cheeseboard

Daily scrape of the Cheese Board Collective pizza schedule, for fun future stats (ingredient frequency, favorite combos, etc). Salads are ignored on purpose.

A GitHub Action runs scrape_cheeseboard.py once a day. Each pizza is written to data/<pizza-date>.json, keyed by the pizza's own date and upserted — since the site usually lists the whole upcoming week at once, most days get scraped more than once before their date arrives, and the file for that date is simply overwritten each time. The same run also rewrites the table below from data/, showing this week (Monday onwards, including days already gone by) plus anything further ahead the site has posted. Rows only drop off when a new week starts, so the whole week stays readable rather than shrinking a day at a time — the full history lives in data/ regardless.

This week's pizzas

Date	Day	Pizza
2026-11-09	Mon	Closed
2026-11-10	Tue	Organic tomato, red onion, mozzarella, toasted garlic Parmesan, garlic olive oil, oregano, parsley
2026-11-11	Wed	Cremini mushroom, green bell pepper, yellow onion, Valbreso feta, mozzarella, garlic olive oil, parsley
2026-11-12	Thu	Organic zucchini, red onion, herbed fresh ricotta, mozzarella, garlic olive oil, parsley
2026-11-13	Fri	Organic baby chard and kale, leek, mozzarella, Caciocavallo cheese, garlic olive oil, organic orange and lemon zest, parsley
2026-11-14	Sat	House made tomato sauce, fresh mozzarella made in Berkeley by Belfiore, mozzarella, garlic olive oil, basil

Files

• scrape_cheeseboard.py — fetches the schedule page and parses each day's pizza (skips salads).

• scrape-pizza.yml — the GitHub Actions workflow (goes in .github/workflows/).

• tests/ — unit tests against saved HTML fixtures of the real page.

• data/ — one JSON file per pizza date, either a normal day:

  {
    "date": "2026-09-12",
    "weekday": "Sat",
    "closed": false,
    "ingredients_raw": "Organic corn (Avalos Farm), red onion, mozzarella, ...",
    "ingredients": ["Organic corn", "red onion", "mozzarella", "..."],
    "ingredients_normalized": ["organic corn", "red onion", "mozzarella", "..."]
  }
  

  or a closed day (weekly closure, holiday, or next week's schedule not posted yet — the site marks all of these the same way, with no "Pizza" heading and a "The pizzeria is closed today." paragraph):

  { "date": "2026-09-14", "weekday": "Mon", "closed": true }
  

  The scraper only fails (non-zero exit) when it can't find any recognizable day on the page at all — a real sign the site's structure changed, not that the pizzeria happened to be closed.

• ingredients.toml — hand-maintained ingredient taxonomy: maps each scraped ingredients_normalized string to a canonical ingredient and a category (cheese, produce, sauce, ...), so "organic tomato" and "organic heirloom tomato" count as one Tomato. A scrape that brings in an unknown ingredient lists it in the PR body; add it here on the PR branch before merging.

• stats.py + build_site.py + templates/ — build the static stats site (today's pizza with "last seen" context, closest earlier pizza, this week, per-category leaderboards, rarities, most similar pizzas, ingredient-by-week table, and a page per day and per ingredient).

• build-site.yml — builds the site and deploys it to GitHub Pages on every push to main that touches the data or the site, plus once a day so "today" keeps moving.

Local dev

Dependencies and script execution are managed with uv.

uv sync
uv run pytest
uv run scrape_cheeseboard.py --print
uv run scrape_cheeseboard.py --readme README.md  # also refresh the table above
uv run build_site.py                             # build the site into _site/
uv run build_site.py --today 2026-10-15          # ... as if it were another day
uv run build_site.py --list-unmapped             # ingredients missing from ingredients.toml