Modifications to the original Docker image for bids_HCPPipelines
1.7K
CBI's modifications to the BIDS-App version of the HCP Pipelines
It includes a python file used to select the images that are used for Anatomical Pipeline (so that we don't use both normalized and un-normalized versions of the same run). If you want, you can specify other ways of filtering the different files in that file.
It also has its own Dockerfile to copy the python file plus the scanner gradient coefficients file1 into the Docker image, and it also modifies the run.py file to include a call to the filtering of the list of anatomical images.
Finally, it includes a bash script (correct_qa_txt_files.sh) that loops through all the qa.txt files generated by the Docker machine and modifies them so that the paths included in the files are not for the docker container file system, but relative to the subject folder. So if you move the pre-processed data to a different folder/machine, you just need to cd to the subject folder level and just copy and paste the commands in the qa.txt files.
Clone or copy this repository to your local machine, cd to that folder, and run:
docker build -t <name-of-your-docker-image> .
(Note the dot (.) at the end of the command.)
To run the pipelines on your BIDS-organized data, for a single subject:
dataDir=<path-to-your-BIDS-organized-study>
outDir=<where-you-want-your-pre-processed-images>
subjectID=01
docker run --rm -v ${dataDir}:/data:ro \
-v ${outDir}:/out \
<name-of-your-docker-image> \
/data \
/out \
participant \
--participant_label ${subjectID}
--license_key LICENSE_KEY
(For more information on the options for the pipelines, see the BIDS-App version of the HCP Pipelines).
Now, to correct the paths in the qa.txt files generated by the pipelines, run:
. <path-to-your-local-copy-of-the-repository>/correct_qa_txt_files.sh \
${outDir}/sub-${subjectID}
(Note the dot (.) and space at the beginning of the command.)
1 The gradient coefficients file is not included in this distribution, since it is Siemens proprietary information.
Content type
Image
Digest
Size
3.4 GB
Last updated
almost 6 years ago
docker pull cbinyu/bids_hcppipelines