Closed mih closed 5 years ago
thanks @mih - perhaps we can create a setup script that people can run in the morning, to update any necessary pieces (e.g., broken image). this would be as opposed to downloading the full image.
I have now published the two dataset that are produced by this analysis on GitHub (with all data). You can find them here:
As the analysis capture the entire trace, it is sufficient to install the latter and you have access to all pieces through this one.
Little test whether our analysis setup can reproduce bit identical results:
% datalad install https://github.com/myyoda/demo-dataset-glmanalysis
% cd demo-dataset-glmanalysis
% datalad rerun --since= --onto 78d6b5511a9981c1da37de340badc4ead7bee0ad -b checkme
% git diff master --stat
sub-02/1stlvl_design.fsf | 2 +-
sub-02/1stlvl_glm.feat/design.fsf | 2 +-
sub-02/1stlvl_glm.feat/logs/feat0 | 2 +-
sub-02/1stlvl_glm.feat/logs/{feat0_init.e8712 => feat0_init.e30675} | 0
sub-02/1stlvl_glm.feat/logs/{feat0_init.o8712 => feat0_init.o30675} | 0
sub-02/1stlvl_glm.feat/logs/feat1 | 2 +-
sub-02/1stlvl_glm.feat/logs/feat1a_init | 2 +-
sub-02/1stlvl_glm.feat/logs/{feat2_pre.e8813 => feat2_pre.e30756} | 0
sub-02/1stlvl_glm.feat/logs/{feat2_pre.o8813 => feat2_pre.o30756} | 0
sub-02/1stlvl_glm.feat/logs/{feat3_film.e9360 => feat3_film.e31256} | 0
sub-02/1stlvl_glm.feat/logs/{feat3_film.o9360 => feat3_film.o31256} | 0
sub-02/1stlvl_glm.feat/logs/feat3_stats | 2 +-
sub-02/1stlvl_glm.feat/logs/{feat4_post.e9806 => feat4_post.e31630} | 0
sub-02/1stlvl_glm.feat/logs/{feat4_post.o9806 => feat4_post.o31630} | 0
sub-02/1stlvl_glm.feat/logs/{feat5_stop.e10442 => feat5_stop.e32267} | 0
sub-02/1stlvl_glm.feat/logs/{feat5_stop.o10442 => feat5_stop.o32267} | 0
sub-02/1stlvl_glm.feat/report.html | 2 +-
sub-02/1stlvl_glm.feat/report_log.html | 2 +-
18 files changed, 8 insertions(+), 8 deletions(-)
Just the logs have changed (time stamps, paths and such). Hooray reproducibility.
I guess I will merge it so we have all materials for upcoming sfn2018- clone
I made a new variant of the prep-script. Using VM path for images (avoids some downloads) and disables all the stuff meant for interactive demos. What follows is the code I ran inside the VM after importing the pristine OVA (
e789b13556ca0f517743e6a1da4f3fe0 /tmp/reprotraining.ova
):All executed in $HOME inside the image
Notice that I could not use the FSL singularity image in the VM, but had to use shub://ReproNim/ohbm2018-training:fsln (the local one has no proper PATH environment setup).
Here is the full log of running the script: