Commit Graph
10 Commits
Author SHA1 Message Date
Sameh Khamis 2a26d13fff render: wind relief is the globe port's hillshade; the wind layer no longer reaches Cesium
The wind relief now asks the globe port for hillshade surface shading at
0.28 instead of building a globe material itself; the Cesium globe port
draws the same view-directed shading, over terrain normals or the globe's
curvature, and the unbound provider now lists surfaceShading. The wind
layer's canvas flow, bounds, imagery fade and render requests go through
the camera, frame and query ports, so src/layers/wind/rendering.js and
relief.js leave the Cesium allow list.

A refused shading request is taken back so it never shows after Wind is
off, and a camera height that is not a number falls back as before.

The application layers and standalone catalog entries no longer load the
unbound factories themselves: a viewer with no renderer bound
needs render/cesium/unbound loaded first. The contract and current-state
docs say what the Cesium adapter's fields and surface shading do now.
2026-11-09 08:13:04 -08:00
Sameh Khamis fcd1c09565 render: Cesium draws wind-streamline fields; the wind layer's flow uses the draw port
The Cesium draw port now implements draw.field for the windStreamlines
kind (src/render/cesium/drawStreamlines.js). It carries the flow shader,
the 30-degree cells, horizon and frustum culling and the path tracing over
from the wind layer. It moves the highlight on the page's own frames, at
most every 32 ms, and only while the field shows, is unpaused, has alpha
and some cell is in view on a visible page. A grid no path runs through,
or one short of the whole globe, is refused. A rebuild that fails draws
nothing and is never told drawn.

The wind layer's GPU flow (src/layers/wind/gpuRendering.js) now builds the
spec, fades it by camera height and pauses it through the draw, camera and
frame ports, and no longer reaches Cesium, so it leaves the Cesium allow
list. The layer runs no animation loop of its own while the renderer draws
the flow; the canvas fallback is unchanged and still takes over wherever
the field is refused. A new conformance check covers the field's 'drawn'
events.
2026-11-09 07:38:06 -08:00
Sameh Khamis 50cb7bd5f4 render: Cesium draws raster-shell fields; weather shells use the draw port
The Cesium draw port now implements draw.field for the rasterShell kind
and advertises the 'fields' capability. The shell surface (a raised
rectangle with a detail window, drawn in primitive order without depth
writes) moves from the weather layer into src/render/cesium/drawFields.js.
A new rectangle or height is drawn invisibly first and swapped in whole,
and 'drawn' is told in the frame of the swap. A stage of older content
never replaces the drawn one while newer content decodes. A field hidden
mid-stage keeps building it.

The weather shell (src/layers/weather/shellRendering.js) now drives one
field through the draw, camera, frame and globe ports and no longer
imports Cesium, so it leaves the Cesium allow list. The wind shell on
photoreal hosts also draws through draw.field. The conformance page runs
each visual scene on its own, so a kind the Cesium renderer does not draw
yet (scalarRaster) fails only its own scene.
2026-11-09 07:02:54 -08:00
Sameh Khamis 03ff8f291f Read the wind layer's map-centre reading and its marker through the query and frame ports
The wind reading at the map centre now asks the renderer's query port where
the ellipsoid lies under the middle of the canvas and converts it with the app's
WGS84 helpers, instead of Cesium's camera pick and Cartographic. The
inspection marker projects its point with the query port, hides it beyond
the ellipsoid's horizon with isOccluded, and follows rendered frames through
the frame port's postRender phase instead of the scene's event. The layer
entry no longer carries a Cesium namespace: the rendering part, which stays
Cesium-bound, defaults it, owns the fallback globe imagery host, and loads
the adapter factories a viewer with no renderer bound gets its ports from.

The wind entry, its inspection reading and its marker leave the allow-list.
They still read the viewer's canvas size and container element.
2026-11-07 01:50:42 -08:00
Sameh Khamis 76880f6eae Draw the weather and wind imagery through the renderer's globe port
Radar, satellite clouds and lightning frames are imagery layers of the
globe port instead of Cesium providers in the viewer's collection. A frame
stages as a transparent layer on the globe's surface and counts its tiles
through the layer's admit hook; the global infrared mosaic draws each tile
of the geographic grid from the decoded canvas, the regional infrared
transfer runs as the layer's tile transform on the upright tile, and
throttled tiles are fetched again by the renderer. Readiness reads the
port's imagery settling, the camera port gives the view and its move end,
and the frame's credit, which a refreshed manifest can still change in
place, is owed through the credits port while the layer is shown.
Without a host the shown frame is hidden rather than detached.

Wind's colour field is a port layer too, so the draw order is a zOrder per
product (wind, infrared, radar, lightning), above the port's other
imagery, and the raise-to-top helper goes. The Cesium port keeps a tile
transform's bitmaps in straight alpha and closes the copies it spends,
and a provider that fails to build is told as an error of its root tile.

The weather rendering and raster tile modules no longer reach Cesium and
leave the allow-list; the 3D Tiles shell still does.
2026-11-06 20:06:33 -08:00
German KhokhlovandClaude Opus 5.5 20f4cf3763 i18n: fix the LGTM review findings and tidy the branch texts
- i18n scanner: lastError sink, no false sinks on defaults and ternaries
- shared src/i18n/marker.js for modules that only mark strings
- wind field messages move to wind/fields.js; CITY_POIS imports from cityPois.js
- drop review IDs, change history and stale references from comments and docs

Co-Authored-By: Claude Opus 5.5 <[email protected]>
2026-10-16 21:00:28 +04:00
German KhokhlovandClaude Opus 5.5 3157888780 i18n: fix the repeat-review findings
- Status pieces translate in one place (src/ui/statusText.js): a Local
  ADS-B error with a single piece now translates, and the analyst card
  splits empty-layer notes into pieces (Transit, Local ADS-B).
- Translate the ALPR retry chip text, the ECCC error on the Air quality
  card, "Upstream response too large" and the "entry {n}" feed label.
- The unmarked-text check also sees error/loadingLabel/statusMessage/
  status assignments, `status:` properties and showToast?.( /
  this._showToast( calls; the scene import preview marks its pack
  statuses, data and model-facing text carry i18n-ignore with a reason.
- Tests check the real code: the HUD locality formatter moves to
  src/hudLocalityFormat.js, the unsaved-keys test pins the teardown.
- One copy each: city names come from src/cityPois.js, the GNSS HTTP
  template and wind "field unavailable" texts; dead fallbacks removed.
- Leftovers of the reverted vessel-label change and the reformatting of
  detectionHost.test.mjs / keySetup.test.mjs are back to main.

Catalogs: es/ru 2617 messages each, complete.

Co-Authored-By: Claude Opus 5.5 <[email protected]>
2026-10-16 19:28:59 +04:00
German KhokhlovandClaude Opus 5.5 3adce33cc6 i18n: fix the final pre-merge review findings
Fixes the P1 and the confirmed P2 findings of the final review of the
i18n branch.

- CI: the air-quality model no longer imports i18n (portable module),
  so check:boundaries passes again.
- Air quality layer: named in the layer panel, HUD and voice; its
  Weather card, bands and errors translate. A test keeps the layer-name
  table in step with the layer registry.
- Layer statuses that still reached ru/es in English: GNSS, CelesTrak,
  Transit loading and feed errors, launches, perimeters, ALPR, ECCC,
  wind "field unavailable", Local ADS-B streaming lines.
- Voice and HUD: preset ids show city names ("Fly to San Francisco"),
  the HUD city translates, cockpit refusals cover frame, aircraft and
  satellite; empty Transit notes translate on the analyst card.
- tn() keeps a caller's formatted count ("12,345 map objects").
- Cyber HUD: the language button no longer covers the bottom-right
  card or the cockpit position readout.
- CCTV buttons wrap only in non-English locales; English layout matches
  main again.
- Switching language asks first when Provider Settings holds unsaved
  keys.

Catalogs: es/ru 2611 messages each, complete.

Co-Authored-By: Claude Opus 5.5 <[email protected]>
2026-10-16 16:39:25 +04:00
German KhokhlovandClaude Opus 5.5 54dce5b3bd fix(i18n): second screenshot review: remaining English, numbers and wraps
From the re-shot Russian screenshots:
- Translated: weather frame/imagery status errors, HUD summary provenance
  ("НЕДОСТУПНО: СУДА AIS ОНЛАЙН"), Bhote Koshi map labels, cockpit signal
  distances (КМ). The HUD summary fields are separated with " · " in every
  language, so the style name and view band no longer read as one phrase.
- Numbers follow the locale: cockpit altitude ("26 175"), precipitation.
- No breaks inside codes and dates: ADS-B and similar in layer rows,
  Sentinel-2, cyclone issue dates (word joiner; the UI font has no U+2011),
  no-break space after label colons in cyclone details.
- Layout: count lines under ALPR, short unavailable chip (context
  "status chip"), roomier draw segments, tighter CCTV calibration chips,
  two-line beat titles and media card titles, wrapping weather status.
- Wording: ru advisory is «сводка» (was «бюллетень»; "Advisory center" is
  «Центр циклона по сводке»), Street View is «Просмотр улиц», OPACITY
  «НЕПРОЗР.», synced «синхр.», Area «Зона»; the import dialog no longer says
  "this scene" for several scenes. Glossary updated.

Co-Authored-By: Claude Opus 5.5 <[email protected]>
2026-10-13 19:13:12 +04:00
0251c74109 feat(weather): add globe wind, observed weather and cyclone context (#617)
Wind (NOAA GFS and ECMWF IFS: animated particles, colour field, a reading at map centre), observed weather (NOAA MRMS radar, GOES and NESDIS infrared clouds, lightning density) on one shared history timeline, NHC and CPHC cyclone advisories with tracks and cones, and a WEATHER panel in the right rail. On 3D Tiles the observed products render as raised shells with a sharper window around the view.

Co-authored-by: Gustavo Beneduzi <[email protected]>
Co-authored-by: Daniel Slay <[email protected]>
Co-authored-by: Rehaan Delmotra <[email protected]>
Co-authored-by: Cursor Agent <[email protected]>
2026-09-22 17:45:39 -05:00