The foll. should work for testing instead of building docker from scratch docker pull spanta28/coinstac_vbm_docker
This computation runs Voxel Based Morphometry on structural T1 weighted MRI scans(BIDS format) using SPMv12 standalone and MATLAB Runtimev713.
The computation runs serially and each scan takes approximately 10 mins to run on a system with 2.3 GHz,i5 equivalent processor and 8GB RAM. Each scan output directory takes about 150MB space after running this algorithm. Please make sure to have the space and resources to run this algorithm accordingly. Please read the contents of outputs_description.txt from the output directory to learn about the descriptions of output filed produced.
Test bids data is placed in /computation/test_dir/bids_input_data
At the end of the computation , the output is written to stdout in json with the property "success" set to true/false indicating whether the algorithm ran successfully or failed
Dockerfile_build_from_scratch builds everything from scratch, downloads packages from online Dockerfile sources from trendscenter/coinstac_nipype_spm_mcr_afni_centos7 and just uploads the current scripts into the docker.
For vbm+regression, use test/inputspec_vbm_regression.json as inputspec.json
if you use sample data in test directory, example output below: { "data": [ [ "regression_input_files/subID-1_swc1_4mm.nii", "regression_input_files/subID-2_swc1_4mm.nii" ], [ "niftifile" ] ], "covariates": [ [ [ [ "niftifile", "isControl", "age", "sex" ], [ "regression_input_files/subID-1_swc1_4mm.nii", true, 28, "M" ], [ "regression_input_files/subID-2_swc1_4mm.nii", true, 39, "M" ] ] ], [ "isControl", "age", "sex" ], [ "boolean", "number", "string" ] ] }
This is how outputDirectory looks: cd test/output/ ls //* local0/simulatorRun/vbm_outputs.zip local0/simulatorRun/wc1Re.png
local0/simulatorRun/regression_input_files: subID-1_swc1_4mm.nii subID-2_swc1_4mm.nii
local0/simulatorRun/vbm_outputs: outputs_description.txt quality_control_readme.txt subID-1 subID-2 vbm_log.txt
vbm+regression optional params : options_smoothing,reorientation params, regression_resampling_voxel_size etc. regression_resampling_voxel_size can be varied to check the performance of coinstac in terms of how much data flow it can handle. Default value for this is 4mm. For more info. check compspec.json Use inputspec_vbm_regression_all_options.json as inputspec.json
For standalone vbm, use test/inputspec_vbm_standalone.json as inputspec.json
example outputDirectory below: cd test/output/ ls //* local0/simulatorRun/vbm_outputs.zip local0/simulatorRun/wc1Re.png
local0/simulatorRun/vbm_outputs: outputs_description.txt quality_control_readme.txt subID-1 subID-2 vbm_log.txt
For standalone vbm options, use test/inputspec_vbm_standalone_all_options.json as inputspec.json
The optional params info is detailed in vbm optional params description.xlsx Type of info. available is : Name Label Description Json data type Algorithm dev type for inner processing Default value Range (if applicable) Step increments pre=processing step