mikebailey.github.io project instructions

Personal website / AcademicPages Jekyll site. Published at michaelbailey.org via GitHub Pages.

sci-map/ is a mirror, not the source

The sci-map/ directory is a copy of mikebailey/sci-world-map (extracted 2026-06-02; canonical repo was originally named mikebailey/sci-map but had to be renamed to avoid path-shadowing the Jekyll mirror at michaelbailey.org/sci-map/). The canonical version is deployed independently at sci-map.michaelbailey.org; this mirror exists so the page also renders inside the AcademicPages layout at michaelbailey.org/sci-map/.

To update the mirror after canonical changes:

# from this repo root, with ../sci-map/ on your local disk
bash sci-map/sync-from-canonical.sh
git diff -- sci-map
git add sci-map && git commit -m "sync sci-map/ from canonical"
git push

The sync script:

• rsyncs the canonical repo’s docs/ payload (its Pages publishing root since 2026-07; repo-level files like AGENTS.md, etl/, bin/ live above it, unpublished) into sci-map/, excluding index.html, .nojekyll, and CNAME.
• Re-stitches the Jekyll frontmatter (layout: null, permalink: /sci-map/, the Mike-branded title/description), the favicon link, the ← michaelbailey.org back-link, and the #site-back CSS that the canonical version drops.

Don’t hand-edit files under sci-map/ other than the back-link / Jekyll frontmatter — they’ll be overwritten next sync. Changes meant to flow to both sites belong in the canonical repo.

CV / resume: _data/cv.yml is the single source of truth

One YAML file drives three outputs:

To update the CV: edit _data/cv.yml, then regenerate both PDFs:

uv run scripts/cv_to_pdf.py
git add _data/cv.yml files/*.pdf && git commit -m "Update CV"

First run on a new machine needs uv sync && uv run playwright install chromium.

Notes:

• Work entries carry a resume: flag. true includes the entry in the short resume; false means full CV only. The resume also caps each role at two highlights and drops the Summary and Patents sections.
• The two PDFs are committed — GitHub Pages serves them as static downloads from the /cv/ page. They’re generated artifacts that still need to be in git.
• The print pages are sitemap: false and noindex; they exist only as PDF sources and aren’t linked from the nav.
• Structure follows the JSON Resume schema. Empty sections aren’t rendered.

Historical note: this replaced scripts/cv_markdown_to_json.py + update_cv_json.sh, which generated _data/cv.json by parsing a markdown CV in _pages/cv.md. That markdown source was later replaced by a template stub, so the scripts silently parsed an empty document — running them would have wiped the CV. Both were deleted along with cv.json.

Other things in this repo

(Add as needed — keeping this file minimal until there’s a load-bearing detail worth capturing.)