@meri-imperiumi/signalk-passage-briefing 0.8.1 → 0.9.0

package/CHANGELOG.md

@@ -2,6 +2,20 @@
 
 ## [Unreleased]
 
+## [0.9.0] - 2026-10-12
+
+### Added
+- The header always names when the served briefing was made: a quiet `Compiled MM-DD HH:MM` pill next to the connectivity and ship's-time pills, stamped from the payload's `fetchedAt` in ship's time on every load (route and here modes alike, any age — the stale-age banner stays the alarm past 24 h, this is the fact next to it). A briefing with no fetch time hides the pill rather than inventing a stamp; the label re-renders when the vessel's timezone arrives.
+- The header names when the next briefing is scheduled: a quiet `Next MM-DD HH:MM` pill beside the Compiled stamp, showing the state machine's next cron window in ship's time (from `/api/status`, fetched with each briefing load). It hides when the plugin serves none — offline, or no cron scheduled — the same hide-rather-than-invent rule the compile stamp follows, and a hover title notes that a manual update can land earlier.
+- Eclipse and moon-phase sky events (work doc #35): the ephemeris now emits discrete moon phase instants (`moonPhase` notes — "New Moon — darkest skies of the passage", "Full Moon — moonlit night watches") and visible eclipses with almanac's built-in observer visibility — lunar eclipses gated on the Moon up at peak, the Sun below the civil twilight line and the cloud gate; solar eclipses on the Sun above the horizon and the cloud gate. Timeline and here-view sky cards carry them like every other event; nothing new is fetched, all local computation. Like every computed-on-board item they carry no provenance record (work doc #31: absence never implies external authority) — pinned by a regression test so the weather provenance chain never leaks onto sky rows.
+- Reentry watch and visible launches (work doc #36): the briefing now answers "could we see that from here?" for the two rarest sky events. Predicted satellite reentries come from Space-Track's TIP messages (the only machine-readable decay forecast; a free account configured once in the plugin, degrading to off without it), refined by CelesTrak element sets and propagated on board with the same SGP4 propagator as the station passes — the visible-event gate models both regimes the Cluster II SAMBA/TANGO deorbits taught us about: fireball at night out to the geometric horizon of the burn altitude (~1000 km), sunlit smoke streak by day readable only close, everything else correctly silent (the backtest that proved the Aitutaki→Niue crew could never have seen those two is now a permanent fixture test). Visible rocket launches come from Launch Library 2 (no account) and fire only for the classic twilight geometry — pad within ~650 nm and the observer in twilight while the vehicle climbs into sunlight — with the bearing to the pad and an honest windows-slip caveat. Both sources follow the house fetch rules: online-window fetches, data-directory caching (TIP 12 h, launch schedule 6 h, stale cache still answers offline), per-source status in the registry with the unconfigured reentry source recorded as absent, and every failure degrading to no events.
+- Eye height above the water as plugin configuration (`celestial_eye_height_m`, default 0): sunrise, sunset and moonrise/set stamps in `celestialNights` — the timeline's night rows and the conditions-here sky line — correct for the dip of the vessel's raised horizon (a couple of minutes from cockpit height). Twilight lines are altitude angles, not horizon crossings, so they stay put.
+
+### Changed
+- Meteor showers are evaluated per night, not per peak instant (work doc #37): each shower's nominal peak is judged against the nights bracketing it — the peak night plus its neighbors, since real maxima drift ±1 night — at each night's nautical-dusk anchor, with the radiant altitude checked at the **forecasted position of that night** from the simulated track. A shower whose catalog peak hour lands in local daytime now fires for the dark nights around it instead of vanishing (the Eta Aquariids from New Zealand: 14:00 peak hour, radiant near zenith, moon dark — previously unreportable); descriptions word the bracket nights honestly ("climbing toward the peak" / "just past peak") so no night claims a maximum its darkness never held. The evaluation lives in a shared dependency-free module (`public/components/meteors.mjs`) run where the track lives — the simulation worker computes the route timeline's meteor nights from the track exactly like the sail changes, while here-mode payloads keep them from the server at the vessel position; route payloads no longer carry fixed-position `meteor` entries, so nothing double-reports. The static annual calendar stays (IMO publishes tables, not APIs) but moves into the shared module with its thresholds.
+- The test runner no longer uses `--test-force-exit`: the flag was intermittently killing suites whose subtests register on timers, undercounting the run while everything passed. Without it the whole suite exits cleanly and the count is stable.
+- The ephemeris engine is `@openwaters/almanac` (MIT, the maintained TypeScript port of astronomy-engine at a pinned commit, checked against JPL/USNO/Espenak fixtures over 1950–2100), replacing astronomy-engine — same math lineage and accuracy class, 21.9 KB vs 44 KB, and the twilight chain collapses from three chained searches to window event lists. Rise/set stamps shift by seconds (almanac uses the Sun's true semidiameter instead of a fixed −0.833°); everything else stays behaviorally identical. The npm package is ESM-only, loaded through `require(esm)` — the minimum Node moves to 22.12.0.
+
 ## [0.8.1] - 2026-10-08
 
 ### Added

package/README.md

@@ -41,7 +41,7 @@ The Open-Meteo entries below are the fallback weather source: used when no Weath
 - [Open-Meteo Forecast API](https://open-meteo.com/en/docs) — surface wind, gusts, MSL pressure, CAPE and the pressure-layer fields behind the K-index (best-match model)
 - [NOAA SWPC planetary K-index forecast](https://services.swpc.noaa.gov/products/noaa-planetary-k-index-forecast.json) — geomagnetic activity behind the aurora advisories
 - [JPL SBDB query API](https://ssd-api.jpl.nasa.gov/doc/sbdb_query.html) — comet brightness parameters (M1/K1) behind the Sky Notes comets
-- [CelesTrak](https://celestrak.org/NORAD/elements/gp.php?GROUP=stations&FORMAT=tle) — two-line elements for the crewed stations (ISS, Tiangong); bright satellite passes are propagated locally from the fetched TLEs with SGP4, gated to nautical night, clear sky and a 15° haze line
+- [CelesTrak](https://celestrak.org/NORAD/elements/gp.php?GROUP=stations&FORMAT=tle) — two-line elements for the crewed stations (ISS, Tiangong); bright satellite passes are propagated locally from the fetched TLEs with SGP4, gated to nautical night, clear sky and a 15° haze line. The same service's per-catalog queries (`gp.php?CATNR=…`) supply the current elements of decaying satellites named by Space-Track's reentry predictions
 - [NOAA TGFTP radiofax tree](https://tgftp.nws.noaa.gov/fax/) and the [BoM difacs charts](http://www.bom.gov.au/difacs/) — synoptic surface-analysis charts per METAREA zone (work doc #11), per the schedule in [otherfax.txt](https://tgftp.nws.noaa.gov/fax/otherfax.txt)
 - [UKHO Admiralty MSI API](https://msi.admiralty.co.uk/) — structured NAVAREA navigational warnings for the resolved zone; the third rung of the GMDSS bulletin fetch ladder (TGFTP raw text, WMO GMDSS portal, then UKHO MSI JSON) and also accepted verbatim as a `bulletin_urls` source
 - [GDACS RSS](https://www.gdacs.org/xml/rss.xml) — EU JRC global disaster alerts (earthquakes, tropical cyclones, floods, volcanic activity), filtered route-relative by alert level and corridor distance: they ride the unified passage timeline and publish as georeferenced chart notes (work doc #22)
@@ -49,6 +49,8 @@ The Open-Meteo entries below are the fallback weather source: used when no Weath
 - NOAA TGFTP (`tgftp.nws.noaa.gov`) — raw METAREA bulletin text for the resolved GMDSS zone's station (e.g. `FQPS01 NFFN` for XIV)
 - WMO GMDSS portal (`weather.gmdss.org`) — per-zone bulletin pages, fallback when the TGFTP station is not configured
 - api.weather.gov product API — US High Seas Forecast texts (default `bulletin_urls`: HSF NP/EP1/EP2); extra sources can be added via the `bulletin_urls` configuration (plain text or api.weather.gov product URLs; UKHO MSI JSON works too)
+- [Space-Track](https://www.space-track.org) — TIP (Trajectory Impact Prediction) messages, the only machine-readable source of satellite reentry predictions; needs a free account configured in the plugin, without which the reentry watch stays off. The predicted decay is refined on board with the object's CelesTrak elements and the vessel-visible geometry (fireball at night, sunlit smoke streak by day) computed from the propagator and the sky, not taken from the message
+- [Launch Library 2](https://thespacedevs.com/llapi) — upcoming rocket launches with pad coordinates (no account); the briefing names only launches whose twilight geometry would make the plume visible from the vessel's position
 
 All external fetches are online-gated and cached to the plugin data directory, so the last payloads survive the offline hours.
 

package/SPEC.md

@@ -197,7 +197,7 @@ interface UnifiedWeatherPayload {
     bulletinText: string;
   };
   spaceEvents?: {
-    kind: string;               // 'aurora' | 'comet' | 'conjunction' | 'opposition' | 'meteor'
+    kind: string;               // 'aurora' | 'comet' | 'conjunction' | 'opposition' | 'meteor' | 'moonPhase' | 'lunarEclipse' | 'solarEclipse' | 'satellite' | 'reentry' | 'launch'
     timestamp: string;          // ISO timestamp
     tactical: boolean;          // 24h dashboard banner vs strategic sky note
     description: string;
@@ -564,7 +564,14 @@ function filterExceptions(simulationResult) {
       totalMotorHours: simulationResult.motoringHours,
       totalFuelGal: simulationResult.fuelConsumptionGal,
       macroSeaAnomalies: simulationResult.seaStateAnomalies.filter(a => a.steepnessRatio < 3.28),
-      convectiveWarnings: simulationResult.upperAirAnomalies.filter(u => u.cape > 1000 || u.kIndex > 28)
+      convectiveWarnings: simulationResult.upperAirAnomalies.filter(u => u.cape > 1000 || u.kIndex > 28),
+      // Meteor-shower nights evaluated along the track (work doc #37):
+      // per (shower, night) spaceEvents gated at the night's nautical-dusk
+      // anchor, the radiant checked at the track's forecasted position;
+      // route payloads carry no fixed-position `meteor` spaceEvents —
+      // these are the timeline's meteors. Here-mode payloads keep them
+      // in `spaceEvents` (single position, no track)
+      meteorEvents: simulationResult.meteorEvents
     }
   };
 }

package/package.json

@@ -1,12 +1,12 @@
 {
   "name": "@meri-imperiumi/signalk-passage-briefing",
-  "version": "0.8.1",
+  "version": "0.9.0",
   "description": "Offshore passage daily briefing webapp for Signal K",
   "main": "plugin/index.js",
   "scripts": {
     "format": "npx @biomejs/biome check --use-editorconfig=true --write plugin/ tests/ bin/ public/",
     "lint": "npx @biomejs/biome check --use-editorconfig=true plugin/ public/",
-    "test": "node --test --test-force-exit \"tests/*.test.js\""
+    "test": "node --test \"tests/*.test.js\""
   },
   "repository": {
     "type": "git",
@@ -41,11 +41,11 @@
     ]
   },
   "engines": {
-    "node": ">=22.5.0"
+    "node": ">=22.12.0"
   },
   "dependencies": {
+    "@openwaters/almanac": "^0.8.0",
     "@openwaters/maritime-zones": "^0.2.1",
-    "astronomy-engine": "^2.1.19",
     "satellite.js": "^7.1.0"
   },
   "devDependencies": {