johanngb/cpt repository overview

Image to reproduce analyses. An example command to run would be:

docker run -it --rm --mount type=bind,source=<output directory>,destination=/Replication/output johanngb/cpt

with <output directory> being replaced by the directory where you want the output to be written. (Note that the directory path must be absolute, so if you want to output to the current directory, then source=`pwd` will work but source=. will not.) It will take some time to run. On a quad-core (8 thread) i7, runtime was a little less than 2 days. Code is parallelized, and uses all but 1 core (2 threads).

To extract the script files and datasets from the image and copy them to the current directory, type:

docker run --name cpt-temp johanngb/cpt /bin/true
docker cp cpt-temp:/Replication/ .
docker cp cpt-temp:/dockerfiles/cpt.paper_1.02.tar.gz .
docker rm cpt-temp

There will then be an R package cpt.paper_1.0.tar.gz which contains the datasets (and a few helper functions), and a directory "Replication" which contains all of the scripts to reproduce the analyses. The bash script script_run.sh will run all of the R scripts.