HOWTO.md

YODA HOWTOs

This document describes most commonly used commands and workflows to operate on the hierarchy

Find template text still needed to be changed

git grep YODATODO

Add an additional existing input dataset

datalad add -d . -s URL inputs/NAME

e.g.,

datalad add -d . -s ///openfmri/ds000001 inputs/ds000001

Update input dataset(s) to a new version

E.g.

datalad update --merge inputs/ds000001/
datalad save -d . -m "fetched more subjects" inputs/ds000001

Create a new sub-dataset to contain results of preprocessing

datalad create -d . outputs/NAME

e.g.

datalad create -d . outputs/fmriprep

Run a command in a "reproducible" way

Use

datalad run CMD [OPTIONS]

which will create a new commit recording all the changes done by the CMD, so it could later be reran using datalad rerun command. E.g.

datalad run -m "preprocess sample subject" \
  envs/fmriprep inputs/ds000001 outputs/fmriprep/sub-01

Reproduce a full sequence of commands (and commits) ever ran before

datalad rerun --onto= --since= -b rerun-1

which will place reran results into rerun-1 branch for comparison etc

Publish entire hierarchy to HTTP server via SSH

TODO

Publish entire hierarchy to github