Commit Graph
3080 Commits
Author SHA1 Message Date
Sameh Khamis 0ac06dad2d fix(shipwrecks): a reversed period is stored in the order the picker shows it, so changing one end keeps the other
A period set as 1950 to 1900 (a link or the voice tool) showed as 1900–1950
but stayed reversed, so changing "From" to 1920 gave 1920 to 1900, shown as
1900–1920: the end the reader did not touch moved.
2026-11-10 13:25:25 -08:00
Sameh Khamis 38ba8a3b59 perf(shipwrecks): draw the wreck markers from one billboard collection, on the wreck, and rebuild them only on new data
- The ~8,000 markers were one Cesium entity each; they are now one
  BillboardCollection (as FIRMS, vessels and wildlife sightings do), ordered
  at the bottom of the shared sprite stack so moving traffic draws over it.
- eyeOffset.z was the death toll in metres, so a 1,500-death wreck's marker
  sat 1.5 km toward the camera and clipped or drifted when close. It is gone;
  markers are added by rising toll, so a bigger toll still draws on top.
- The page asks every 6 h but the proxy changes data at most daily: the proxy
  now sends a version that changes with either source's data, and the layer
  rebuilds only when it does.
- Adds a controller test: enable/disable, teardown, draw order, unchanged
  snapshots, a late response after turning off, and the period closing an
  open card.
2026-11-10 13:25:03 -08:00
Sameh Khamis 3b52e24140 fix(shipwrecks): a failed or empty source refresh keeps its last good wrecks, and waits 15 minutes before the next try
A refresh with one source down (or Wikidata answering 200 with no results)
was cached as the day's snapshot, replacing the full set; and once the
snapshot was a day old, every request after a failure fetched both sources
again. Each source now keeps its own last good records, served as stale for
up to seven days, and a source that failed is not asked again for 15 minutes,
even when there is no data yet. A Wikidata answer without results, or with no
wreck in it, counts as a failure.
2026-11-10 13:20:49 -08:00
Sameh Khamis a1200efba6 fix(shipwrecks): an end-of-service date (P576) keeps its precision, so a year-only one is not shown as 1 January
The query read P576 through wdt:, which carries no precision, and a missing
precision counted as a day. It now reads the best-rank statement's value with
wikibase:timePrecision, and a value without a precision is trusted to the
year only. Against the live query service the new query takes 23 s and dates
16 more wrecks while every date the old one gave is unchanged.
2026-11-10 13:19:23 -08:00
Sameh Khamis 0da3a28d76 fix(shipwrecks): a Wikidata date before AD 100 reads as ancient text (300 BC), not as the AD year
parseYear took the minus sign of -0299 as a separator, so a wreck from 300 BC
showed as year 300, sat at AD 300 in the period picker and stretched its
range. Years before AD 100 now keep their year as text ("300 BC", "AD 50")
and count as ancient, like EMODnet's period-dated wrecks; the query service's
XSD 1.1 years (0000 is 1 BC) are converted. Merging records now carries that
text with the date, and coarser-than-year precisions give no date.
2026-11-10 13:17:50 -08:00
Sameh Khamis 32e7a1a913 fix(shipwrecks): skip every wrecksite.eu row, even when the name comes after the 80-character cut or only in the link
The exclusion tested the provenance after it was cut to 80 characters, so a
long source list ending in wrecksite.eu kept the row, and a row whose only
trace was a wrecksite.eu page link was kept too. Both raw values are now
checked before anything is cut.
2026-11-10 13:16:42 -08:00
Sameh Khamis 447cc08783 fix(flights): a stalled feed dims the tracked aircraft as STALE
When the whole feed backs off, no poll counts a miss, so the tracked
aircraft read STALE at full brightness while no newer position came in.
It now takes the same stale fade as a missed update, in the flights and
military layers.
2026-11-10 13:03:15 -08:00
Mazeyar Moeini FeizabadiandClaude Fable 5.2 bd1b16e4ee fix(news): on phones the News panel keeps its height in the Context sheet
It shrank under the Context panel, leaving the feed a strip one card high.

Co-Authored-By: Claude Fable 5.2 <[email protected]>
2026-11-10 21:51:34 +01:00
Mazeyar Moeini FeizabadiandClaude Fable 5.2 caa6f6f812 fix(news): the browser keeps Local news an hour at most; docs say what the fixes do
Local kept World News API's rows listed after the server had deleted them.
README names News among the routes refused behind a proxy; a timed-out
request is said to count; the cockpit cached state is near the same place.

Co-Authored-By: Claude Fable 5.2 <[email protected]>
2026-11-10 21:41:29 +01:00
Mazeyar Moeini FeizabadiandClaude Fable 5.2 671a256a6a fix(news): follow-up review of the fixes
Deleted World News API news is not restored by a refresh that fails; a
refund goes only to the day the read was counted in, and not to a request
that timed out; a cockpit brief from another place is dropped, not kept for
a language change; a feed list started over lets scrolling ask again.

Co-Authored-By: Claude Fable 5.2 <[email protected]>
2026-11-10 21:34:58 +01:00
Mazeyar Moeini FeizabadiandClaude Fable 5.2 bb7e7ecad6 fix(news): Hebrew strings for the News layer and the gate modules in the news boundary
main added the he locale and the news provider now imports the same-site gate.

Co-Authored-By: Claude Fable 5.2 <[email protected]>
2026-11-10 21:33:01 +01:00
Mazeyar Moeini FeizabadiandClaude Fable 5.2 9c0ad0a730 fix(cockpit): a failed refresh after a move never shows the last location's brief
Co-Authored-By: Claude Fable 5.2 <[email protected]>
2026-11-10 21:26:48 +01:00
Mazeyar Moeini FeizabadiandClaude Fable 5.2 369b8e924a fix(news): show the News panel in the phone Context sheet only
Co-Authored-By: Claude Fable 5.2 <[email protected]>
2026-11-10 21:26:48 +01:00
Mazeyar Moeini FeizabadiandClaude Fable 5.2 fd9373c09c fix(news): keep the feed cursor in step with the list and drop Local reads for an area left
Co-Authored-By: Claude Fable 5.2 <[email protected]>
2026-11-10 21:26:48 +01:00
Sameh Khamis 05daaf267a Merge branch 'main' into feat/transit-icon-heading 2026-11-10 12:24:46 -08:00
Mazeyar Moeini FeizabadiandClaude Fable 5.2 3342d7a07a Merge origin/main into review/pr-1593
Conflicts: docs/CURRENT-STATE.md layer table, src/ui/layerPanel.js imports,
src/data/dataCredits.test.mjs imports; both sides kept.

Co-Authored-By: Claude Fable 5.2 <[email protected]>
2026-11-10 21:20:15 +01:00
German Khokhlov c76832ca69 Street Level: drop the unused coverageLines diagnostics
Nothing reads the drawn coverage lines: the browser gate aims elsewhere, so
lineCoverage's drawnLines, the providers' coverageLines and the layer's
coverageLines go, with their tests and the doc mention.
2026-11-10 23:58:03 +04:00
German Khokhlov e34b5103c6 Street Level: shorter comments and docs
- Comments in the action, the layer, embed, the Mapillary policy and the
  controls say why in fewer words; restatements of the code are gone.
- CURRENT-STATE's "Show me X on Street View" section is a list by topic and
  no longer says the browser gate aims at coverageLines.
- TOOLS.md's street-level view paragraph is shorter.
2026-11-10 23:54:16 +04:00
German Khokhlov f3996da0af Street Level: a landed link, a replaced photo and a retired lookup keep their state
- A link's photo step compares with the navigation token its own restore
  flight owned, so a navigation after the flight landed (while the layers
  still load) keeps the camera.
- A request that switches the source keeps the open photo up until it is
  replaced (holdViewer), so EXPAND stays across sources.
- A lookup given up while it looked leaves the open photo and the panel as
  they were, also when it found nothing or failed.
- A failed replacement shows its error inside the expanded viewer.
- A replacing lookup stops a playing Mapillary sequence before the found
  photo opens.
- The action refuses coordinates outside the globe, keeps a name given with
  coordinates as its label, and fly: false keeps the camera only for
  coordinates (its test-only entity and query branches are gone).
- Numbered entities carry their label; a view's street-level point is a pair.
- Mapillary's nearest search always asks for 50 m and 100 images.
- Tests: the lifecycle reproductions, PLAY on replacement, out-of-range
  coordinates, half a point, and the named label; three layer tests share
  enabledLayer; the second payload digest is gone.
- Docs: the tool count, the panel freshness rule, armed: false, the photo
  replacement, the browser-gate comment and a shorter changelog fragment.
2026-11-10 23:48:57 +04:00
Sameh Khamis d1cd7f63ef Merge branch 'main' into sameh/route-corridor 2026-11-10 11:45:29 -08:00
Sameh Khamis 486a78ad8e fix(flights): a long flight keeps its route when it is up to 15% of the route's length off the great circle
The route check allowed 200 km off the great circle between the airports, so a
coast-to-coast flight routed a few hundred kilometres around weather lost its
origin and destination (JFK–SFO over Kansas, 384 km south). The tolerance is
now 200 km or 15% of the route's length, whichever is larger: 623 km for
JFK–SFO, and still 200 km for a short hop, so a wrong leg stays hidden.
2026-11-10 11:37:04 -08:00
Sameh Khamis 8fec5a76bf feat(flights): a tracked flight's trail splits at gaps, follows the drawn aircraft, and a lost aircraft coasts then reads SIGNAL LOST
The trail no longer joins a day of history into one line. It lifts the pen
across a gap of more than 48 minutes, an implied speed over 1,500 kt, or a
report without altitude (a parked aircraft within 0.1 nm for under 3 h stays
joined), and every history point is kept as its own segment. History never
gets an invented altitude. The live end is the path the aircraft was drawn
along, thinned to its shape, so it no longer cuts corners through turns; a
late report trims what was drawn after it. Old segments are frozen in chunks,
so a 2,000-point trail costs about 0.1 ms a frame.

A tracked aircraft that misses an update shows STALE and is dimmed. It keeps
flying on its last speed and heading for five minutes (at most 1,500 kt),
then stops, dimmed, labelled SIGNAL LOST with the time since its last report,
and is kept for ten minutes after that report so the camera keeps its target.
Aircraft on or near known ground hold at their last report instead of
coasting. Military flights follow the same rules.
2026-11-10 11:30:59 -08:00
Sameh Khamis 6eb0729832 fix(firms): translate the snapshot-end label for Hebrew 2026-11-10 10:35:16 -08:00
Sameh Khamis bb45aacb24 Merge branch 'main' into sameh/firms-window-label 2026-11-10 10:27:01 -08:00
Sameh Khamis 4972aac345 Read a material's held texture through internals.js, pinned to Cesium's Material
dropPendingTexture reads the private material._textures; it now does so
through internals.js like every other private field. The internals test
fails when the installed Cesium no longer declares _textures or no
longer takes a uniform's texture as Material updates.
2026-11-10 09:50:22 -08:00
Sameh Khamis f1817756ed Read conformance pixels where they are drawn, and skip them only by the harness
Checks that read drawn pixels no longer skip when nothing is drawn: a
harness says pixels: false when its renderer draws only models into
pixels, as the stub, and a renderer that draws an icon, an
outlined line or a raster shell as nothing fails. The shell's patch is
read where the ground below projects, in the units readPixels crops in.
The contract notes what changed in version 0 besides additions, and
loses two stale lines.
2026-11-10 09:46:12 -08:00
Sameh Khamis d5e3968caa Draw outlined draped lines with their outline; relay pinches for an early Ctrl+wheel
A draped solid line with an outline went into a colour batch and lost its
edge; it now draws by its own outline material. A camera controller that
binds Ctrl+wheel before the controls are made relays trackpad pinches from
the start, not only after the binding is set again. A raster shell image
zero pixels wide is left to the material, as before.
2026-11-10 09:46:11 -08:00
Sameh Khamis 8ca657e63f Run camera.setView.exact on the Cesium renderer in the conformance page 2026-11-10 09:46:11 -08:00
Sameh Khamis c56ff090e6 Destroy an image fill's texture its material never took
A material takes a texture, and destroys it later, only when it next
updates. A canvas fill replaced by one of another size twice before a
frame, or disposed before one, left the texture in between to no one.
Such a texture is now destroyed when it is replaced or its shapes go.
2026-11-10 09:46:03 -08:00
Sameh Khamis f834903950 Place the stub's camera exactly from a direction and up
The contract places a view given as direction and up exactly, but the
stub turned it into heading, pitch and roll first, so a thousand small
turns straight down drifted by a hundredth. The new check
camera.setView.exact holds the stub and Cesium to it; both pass.
2026-11-10 09:45:49 -08:00
Sameh Khamis ac0496a785 Keep the range ring's 80 ms step while a drag wakes its pulses
A user's drag woke the pulses on every frame it drew, and each wake
stepped the fronts at once without noting the step, so a drag stepped
every frame and some frames twice. A wake now starts resting pulses at
once and leaves running ones to their Cyber sonar interval; a subject
move that steps notes it too.
2026-11-10 09:45:49 -08:00
German Khokhlov b6c2ba379b i18n(he): translate the photo credit added on main 2026-11-10 21:26:15 +04:00
German Khokhlov 2919e90c63 Merge branch 'main' into feat/hebrew-i18n 2026-11-10 21:24:21 +04:00
Sameh Khamis 3f194d0eec Check that a raster shell draws its image, not only that it says so
draw.field.rasterShell now reads the middle of the frame, which the shell
covers, once 'drawn' has come. A renderer that draws a probe line into
pixels must draw the shell there too; one that draws neither, like the
stub, leaves the pixels unread.
2026-11-10 09:23:43 -08:00
German Khokhlov 7571e42556 i18n(he): apply the second translation review
- RECORD is «רשומה», a record rather than a recording.
- Time abbreviations end in the geresh again (דק׳, שנ׳, שע׳, ימ׳), and the
  counted forms agree with 1.
- Numbers after a Latin name, ordinals and units keep their order in RTL.
- Meaning, agreement and domain terms fixed across the catalog: cyclone
  advisories and storm centre, sensor swath, track-up, cinematic, tropical
  depression, pilot and sailing vessels, chips, metered billing, radio static,
  airspace, NOT AN ALL-CLEAR, and the remaining review items.
- LOOK and OFF-NADIR are translated; {value} KM stays Latin as an instrument
  readout, like KM/S.
- rtl.css shares the Arabic and Hebrew letter-spacing reset and drops the
  duplicated Arabic rules; he.md and its glossary follow the catalog.
2026-11-10 21:23:28 +04:00
Sameh Khamis 72a322f7ce Record Cesium references for the quad, mesh and field scenes
Every conformance scene now has a reference. The quad and mesh views look
at their item's middle, where the mesh's base was cut off at the bottom of
the frame before.
2026-11-10 09:07:51 -08:00
Sameh Khamis 46d44eed89 Draw a one pixel wide raster shell image
The shell's material reads an image width of 1 as Cesium's 1x1 default
texture, nothing arrived yet, so a one pixel wide image never drew while
'drawn' was still told. Such an image is drawn as a copy two pixels wide,
which samples the same.
2026-11-10 09:07:51 -08:00
Sameh Khamis 149f8aad4d Settle a line outline's width as the whole outline, half on each side
A solid line's outline.widthPx is the outline in all, half on each side and
inside the line's width, as Cesium draws it. The range ring's pulse fronts
pass their halo width whole again, so they keep main's look, and the
conformance check draw.lines.outline pins the width. The lines scene's
outlined line keeps its 2 px edge each side and its references are
re-recorded.
2026-11-10 08:59:36 -08:00
Sameh Khamis 94138d9698 Draw CCTV lines on a renderer without meshes or quads
Meshes and quads are capabilities, so the CCTV layer asks for its viewshed
and plane sets only where the renderer has them and draws the rest of its
items without them.
2026-11-10 08:59:28 -08:00
Sameh Khamis 4ef0869a87 fix(firms): drop LIVE from the row's source and the voice answer for a snapshot that ends at a set time
The catalog source 'NASA FIRMS · LIVE' still led the row; getStats() now
names the source 'NASA FIRMS' for such a snapshot. get_active_fires says
"in the 24 hours to" its end and reports window_end. The end is checked
to be a printable time, cleared under KEY REQUIRED, the test moves next
to the layer, and the changelog and CURRENT-STATE describe it.
2026-11-10 08:19:40 -08:00
Sameh Khamis 43706b8dd0 fix(firms): label a snapshot that ends at a set time with that time, not LIVE
A FIRMS source may answer the 24 hours to a set time, naming its end as
`at` (epoch ms). The layer row then reads "24 h to MM-DD HH:MM UTC"
instead of "LIVE · updated ...", and getStats() reports `at`. A
snapshot without `at` (the trailing 24 hours to now, as the dev proxy
serves) keeps the LIVE label. Translated in every catalog.
2026-11-10 08:11:58 -08:00
Sameh Khamis 3bca55831f Step the range ring's fronts and band on the Cyber sonar interval while the camera moves
The fronts and band are fixed in the world, so a camera move needs nothing
of them: they now step 80 ms apart however many frames are drawn, as a
parked view's do, and look the same moving or parked. Each frame checks
only whether the camera crossed the rim height, which shows or hides them
at once.
2026-11-10 08:07:22 -08:00
Sameh Khamis e4af6c8ed7 Cut the range ring's frame cost while the camera moves
The pulse level is worked out once a step, so the camera pose is read
once rather than on every call. The band's alpha is tabulated along the
radius and each pixel looks it up (0.78 ms a paint down to 0.21 ms in
Node). A line sent again where it already runs keeps its positions, so the
polyline collection does not rewrite the ring's vertices on every step.

The band test now compares images byte by byte. A failing deepEqual on a
256x256 image builds a diff of every byte, and that ran a test process
out of memory.
2026-11-10 07:59:23 -08:00
Sameh Khamis 74d6d97b29 Draw the Global Context range ring and its pulses through the draw port
The ring, its sonar fronts and the band behind them leave Cesium: the
ring and each front are one solid line with an outline, the band is one
flat unpickable ellipse at the rim filled with a canvas painted from the
pulse radii. rangeRingCesium.js goes, and the Cesium allow-list is empty.

Solid lines take an optional outline ({color, widthPx}, inside the
line's width on each side), drawn in the line's own pass so a
translucent halo never darkens where it meets the core; both colours
change in place, and outlined lines never settle into the static batch.
An unclamped image fill takes a new canvas in place: copied into its
texture when the size is the same.

A parked view steps the pulses from the 80 ms timer before it asks for
the frame that draws them, so it draws one frame a step; frames the
camera moves in, or held drawing, step them every frame, and the band is
painted at most every 80 ms. The lines reference scene gains an outlined
line.
2026-11-10 07:48:07 -08:00
Said Turksever 9c73939425 Street Level: leave an interrupted link and an overtaken view alone (gekh review)
- A pasted or start-up link opens its photos only when its restore's camera
  landed: a drag or search during the restore flight keeps the camera.
- A posted view whose photo step was cancelled or superseded no longer closes
  the viewer or flies; one that found a photo it could not load keeps the error.
- openNearest gains `replace` (keep the open photo while it looks, close it
  only when nothing opens, so EXPAND stays) and `proceed()` (asked before a
  found photo loads); the action uses both, so a navigation, a gesture or a
  hiding mode during the lookup loads no billed panorama.
- Mapillary's timeout message is a msg() string with all eleven translations.
2026-11-10 15:12:50 +00:00
Said Turksever 5370b8a268 Merge upstream main into pr/street-level-show-action
gevActions.test.mjs: keep the PR's street-level tests and main's analyst display tests.
2026-11-10 14:59:39 +00:00
Said Turksever 9bd2b7ca37 Merge branch 'main' into feat/viewport-context-decision-upstream 2026-11-10 12:14:56 +00:00
Sameh Khamis c8346c9fda Keep shape radii, rim gradients, depth-fail looks, sweep cadence and step camera moves out of the renderer contract
Shapes take their radii in metres; the world annotation ring follows the
camera height itself (ringRadiusAt), resizing on preRender. The radial rim
preset and depthBias go: the launch pad zone draws a canvas image fill made
by the layer. Solid and dashed lines take one depthFail look (solid or
dash) in place of depthFailColor and depthFailDash. sonarCadenceHz goes from
the quality settings: nothing set it, and the sweep moves every 80 ms.

Camera move, look, rotateAround and zoom leave the camera port: the app's pans,
in-place turns, globe turns and zooms are setView views summed over the pose
(geo/eyeMoves), and the Cesium port places a view by direction and up
exactly rather than through heading, pitch and roll. The conformance driver
runs those script steps through setView; the lookAt script turns in place
where it turned about the frame-local axis, and its reference is recorded
again.
2026-11-10 04:08:41 -08:00
German Khokhlov 353eebf224 refactor(tools): normalize ALPR text once, in the records
Every camera find_alpr_cameras returns comes from the tile source, which builds
it through normalizeAlprNode, so the tags are already one line and capped.
Drop the second pass in the tool, its duplicate test and the boundary entry,
and the function check normalizeDisplayText never sees.
2026-11-10 15:37:16 +04:00
Sameh Khamis 7a376ebefc Make quality settings detail only and fold ctrl-wheel zoom into the camera's zoom inputs
Quality no longer switches clouds or water (their ports are gone) and its
atmosphere and post-processing settings are low or high detail; whether the
sky or a pass shows stays with globe.sky and effects. The separate
input.ctrlWheelZoom switch is gone: 'ctrlWheel' in the camera's zoomInputs
also relays trackpad pinches, which the application turns on.
2026-11-10 03:35:35 -08:00