Run BIDS Apps over many studies on an HPC cluster, with BABS running the jobs.
mechababs works on a BIDS study, adding derivatives to it, and records how each one was made inside the study itself: the exact mechababs and babs that ran, pinned in a lock, and every orchestration step as a datalad run record.
A campaign is one such run's recipe, your app and cluster configs plus that locked environment, and mechababs iterate moves the study toward what you declared, one pass at a time.
A superstudy groups many studies so one campaign sweeps them all.
datalad clone https://github.com/OpenNeuroStudies/study-ds000001 && cd study-ds000001
uvx --from git+https://github.com/con/mechababs@main mechababs campaign init demo \
--apps ~/config/SimBIDS-0.0.3.yaml --cluster ~/config/your-site.yaml \
--babs https://github.com/PennLINC/babs.git@main --limit 2
source .mechababs/campaigns/demo/env.sh
mechababs add-dataset --sourcedata sourcedata/ds000001
mechababs iterate # repeat until every cell is mergedThe quickstart walks through it, configs included.
New cluster? Put the prerequisites in place (installation.md), then validate your cluster config with mechababs test-cluster (tutorial).
- docs/quickstart.md — one study through one campaign, end to end.
- docs/overview.md — the concepts: the study as the unit, the campaign as a recipe recorded in it, the reconciler.
- docs/glossary.md — the words, defined once.
- docs/reference.md — every command and flag, selection, and the two config files.
- docs/interventions.md — when a cell fails: find it, repair it in place, or redo it.
- docs/installation.md — HPC prerequisites: PATH tools, scratch, the container dataset.
- docs/cluster-config-and-testing-tutorial.md — add your cluster and validate it.
- docs/output_structure.md — the on-disk shape of a study, a superstudy, a campaign, and a derivative.
- docs/spec.md — the design decisions of record, and docs/use_cases.md, the user stories they answer to.
- CONTRIBUTORS.md — developing and testing mechababs itself.
- CLAUDE.md — contributor conventions and issue tracking.
- Open work and milestones live in the GitHub tracker.
- OpenNeuroStudies — the superstudy mechababs feeds
- OpenNeuroDerivatives — derivative mirrors + the fmriprep opinions repo
- BABS — the execution engine
- ReproNim/containers — container datasets
- FAIRly Big processing workflow — the pattern BABS implements
- STAMPED principles — the guiding principles