Physics Reproducibility Archive

Contract

• Input: physics result, data, and code.
• Output: Zenodo DOI + GitHub release + checklist.
• Side effects: none.
• Dependencies: none.
• Stop condition: DOI assigned; tag created; checklist complete.
• Risk: low.
• Boundary: archives and links; does not execute long computations.

Physics Reproducibility and Archival

Make a physics result reproducible and archive it with a DOI — data, code, and environment — for compliance with journals and funders.

When to use

• A paper is being submitted to a journal that requires data availability.
• A funder (DOE, NSF, ERC) requires open data / code.
• A reviewer or collaborator needs to reproduce a result.

Process

1. Inventory the artifacts

List:

• Raw data — unprocessed files from the experiment / simulation.
• Processed data — calibrated and cleaned.
• Analysis code — scripts that produce figures and tables.
• Simulation code — input parameters, source, build instructions.
• Environment — library versions, OS, compilers.
• Figures and tables — with captions that cite the source data.

Completion criterion: inventory saved as REPRODUCIBILITY.md.

2. Prepare for archival

• Data: clean, documented schema; remove PII; check file formats (open preferred: CSV, HDF5, FITS; avoid proprietary).
• Code: add README, setup.py / pyproject.toml, test that it runs.
• Documentation: usage instructions, expected output, hardware requirements.
• Licence: MIT, Apache 2.0, or CC0 for data; GPL for code (choose before upload).

Completion criterion: inventory items prepared; licence chosen.

3. Containerise

• Docker: Dockerfile with base image, tool versions, code, data.
• Apptainer: .def file for HPC clusters.
• Include: exact tool versions (not latest); build instructions; run command.

Completion criterion: container builds; produces expected output on a fresh host.

4. Archive on Zenodo

Create a Zenodo deposit:

• Title: descriptive and unique (not just "Data for paper").
• Authors: all contributors, with ORCIDs.
• Keywords: physics terms (e.g. "quantum optics", "cosmology", "plasma physics").
• Licence: stated.
• Related identifiers: link to the paper (arXiv DOI or journal DOI).
• Upload: data + code + container (or link to Docker Hub / GHCR).

Completion criterion: Zenodo record created; DOI assigned.

5. GitHub release

• Tag the repository with the paper version (e.g. v1.0).
• Write a release note: what changed, what is archived.
• Link to the Zenodo DOI in the release.

Completion criterion: GitHub release created; DOI linked.

6. Checklist and compliance

• Journal requirements — which journals require data / code? (Nature, Science, APS, IOP all have policies).
• Funder requirements — NSF, DOE, ERC open-data mandates.
• FAIR compliance — check Findable, Accessible, Interoperable, Reusable.
• Supplemental material — prepare for journal submission (separate from archival).

Completion criterion: checklist saved; compliance status per requirement.

Notes

• Pair with physics-simulation-setup for simulation containerization.
• Pair with physics-experimental-notebook for experimental data archival.
• Pair with physics-writing-revtex for journal submission.