Resilient OpenNeuro staging
Large OpenNeuro MEG runs may contain a small number of corrupt or unreadable subject FIF files. The standard neureptrace-openneuro-meg stage command is intentionally strict and aborts on the first failed subject. For exploratory full-cohort screening, use the resilient wrapper instead.
Installed environments expose the wrapper as both neureptrace-openneuro-resilient and the grouped neureptrace openneuro-resilient command. The equivalent python -m neureptrace.openneuro_resilient form remains useful from a source checkout.
neureptrace-openneuro-resilient stage \
--dataset ds006629 \
--bids-root "$OPENNEURO_DATA_ROOT" \
--staged-dir "$NEUREPTRACE_OPENNEURO_STAGED_DIR" \
--subjects all \
--runs all \
--summary-out outputs/openneuro_ds006629_full/stage_summary.csv \
--failure-summary-out outputs/openneuro_ds006629_full/stage_failures.csv \
--successful-subjects-out outputs/openneuro_ds006629_full/successful_subjects.txt \
--skip-failed-subjects \
--min-successful-subjects 2 \
--on-mismatch warn \
--overwrite
The same invocation can be routed through the grouped CLI when a workflow uses one command namespace:
neureptrace openneuro-resilient stage \
--dataset ds006629 \
--bids-root "$OPENNEURO_DATA_ROOT" \
--staged-dir "$NEUREPTRACE_OPENNEURO_STAGED_DIR" \
--subjects all \
--runs all \
--summary-out outputs/openneuro_ds006629_full/stage_summary.csv \
--failure-summary-out outputs/openneuro_ds006629_full/stage_failures.csv \
--successful-subjects-out outputs/openneuro_ds006629_full/successful_subjects.txt \
--skip-failed-subjects \
--min-successful-subjects 2 \
--on-mismatch warn \
--overwrite
The wrapper keeps the same staging options as the strict command, but records failed subjects in stage_failures.csv and writes the successful cohort as a compact participants.ids override. For example, if sub-06 fails in ds006629, successful_subjects.txt contains the remaining numeric participant ids so the decode config can be run only on the subjects that were staged successfully.
In GitHub Actions, pass --github-env "$GITHUB_ENV" to append updated OPENNEURO_SUBJECTS, OPENNEURO_N_SPLITS, and OPENNEURO_FAILED_SUBJECTS values for subsequent decode steps.