From the test-engineering review of #768: the rendering and panel code had
thin unit coverage (coverage.js 29% of lines, sequences/viewer/selection
~30%, cameraFollow 50%, streetLevelControls only source-text checks), and
two shipped bugs were caught only by the manual browser gate.
- coverage.test.mjs drives the real createCoverage against a fake scene and
a source whose tile requests the test resolves. It reproduces gekh's P0 #2:
a request superseded by retire() finishes after a newer one for the same
tile. Without the ownership checks it drops the newer request's key and a
refresh starts a duplicate whose lines nothing can remove (fails on the old
loadTile, passes now). It also covers the draped-to-cast swap keeping both
line sets on the globe until the cast one is built and clear() removing
both mid-swap, and a tile with nothing castable not being rebuilt.
- coverageFixture.mjs encodes Mapillary-shaped coverage tiles (sequence
lines, overview points) from lon/lat; the tests and the hermetic gate use it.
- cameraFollow.test.mjs pins the underground-camera fix: a -14,886 m sample
is ignored for bare earth, a plausible sample wins, and following stands at
eye height. Three of the four fail on the pre-fix module (the first reports
"centre at -14884 m").
- streetLevelControls.test.mjs is a contract test through the real
DataLayerManager and LayerStateCoordinator: a 360 deg click and a SINCE
release reach the durable layer state and the encoded share link under
Street Level's token. Against a stand-in layer it also covers the chip rule
(darkening the last lit chip turns the layer off, provider kept on; lighting
a dark chip is a user params request, then enable), the legend rebuilding
on a same-count swap (review #9) and the viewer resizing only on open
(review #8). With the old direct-to-layer routing, three of these fail,
including both share-link tests.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Keeps the voice-core prompt rules and adds main's world-query and
vessels-vs-visuals rules; the tool catalog wraps the deixis-aware runner.
Also fixes two things the branch had before the merge:
- Live voice could not start: OpenAI rejects track_entity's top-level
anyOf. Model-facing tools now omit top-level combinators; the runner
and benchmark still require a query or a referent.
- The Contacts-centred analyst answer dropped the viewport coverage note.
The flight, military and vessel feeds were named after the providers the
bundled server uses (/api/opensky, /api/adsblol/mil, /api/ais-live and
their track routes; createOpenSkySource, createAdsbLolSource,
createAisStreamSource). Any server should be able to answer a feed from
its own provider, so the feeds are now:
/api/flights /api/flights/track?icao24=
/api/military /api/military/track?hex=
/api/vessels /api/vessels/track?mmsi=
with createFlightSource, createMilitarySource and createVesselSource. The
old routes are removed. Each server names its provider in X-Feed-Source
(the bundled server sends OpenSky Network, adsb.lol and AISStream, so its
attribution is unchanged) and the military feed's cache state and age in
X-Feed-Cache and X-Feed-Age-Ms. Sources report the provider the server
names and fall back to the feed's name.
The vessel source accepts an optional area and sends it as lat, lon and
radius_km. The vessel layer asks for the area the camera shows (centered
on the middle of the screen, sized by camera height), or every vessel
from high enough up, and asks again when the view moves or zooms by a
quarter of that radius. vessels_in_area names its area too. Servers that
hold every vessel may ignore the area; callers still keep only what is
inside it.
The submarine cable source and the infrastructure data URLs built their
asset URLs from import.meta.url while the module loaded. Importing the
tool services therefore failed in runtimes without a usable module URL,
even for tools that never read those files. Build the URLs when they are
read; Vite still rewrites them in browser builds.
Brings in 123 commits from main, including the v0.2.0 and v0.2.1 releases
(MCP server and tool catalog, server security hardening, build hygiene).
Conflicts:
- package.json: kept main's script order and added this PR's
qa:panel-resize and qa:street-level; exports keep both sides (main's
tools/view entries and this PR's street-level and Mapillary entries).
- src/tooling/viteBuild.test.mjs: kept this PR's Mapillary-token define
assertion and main's new wildcard-host allowedHosts assertions.
- CHANGELOG.md: main now has release sections; this PR's entries go under
[Unreleased].
Checks after the merge: doctor, format:check, check:boundaries,
layer-token:check (28 published, 1 new: street-level 0), npm test (5726
pass), build, the Windows job's focused tests (70 pass), qa:panel-resize,
and the keyless street-level gate. The new CSP allows Mapillary (https:
connect/img, blob workers). The keyed gate stops at coverage locally
because the dev .env Mapillary token is invalid (upstream 403), not because
of the merge.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
loadBundledJson's callers passed a literal JSON import as a Node fallback,
which made the production build emit a JavaScript copy of every pack beside
the JSON the browser fetches. The loader now reads file: URLs itself under
Node (through process.getBuiltinModule, synchronously, as promptly as the
cached import it replaces), and the callers pass only the URL. On a Node too
old to provide getBuiltinModule it says that Node 24.14 or newer is needed.
dist/ drops from 54 MB to 42 MB. A test keeps app code from importing a pack
as a module.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A sweep for the same pattern found more answers presented as current and
complete: launches and satellite orbits served from a stale proxy cache, a
regional aircraft fallback read as worldwide, license plate cameras trimmed
or partly loaded, and a radio directory of popular stations read as every
station. The launch and satellite sources now report the proxies' stale
markers, and each answer says what it covers.
Traffic figures for a radius area counted roads anywhere in its bounding
box, and a tile that failed to load silently changed them. Roads are now cut
to the circle, and the flow source reports per request whether any tile
failed so the answer can say its figures are partial.
The installation source keeps one named site per display cell for the map.
Queries now ask for every site and the exact box, and report when the
source returned only part of the mapped sites.
Reworks #568 to the shape asked for in review.
Generalized from "AQHI" to a provider-backed Air quality layer, with ECCC's
AQHI as the first registered provider. Readings are carried as
{ provider, scale, value, band } rather than a bare number, because
air-quality indices are not interchangeable: AQHI runs 1-10+ on health risk,
the US AQI runs 0-500 on concentration breakpoints, and Europe's EAQI is a
five-step category. Naming the index a value belongs to is what lets another
network be added without inventing a shared numeric meaning. Coverage is
stated as "Canada only" on the provider, the layer row and the card.
Adds the WEATHER panel readout (readout: true) carrying the band legend and
the worst readings currently in view, with getRowControls/setRowControlsListener
following the wind and cyclone cards. An unresolvable view rectangle degrades
to the worst readings overall rather than to a blank card, and a rectangle
crossing the antimeridian is handled.
update() now takes the earthquakes layer's abort-and-guard shape: it aborts any
in-flight request, passes the signal through to the source, and refuses to
touch the scene when the request was superseded, aborted, or the layer was
disabled while it was outstanding. disable() and destroy() abort too. Tests
cover the late-response, superseded-refresh and disabled-update paths.
Rebased on main: the source is registered in src/sources/reference.js, and the
allocation profile was re-measured against the current worker — median 3,900
B/frame, max 4,114, 85.7 B/candidate, inside the shared 154 ceiling, budgeted
at 5,200 for 26% headroom. The earlier figures no longer held.
Freshness is now judged on the observation time rather than on when the last
response arrived. lastUpdate only says the feed answered; a station that keeps
returning a reading from four hours ago would otherwise be narrated as live by
the HUD and the voice provenance envelope. getStats() reports observedAt and
stale, so layerFeedState() calls the layer stale past two missed hourly
publications, and degraded when a refresh fails over readings that are still
good.
The share-link token is 0, which is what layer-token:next allocates. Taking it
moves the ledger fixtures that had hard-coded 0 as the free digit and 3 as the
one after; those shift to 3 and 4. The pinned equality between the reservation
ledger and LEGACY_LAYER_STATE_TOKENS also had to become "the closed legacy set
plus air-quality", since that set is documented as never growing. Happy to
re-pin this differently if you would rather the fixtures derive the provisional
digits from nextLayerStateToken() so the next layer does not move them again.
The layer and its source are registered as ./layers/air-quality and
./layers/air-quality/source, each with its own boundary group, following the
perimeters pair. check:boundaries requires the package exports and the group
exports to be the same set, so a layer that ships a source needs both.
P0
- tileMath: lonToTileX(180) wrapped to column 0, so a whole-world view at
z1 loaded only the western column; 180 is now the last column.
- coverage.loadTile: a superseded request (after a retire or clear) still
deleted the newer request's pending entry and loading count in `finally`,
letting a duplicate load attach lines that nothing could remove. Only the
request that still owns the key attaches its tile or settles the key.
P1
- view/tileMath: a view across the date line read as a ~358 deg box and
loaded tiles on the far side of the globe. visibleBbox now returns
west > east for such views and tilesForBbox splits them at +/-180;
viewCentre averages across the line.
- decode: the parts of a multi-line sequence were joined into one line,
drawing straight segments over capture gaps. Each part is now its own
line (`parts`); later parts carry `~n` in their pick id, and picking or
recolouring any part acts on the whole sequence.
- viewerHost: a prewarmed MapillaryJS viewer was never destroyed, leaking
its WebGL context. Prewarmed adapters are tracked and released on
unmount, at once when the layer went off mid-load, and per provider when
only that provider is switched off.
- nearestImage took the first record the API returned; it now takes the
closest one that passes the filter.
P2
- groundCast: the cache was cleared after a request counted its corners,
dropping ones it relied on; it is now cleared first.
- streetLevelControls: the viewer was resized on every render (every frame
while a panorama is dragged); the ResizeObserver handles size changes and
the viewer is resized once when it opens. The legend now rebuilds when its
entries change, not only their count.
- panelPositionControls: the portable-panel paths this PR added called the
CCTV viewport sync directly; they go through onPanelResized, which the
shell routes to CCTV. The "Compose camera panel controls" JSDoc is back on
_initCctvPanel.
New tests: zoom-1 columns, date-line tiles and bbox, multi-part decode and
part pick ids, nearest image, prewarmed viewer release, and the cache
ordering (fails on the old order).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
keep contacts counts, rows, scope and provenance on one retained snapshot
report capped cohorts as lower bounds
separate sibling tool cancellation from newer user turns
align track_entity validation across runtime and benchmark paths
The raycast challenge on #768 showed bare-earth casting misplacing roads
that are not at grade: lines floated over freeways in trenches and steep
streets (rendered line pixels floating: Austin 8.5%, SF 27.1%, LA 43.8%),
because a ~111 m terrain grid cannot see a 30 m-wide cut or a slope between
its corners.
meshSampler.js samples the rendered Google 3D surface (scene.sampleHeight,
overlays excluded) once per ~11 m cell within 900 m of the camera, nearest
first, in idle time within a 6 ms budget; misses are retried, not latched.
groundCast.refineHeights blends it in, per point along a line:
- mesh up to 3 m above bare earth is the road: follow it, 1 m above;
- higher mesh (canopy, deck, roof): carry the road's offset from bare earth
across from the road samples on both sides (up to 8 points);
- otherwise bare earth + 2 m, as before.
Lines, cones and the marker use it; tiles touched by new samples are
redrawn one per idle slice with the no-blink swap.
Rendered line pixels on the surface / floating, before -> after:
Austin 90.0 -> 95.3% / 8.5 -> 0.7%; SF 69.0 -> 86.4% / 27.1 -> 9.1%;
LA 52.3 -> 91.0% / 43.8 -> 4.0%; Innsbruck 95.0 -> 95.7% / 0.2 -> 0.2%;
Sacramento 80.7 -> 79.3% / 3.8 -> 2.7%; buried lines stay under 1%.
Cones floating in SF 44 -> 5%, buried at LA 59 -> 28%. Main-thread long
tasks added during refinement: at most 134 ms (Austin, densest tile).
Roads on elevated decks are still drawn at ground level under the deck.
Unit tests cover the blend (trench, slope, canopy bridge, deck, bad probe),
the finer cast with meshAt, and the sampler (once per cell, nearest first,
overlays excluded, far cells skipped, misses not latched, idle when off).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
From the code review of #768:
- view.js measured the camera's height above ground with globe.getHeight,
but Google 3D hides the globe, so coverage picked its zoom from the
ellipsoidal height. Over a high city (Denver, ~1,600 m) a camera 500 m up
read 2,100 m: coverage stayed at z12/z13, whose tiles are too big to cast,
and lines stayed draped on the roofs. With the globe hidden, the bare-earth
height from the ground caster now answers; coverage re-picks its zoom when
the surface mode changes. Verified in Denver: terrain mode, z14, cast lines.
The core's surface check uses the same helper (and no longer subtracts the
2 m lift from the camera height).
- castTile rebuilt a tile even when no line could be cast (terrain proxy
down, tile too big); it now leaves such tiles as they are.
- The feature gate read Street Level's share token as a literal `0`; it now
takes it from LAYER_STATE_REGISTRY, since a rebase can reassign it.
New view.test.mjs covers the shown-globe, hidden-globe and no-sample cases.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Add getImage() to the weather source. It reads one frame through the
bounded image route the layer already uses and returns its content type
and bytes, so code outside the layer can show the current map.
layerState.js held both the share-link layer codec and the coordinator
that restores layer choices into the running app and persists them to
browser storage. Move the coordinator, unchanged, to
layerStateCoordinator.js so the codec stays a pure data module that code
outside the browser can use to write share links. The registry stays in
layerState.js, where the token tooling reads it, and the aircraft
barrel re-exports both.
Add getFeeds() to the transit source. It reads the public catalog route
the server already serves, so code outside the layer can find feeds
near a location without loading the layer.
The bike-share layer parsed station documents inside its model factory.
Move the parsers unchanged to src/sources/gbfsStations.js and have the
model import them, so code outside the layer can read GBFS stations
without constructing the layer.
Add getFrame(camera) to the CCTV source. It reads one frame through the
registered frame endpoint the layer already uses and returns its content
type and bytes, so code outside the layer can show a camera's view.
The satellites layer parsed catalog text inside its Cesium-backed orbit
owner. Move the parser unchanged to src/sources/tle.js, add a catalog
number helper, and have the layer delegate to it so code outside the
layer can read TLE text without loading rendering.
Route every application-owned presentation string through the i18n
core: t() call sites across the app/layers/ui/data/voice surfaces,
data-i18n* attributes in the src/ui/templates/*.html markup (applied at
boot by applyDocumentTranslations from src/standalone/application.js),
and the dock locale selector (runtime-rendered from the offered pair —
a single EN button while only English ships).
Rebased onto current main, where the extraction follows the module
splits: the shell wiring now lands in the owner modules
(panelChrome, visualSettings, locationNavigation, shareRestoration,
shellFeedback, applicationShell) and the voice status corpus in the
realtime* modules; scene status lives in director.js +
playbackAdapter.js.
Modules shared with the node runtime (loadingFeedback,
installationFeedback, regionalModel, keySetupCore, vessels ingestion)
stay environment-neutral English and localize at their browser render
sites (chip maps in shellFeedback/installations controls, the weather
table in cockpitBriefing, keySetupRequirementEnvVars compositions), so
the package-boundary graph stays clean; the browser boundary groups
now own the i18n modules they reach (scripts/package-boundaries.json).
- Voice layer manifest (src/voice/layerManifest.js) generates the layer enums,
aliases and analyst field hints; a test fails when a catalog layer has no entry.
- analyst_query validates fields, operators, units, values and scopes, counts up
to 250,000 loaded records per layer, reports complete/partial/unanswered, tells
LAYER_OFF / NOT_READY / FEED_UNAVAILABLE apart from zero, refuses mismatched
follow-ups and never lets a superseded query replace follow-up memory. Seven
more layers expose analyst records.
- Tool results carry a code-built say line and a display card; the voice card in
the mic control shows captions, plan steps and the result
(src/voice/resultDisplay.js is the one adapter). Slow tools narrate their
current step out of band; holding Space while the assistant talks cuts it off.
Turn spans go to the debug log; caption transcription is metered; the log
omits what was said unless GEV_VOICE_LOG_CONTENT=1.
- Point and ask: the Space-down pointer snapshot resolves "this", "that one" and
"around here" through 'pointer' targets, and "the second one" through
referent:n from the last numbered list.
- Prompt states each policy once (instructions + tools 48.5k -> 46.7k chars).
- Voice benchmark harness (scripts/qa-voice-bench.mjs, scripts/voice-bench/)
shares its phrase table and grading with scripts/qa-voice-routing.mjs.
Traffic roads come from OpenFreeMap vector tiles, or TomTom flow tiles with a
key (Hybrid by default). Military areas use OpenFreeMap polygons with a bundled
worldwide name index; wide views show named points. ALPR reads an hourly
OpenStreetMap extract. Cockpit place context and country, state and US county
outlines use bundled Natural Earth and US Census data. Overpass is used only
when an operator configures their own endpoint.
Also: camera arrivals no longer land below ground, keyless terrain retries on
429, bounded relation geometry for configured Overpass, one shared
OpenStreetMap credit, and separate tile and API transports for embedders.
Co-authored-by: theGaryLarson <[email protected]>
Co-authored-by: Milan Khanal <[email protected]>
As more sources arrive, each one reads the same everywhere it appears: its
PROVIDERS chip, coverage lines, overview points, image cones and a single
legend swatch. The palette lives in one place, PROVIDER_COLORS in
src/layers/streetLevel/policy.js: Mapillary green (#05cb63), Panoramax
purple, Google Street View blue.
Mapillary drops its age dimming and magenta panoramas: lines and overview
points are Mapillary green, cones are green too and 360° images keep their
ring shape; the 360°/FLAT filter still separates them, and the selected
sequence stays GEV cyan.
The provider contract loses `legend`: `colors.coverage` is the source's one
colour, and the core composes the legend as one swatch per active source
(its name, its colour) followed by Selected. The chip group takes an
optional per-chip colour as a `--chip-color` CSS variable, written only
when it changes; Street Level chips show it as a swatch dot and, while on,
as the chip's border, text and a faint fill.
The feature gate checks the legend has one swatch per source plus Selected
and that the Mapillary chip carries #05cb63.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Opening a photo frames it from 140 m behind using scene.sampleHeight for
the ground. Before the Google 3D tiles under the photo have loaded, that
sample can be kilometres wrong: an Austin 360 photo sampled -14,886 m and
the camera flew to -14,811 m. A sample now counts only when it is a
plausible surface height and, once the bare-earth height is known, lies
within -15/+80 m of it; otherwise the bare-earth height is used. FOLLOW's
eye height goes through the same check.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Draped ground polylines and clamped billboards land on the top of the
Google Photorealistic 3D mesh, so on a street under trees or an overpass
the coverage seemed to float at roof height.
At street zoom on Google 3D (camera below 1,400 m, back above 1,800 m)
the core switches to a `terrain` surface mode. groundCast.js samples the
existing terrain service on the ground-floor grid (0.001 deg), interpolates
bilinearly and adds 2 m; Mapillary lines become plain polylines at those
heights, and cones and the position marker stand on the street, so the
mesh hides what is behind buildings and trees. A tile draws draped first
and swaps to its cast lines once the heights arrive, keeping the old lines
until the new primitives are ready. Lookups are once per tile and aborted
when the tile is dropped; a geoid fallback counts as unknown and leaves
lines draped. Other map stacks and higher views are unchanged.
Providers get getSurface() and groundCaster in their context and an
optional setSurface(mode); getUIState() exposes `surface`. The QA gate
checks draped on Esri and terrain on Google 3D.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
`mapAnalystRecord` has always published `verticalRateMps`, but the tracked
seam — `getTrackedInfo()` / `getTrackedSubject()`, built by `_describeFlight`
— did not, so anything needing climb/descent had to reach past the public
descriptor into the raw poll record.
Both layers now map it, which matters because they store it under different
keys: the civil records write `verticalRate` (records.js), the military ones
write `verticalRateMps`. Mapping only one would leave
`resolveTrackedAircraftInfo()` handing back two different shapes, and a
consumer reading the field off a civil contact would silently read every
military contact as level flight.
Reported, never derived. `stickyNumber` seeds the field with null, so an
aircraft that has not transmitted a rate publishes null rather than a
level-flight 0 — absent is not level, and a flight-phase classifier that
cannot tell them apart calls a climbing aircraft cruise.
Tests: both layers, both halves (a reported rate arrives with its sign; an
absent one stays null). Mutation-verified — removing either field or
defaulting it to 0 turns them red.
`docs/CURRENT-STATE.md` records the descriptor field, the two record keys it
reconciles, and the reported-never-derived rule.
A PTZ/varifocal datasheet FOV is a capability range, not the lens's
current zoom, so replacing the frustum's estimate with its wide end
presented unknown optical state as known. Parse the full horizontal
range (ignoring vertical, "combined" panoramic and lens-note figures),
carry it as specFovMinDeg/specFovMaxDeg, and substitute the datasheet
FOV only for fixed single-lens hardware. The HUD labels the range as
capability beside the estimate. This also stops the Q6000-E Mk II's
"360 combined" figure from being read as a 360-degree lens.
The loader and normalizer carried the hardware model, but the /sources
response projection and the browser catalog both whitelist fields and
dropped it, so no client consumer ever saw it. Project model and
manufacturer (manufacturer disambiguates bare model strings like "3950")
at both boundaries, and add a test that drives the real middleware and
browser catalog end to end.
FOLLOW puts the globe camera at eye height inside the street-level view; on
flat imagery or terrain there is nothing to stand among. The layer now takes
the application map stack controller (attachMapStackController), subscribes
to stack switches, offers FOLLOW only while the stack is 'photoreal', disables
the button with a MAP SOURCE hint elsewhere, and stops following when the
stack leaves Google 3D. Selection uninstall is a no-op when never installed.
The QA gate checks the disabled path and, when Google 3D is available, that
leaving it stops following; its caption-timeout dump now forces one render to
tell a missed update from a render fault.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- Only a user's own collapse docks a floating window; cockpit entry,
restores and the Cyber accordion no longer dock it or delete its saved
place, and the accordion leaves floating windows alone.
- The viewer host records an adapter as active only after its mount
succeeds, so a library that failed to load once is retried; concurrent
opens share one mount, and unmounting mid-mount leaves nothing active.
- openNearest tries every active provider; one provider's error (no key,
offline) no longer hides the others' imagery.
- The SINCE readout shows its cut-off date as soon as the committed value
matches the thumb, even while the slider keeps focus.
- A floating window that was only dragged is bounded to the viewport and
scrolls instead of running off-screen.
- getStats summarises coverage (summarizeCoverage) instead of building a
full UI snapshot every second; disable() uses credits.hideAll(); smaller
cleanups in the viewer host and sequence snapshot.
- Unit tests: panel docking and header double press, viewer-host mount
retry/sharing/cancel, coverage summary.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Street Level is now a provider-neutral layer modelled on the iD editor's
photo overlay. The core owns what every imagery provider shares: the
enable state and a switch per provider, the imagery filter (360°/flat and
captured-since in relative days), one click handler routed by pick prefix,
one viewer host that mounts whichever provider's viewer adapter owns the
open image, the position marker, camera follow, one on-globe credit per
active provider, and share-link parameters (one bit per provider plus the
filter). Providers implement the contract in registry.js: coverage,
sequences, nearest-image lookup and a viewer adapter emitting a common
pose. Mapillary is the only registered provider; Google Street View,
KartaView and Panoramax slot in beside it.
Panel: a PROVIDERS chip row (one chip per registered provider, keyless
chips read as errors), relative-day SINCE options, a provider-neutral
caption whose link label follows the provider, and no vendor mark in the
header. The panel inner is a rail scroller like CCTV's.
Tests cover the registry, params codec, filter, pick router, UI-state
composition, the provider definition and coverage-tile decoding; the QA
harness switches the provider off and on and checks the filter.
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Pure move: the layer id, panel id, CSS prefix, classes, exports and config
paths become provider-neutral (street-level, #street-level-panel, sl-);
the Mapillary-specific modules move under providers/mapillary/. Behaviour
is unchanged. The two sprite collections now register under distinct ids
instead of overwriting each other under 'mapillary'.
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Street Level is imagery only from here on: coverage, sequences and the
embedded viewer. The Claude planner, map-feature and traffic-sign tiles,
the NDJSON feature stream, sprites, detection outlines and analyst records
go, along with @anthropic-ai/sdk, zod and the Anthropic key entry.
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Upstream took share token '2' for fire perimeters, so Street Level moves to
'3'. mapillary-js joins upstream's optimizeDeps include list instead of
overriding it; catalog and registry tests count the new upstream layers.
Co-Authored-By: Claude Fable 5.1 <[email protected]>
The panel no longer shows a separate keyless call-to-action; the fieldset
gate and the layer-row tooltip already cover the missing-key case. The
planner flag is tri-state (null until status answers).
Co-Authored-By: Claude Fable 5.1 <[email protected]>
Some catalogs publish a full product title rather than the bare model, e.g.
"AXIS Q6155-E PTZ Dome Network Camera" — which the exact model/alias match
misses. Add a fallback: after the exact match, scan for a record whose
DISTINCTIVE token (a >=4-char model/alias word containing a digit or hyphen,
so "dome"/"network"/"ptz"/"hd" never qualify) appears as a whole word in the
query. The most specific (longest) matched token wins; if two records tie on
the longest matched token the result is ambiguous and returns null, so the
lookup never guesses. Exact matches are unchanged and still take precedence.
Grow the enrichment table from the 4-model starter to the hardware the
currently integrated public catalogs actually publish, so identified
cameras render datasheet FOV cones and spec cards instead of estimates:
- King County (WA) road cameras: Cohu 3950 / 3960 (3965) / 3920 iDome
(3925), Bosch AUTODOME IP 5000i.
- Sarasota County (FL) + Sioux Falls (SD) open data: Bosch AUTODOME IP
starlight 7000i / starlight 7000 HD / dynamic 7000 HD, the VG4 AUTODOME
(500i), Bosch AUTODOME 800 (VG5-836), and Axis Q6155-E, Q6054-E,
Q6000-E Mk II, and P3707-PE.
All records are vendored verbatim from the CCTV Camera Database (CC0),
each transcribed from a manufacturer datasheet. Analog Cohu positioners
carry no fov_deg (their datasheets publish none), so they enrich the HUD
specs without a datasheet cone; every other model carries a real FOV.
Adds a test asserting the vendored table covers each integrated catalog's
published hardware strings.
When a CCTV source pack (or a seed) names its hardware `model`, resolve it
against a camera-model spec table: the coverage frustum then uses the
manufacturer's horizontal FOV instead of the pose estimate, and the HUD meta
line gains the model's resolution, night-vision, and PTZ specs. Cameras with
no identified model are unchanged — unknown never implies anything, matching
the layer's existing honesty about RAW PRIOR poses.
Spec records are vendored verbatim from the CCTV Camera Database
(https://github.com/ch-bas/cctv-camera-database, CC0-1.0), an open dataset of
12,000+ camera models transcribed from manufacturer datasheets. This lands a
small starter table plus the join + parsing helpers; source packs that publish
their hardware can grow it. Records describe camera products (optics,
resolution, night vision) only — no endpoints or access data.
- src/layers/cctv/modelSpecs.js: spec table, case-insensitive/alias lookup,
wide-end FOV parser, HUD summary formatter
- src/layers/cctv/modelSpecs.test.mjs: lookup/alias/unknown, FOV parsing
across the real fov string shapes, summary rendering
- src/layers/cctv/catalog.js: apply enrichment in both catalog builders
- src/layers/cctv/presentation.js: label a datasheet FOV, append the spec
summary, and pass fovSource + spec through the public camera state
502 (every mirror unreachable) and 503 (the local limiter, before any mirror
is asked) come from our proxy, so they no longer read as a mirror refusing
the query. Credit the fix in CHANGELOG and note the mapping in CURRENT-STATE.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
