datalad-container is a DataLad extension that integrates containerized execution into DataLad workflows. Its purpose:
- Let you register, version, and manage container images (Docker, Singularity/Apptainer, etc.) alongside dataset content.
- Enable transparent execution of commands inside containers via DataLad’s run/rerun commands so processing is recorded with provenance (which container, inputs, outputs, command).
- Provide commands to wrap container execution, discover available containers, and record container metadata in the dataset.
- Improve computational reproducibility by coupling analysis code, data, and the exact container runtime used to run analyses.
Installable as a separate Python package and used to make DataLad-managed analyses reproducible and portable.