{{repo-name}}

This is a paper-template repository. Replace every {{placeholder}} with the values for your new paper, then git init and start writing. See # What to fill in below for the full checklist.

Manuscript: {{paper-title}}. {{authors}}, CESI LINEACT, {{year}}. In preparation for submission to {{journal}}.

Headline result (replace before going public)

Statistic	Value
Primary effect-size	{{X.XX}} (CI [{{LO}}, {{HI}}])
Significance	exact p = {{0.0XX}} ({{n}}! permutation)
Cross-validation protocol	leave-{{group}}-out, K-fold
Sample size	{{N}} observations across {{m}} groups

What's in here

{{repo-name}}/
├── paper.tex                     # Main manuscript (iopjournal class)
├── paper_si.tex                  # Supplementary Information
├── iopjournal.cls                # IOP journal class (do not edit, do not store in subfolders for submission)
├── orcid.pdf                     # ORCID icon embedded by iopjournal's \orcid{} macro
├── cover_letter.md               # Cover letter draft
├── .zenodo.json                  # Zenodo deposit metadata
├── CITATION.cff                  # Citation File Format
├── references/references.bib     # BibTeX
├── experiments/                  # d01 ... dNN numbered, reproducible
│   ├── _plot_style.py              # Publication-quality plot helper (Paul Tol palette)
│   ├── d01_pilot.py                # Smoke-test pilot
│   └── README.md                   # Script-numbering convention
├── figures/                      # All publication figures (PDF)
├── outputs/                      # Per-experiment JSON / CSV / NPZ
├── drafts/                       # Earlier drafts, design notes (gitignored content welcome)
└── README.md                     # This file

Reproducing the paper

# Build the manuscript (figures must be in figures/, iopjournal.cls at repo root)
pdflatex paper.tex
bibtex   paper
pdflatex paper.tex
pdflatex paper.tex

# Re-run a numbered experiment (example)
python3.12 experiments/d01_pilot.py

Master random seed is pinned to SEED = 42 in every stochastic script. Heavy caches (precomputed features, eigenvector dumps) are NOT committed; they are deterministically regenerated by the corresponding script on first run.

Zenodo

A versioned Zenodo DOI is auto-minted on every GitHub Release via the GitHub-Zenodo bridge. The deposit metadata lives in .zenodo.json. To activate on a fresh repo :

1. Sign in to https://zenodo.org with the GitHub account that owns this repository.
2. Enable the repository under Settings → GitHub integration.
3. Push a release tag (git tag v0.1 && git push --tags); Zenodo will mint the DOI on the first GitHub Release.
4. Substitute the issued concept DOI into the \data{...} block of paper.tex and into the doi: field of CITATION.cff.

How to cite

A machine-readable citation is provided in CITATION.cff. Plain BibTeX :

@unpublished{{{cite-key}},
  author = {{{authors-bibtex}}},
  title  = {{{paper-title}}},
  note   = {Manuscript in preparation, CESI LINEACT, {{year}}.
            \url{https://github.com/cycling-data-lab/{{repo-name}}}},
  year   = {{{year}}}
}

Sibling repos

• materials-applicability-bound — structural lower bound on the applicability-domain gap (MLST submission), the methodological foundation of the GSP framework used throughout the org.
• bikeshare-demand-forecasting — empirical anchor on bike-share demand prediction.
• bikeshare-gsp-tools — graph-signal-processing toolkit (spectral bounds, D-optimal siting).

License

MIT.

Contact

{{lead-author-name}} — {{lead-author-email}} — ORCID {{co-author-name}} — ORCID

---

What to fill in (delete this section before going public)

A new paper repo created from this template needs the following placeholders replaced :

Placeholder	Where	Example value
{{repo-name}}	README, .zenodo.json, CITATION.cff	materials-applicability-bound
{{paper-title}}	README, paper.tex, .zenodo.json, CITATION.cff, cover_letter.md	A structural lower bound on …
{{authors}} / {{authors-bibtex}}	README, paper.tex, CITATION.cff	R. Foss\'e and G. Pallares
{{year}}	README, paper.tex, .zenodo.json, CITATION.cff	2026
{{journal}}	README, paper.tex, cover_letter.md	Machine Learning: Science and Technology
{{cite-key}}	README, BibTeX	FossePallares2026ApplicabilityBound
{{lead-author-*}} / {{co-author-*}}	README	name / email / ORCID URL
Headline result table	README §Headline	actual numbers when known
paper.tex title / author / abstract / keywords / data block	paper.tex	fill from your own paper
paper.tex \graphicspath{}	paper.tex	usually {{figures/}}
paper_si.tex same fields	paper_si.tex	same as paper.tex
cover_letter.md placeholders	cover_letter.md	journal name, suggested reviewers
.zenodo.json author block, title, keywords	.zenodo.json	fill from .your own paper
CITATION.cff author block, title, journal	CITATION.cff	same
requirements.txt	requirements.txt	pin the packages you actually use
references/references.bib	references.bib	start adding entries as you cite them
experiments/d01_pilot.py	experiments/	rename / rewrite for your first experiment

After the substitutions, run :

grep -rnE "\{\{[a-z_-]+\}\}" .   # this should return zero matches before you push v0.1

Then git init, push to GitHub under cycling-data-lab/, enable Zenodo, and start writing.