aug 23 18:07 awpy/base.nii.gz

How do I disable the compressor, or work around it?

5

I have tried including “unsetenv AFNI_COMPRESSOR” at the start of the proc.SUBJ script but it hasn’t prevented the files to be compressed as .gz. Is there another way to unset the AFNI_COMPRESSOR variable?

Thanks!

6

What shell are you using? echo $0

Check your ~/.cshrc file, and if you have one, ~/.afnirc.
If you are using bash, check those start-up files.

Note that even if you unset that variable in the script,
the script will start sub-shells. So if those apply it, or
if it is in your .afnirc file, it will still be used.

  • rick
7

The default shell is bash but I run the proc.SUBJ script using the suggested tcsh proc.SUBJ -xef
I may have to contact the root user to change the local environment. Thanks for helping me get to the bottom of this.

8

It shouldn’t be so much much. What does this show?

grep AFNI_COMPRESSOR ~/.afnirc ~/.cshrc ~/.bash*

  • rick
9

Here’s the output.

[jgowin@muso ~]$ grep AFNI_COMPRESSOR ~/.afnirc ~/.cshrc ~/.bash*
/home/jgowin/.afnirc:unset AFNI_COMPRESSOR
grep: /home/jgowin/.cshrc: No such file or directory
/home/jgowin/.bash_history:AFNI_COMPRESSOR
/home/jgowin/.bash_history:unset AFNI_COMPRESSOR
/home/jgowin/.bash_history:unsetenv AFNI_COMPRESSOR
/home/jgowin/.bashrc:unset AFNI_COMPRESSOR
/home/jgowin/.bashrc~:unset AFNI_COMPRESSOR

Does any of that look problematic?

10

I updated the files and now here is what I get.

[jgowin@muso ~]$ grep AFNI_COMPRESSOR ~/.afnirc ~/.cshrc ~/.bash*
/home/jgowin/.afnirc:AFNI_COMPRESSOR = NONE
/home/jgowin/.cshrc:setenv AFNI_COMPRESSOR NONE
/home/jgowin/.bashrc:AFNI_COMPRESSOR=NONE
/home/jgowin/.bashrc:export AFNI_COMPRESSOR

Now, most of my files are saving uncompressed as .BRIK files. However, at the @auto_tlrc stage, it is still saving as .nii.gz and failing to find base.nii since only base.nii.gz exists.

Do I need to edit my rc files?

11

I just saw this in my output.proc file:

auto_warp.py -base MNI_avg152T1+tlrc -input 0322201401_ns+orig -skull_strip_input no
#++ auto_warp.py version: 0.03
– clearing AFNI_COMPRESSOR …

Do I need to prevent auto_warp.py from clearing my AFNI_COMPRESSOR setting?

12

Sorry, but I have been away a bit.

Are you at the NIH? If so, maybe I could look
more closely at your setup?

  • rick