will
60f3bcc4db
Fixed a real audit-methodology bug first: the bulk gap-audit was sampling each saint's own date via prayVespera, which reflects whatever saint claims First Vespers that evening (often a neighboring day's), not the sampled saint's own category. Redid it via prayLaudes (no anticipation). That surfaced the real finding: common-of-a-confessor (37 saints, this app's largest remaining gap) isn't a distinct liturgical Common -- it's common-of-a-confessor-not-bishop under a different label. St. Joseph's and St. Thomas Aquinas's chapter is byte-identical to the already- authored Sir 31:8-9 text, and several of the 37 saints' own file comments already say "vide C5" (Confessor Not a Bishop) directly. Not relabeled across the 37 saint files, since matins-hymn-common-of-a-confessor.yml and benedictus-antiphon-common-of-a-confessor.yml already establish this plain id as the real, working one for those saints' Matins/Benedictus content -- duplicated the Lauds/Vespers bundle under this id too, to match that existing convention. Updates two existing tests whose dates collided with this fix: the FERIAL_MONDAY constant (moved off St. Raymond Nonnatus's own day) and a new Vespers proof on St. Joseph's First Vespers. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014axryJBrRswYh2niA7WCUc
vu
A static, offline-first PWA for praying the Divine Office — Matins, Lauds, Prime, Terce, Sext, None, Vespers, and Compline — in bilingual Latin/English, for any given date.
This isn't a reconstruction of any single historical breviary. It's a personal design that
draws on several historical sources without reproducing any one of them exactly: hour
structure chiefly from the Monastic Tridentinum (1617), precedence/rubrics from Divino
Afflatu, and a calendar merged from both the Tridentine 1906 and Divino Afflatu calendars,
with more generous handling of commemorations and octaves than any one historical rite used.
See CLAUDE.md (gitignored, not shipped in this repo) or the project's own notes for the full
rationale if you're extending it.
There's no backend and no database — everything is a pre-built static bundle, and the app is designed to keep working fully offline once it's been loaded once.
Development
Requires Node.js (developed against Node 20+).
npm install
npm run dev # start the Vite dev server
npm run build # type-check (tsc --noEmit) and produce a production build in dist/
npm run preview # serve the production build locally
npm test # run the test suite (vitest)
Run a single test file or a subset by name:
npx vitest run tests/hours/vespers.test.ts
npx vitest run -t "some test name"
Running in production
The app is entirely static once built, so "running it" just means serving dist/ — any static
file host works, as long as it falls back to index.html for unknown paths (the app does its
own client-side routing for URLs like /2026-08-09/prime).
Docker
A Dockerfile builds the static bundle with Node and serves it with Caddy:
docker build -t vu .
docker run -p 8080:80 vu
Or with Docker Compose (docker-compose.yml is included):
docker compose up -d --build
This serves the app on http://localhost:8080. Put a reverse proxy/TLS terminator in front of
it for a real deployment; the container itself just listens on port 80 internally and has no
configuration, secrets, or external dependencies.
Without Docker
Any static file server works — build the app, then serve dist/ with SPA-style fallback
routing. For example, with Caddy directly:
npm run build
caddy file-server --root dist --listen :8080
(For a real deployment you'll want a fallback rule so unknown paths resolve to index.html —
see the Caddyfile in this repo for the pattern.)