SPM12-based pipelines require a little extra work to get them compiled and working in a container. Freesurfer's Freeview is also included here, as it's very handy for creating the PDF QA report. This example shows three different ways of creating image displays for the QA PDF.
See https://github.com/baxpr/demo-singularity-matlab-fsl for a lot of detailed info about putting Matlab code into singularity containers. This example uses the same structure.
A licensed Matlab installation is required to compile the Matlab code, but is not needed to run the compiled executable in the container.
SPM12 (https://www.fil.ion.ucl.ac.uk/spm/software/spm12/) is not in this repository and must
be installed separately on the compilation host. Edit matlab/compile_matlab.sh to point
to it.
SPM requires jumping an extra hurdle at the compilation step - we use a modified version
of SPM's own compiler function spm_make_standalone.m, found at
matlab/spm_make_standalone_local.m. This process captures a lot of dependencies that
could otherwise easily be left out of the executable, with the resulting symptom that it
compiles just fine but fails at run time with various cryptic error messages. In addition
to SPM12, everything in the matlab/src directory is included in the path at compile time.
If Matlab toolboxes are used, they will need to be added to the list of included toolboxes
in matlab/spm_make_standalone_local.m.
The compiled Matlab executable is stored on github using git LFS. A regular git clone will download a pointer text file instead of the executable binary. The result of building a container from that will be a cryptic error message - so, compile it yourself. Or, if storing on github, download it manually and replace the pointer text file, or include this step in the Singularity file if helpful - example here: https://github.com/baxpr/gf-fmri/blob/47b0552/Singularity.v1.3.4#L65
Freesurfer requires a license to run: https://surfer.nmr.mgh.harvard.edu/fswiki/DownloadAndInstall#License Best practice is to store your license file on the host that will run the container, and bind it to the container at runtime - NOT to include your own license file in the container itself.