skyBackdrop/BACKDROPS and the tests that held login text against sky bounds are gone (the bounds were false: the look is tuned by eye). The vertex-carry guard asserts placement only. d3-celestial's BSD-3 text is in THIRD_PARTY_NOTICES.md.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
About 31 degrees of sky in a 16:9 cell. The vertex-star guard drops to 0.8:
low vertices sit in the moonlit haze at this framing.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Under a full moon at the new framing every constellation-line vertex in the sky strip now carries a drawn star (was 24 of 34); a test pins it.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
windowFor places the plane so its horizon row lands at HORIZON_AT of the height, the same for every sky state and viewport; the plane runs below the viewport bottom (negative bottom) and the near plain is cropped. Every consumer already read the camera and plane from windowFor; the sheet's card placement now reads plane.bottom too, and takes ?h= for a phone-shaped cell. Tests pin the constant horizon, ~35 degrees of sky in a 16:9 cell, and plane coverage with baked ground to spare at 1024x768, 1920x1080, 375x812.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The night set renders new, first-quarter and full moon at four hours each plus deep dusk. Night exposure keys the sky to where the curve wants it so the per-channel curve does not turn the scotopic blue shift into a saturated sky.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
pickSkyState (the menu's pick): Night while the choice is Night toggles the lines, stored in ratatoskr.skyLines (storage wrapped), told to listeners as a sky change; the rebuild key carries ':lines' at night so the scene is rebuilt with the option. Sky menu Data section credits BSC5 and d3-celestial (BSD-3-Clause).
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
nightBodies.ts places the catalogue stars for the sky's own moment, adds their radiance (extinction, B-V colour, point spread) to the dome behind them and takes the sky's display, so twilight hides them; the moon is drawn with its terminator and earthshine; lines over the stars when asked. stars.ts keeps the twinkle schedule, now a gain on a star's flux. Ridge.ts bakes the bodies and takes a lines option.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
nightSky.ts holds the night's photometry (moon magnitude by phase, Krisciunas & Schaefer; airglow; star radiance; colour from B-V). The moon gets its own atmosphere bake for the night state and lights the dome, the clouds, the haze lookup and the cards (replacing nightAmbient and the cel bands' sun at night). Night loses its fixed knots and 0.043 ceiling: it adapts like day with the targets scaled by partial adaptation, a scotopic grade (Jensen 2001 luminance, blue shift) before exposure in CPU and shader, NIGHT_CEILING tunable (1). The ink-ceiling test for night is dropped with the ceiling.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
2887 stars (Vmag <= 5.5) and 150 line chains, 21 KB packed. Placement precesses J2000 to the date and uses sun.js equatorialToHorizontal; tests hold five stars and the moon to astropy's altitude/azimuth within half a degree.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
sun.js exports daysSinceJ2000, gmstRad, sunEquatorial and equatorialToHorizontal (sunPosition is now their composition, unchanged in value). moon.js adds the Astronomical Almanac low-precision moon, phase from elongation (Meeus 48.3), topocentric position. Tests against Meeus ex. 47.a and USNO phases.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
RidgeLook gains celBands/celSoftness/celBandLevels and a coarse ambient step; DEFAULT_LOOK is painterly with cel 3, PHYSICAL_LOOK keeps the old Lambert look. Sheet presets: physical, swiss, poster, painterly, painterly-cel-2/3/4.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Per-sky monotone cubic through five knots (black lift, p5/p50/p95 of the sky grid to target luminances, white or the night ceiling), shared by sky and cards. Night gets a fixed-knot curve and a RidgeLook.nightAmbient term.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
reliefWeight, reliefLight, hazeTint, nearWarm, contrastByDistance,
cardToneStep, cardGradient, inkScale, groundNoise, cloudShadows; all default to
today's look. makePhysicalSky takes an optional sun for the look sheet.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The look constants (ink geometry, ink keeps, sun cut-off) become RidgeLook with
DEFAULT_LOOK; lightCards and lightingOf take an optional look. hazeScale and
ambientScale are new axes (default 1 = today). A recording-GL test proves an
override reaches the shader uniforms.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The face-up albedo mutant still survives (no fixture has a wall under a face steep enough to move the
albedo: not yet reached); the card-0 wall mutant is killed by the near-hill fixture.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
A near hill makes every row of card 0 lie below its first sample (so a flagged card 0 would show);
a ramp under card 1's first sample makes its face steep (so albedo from the face normal would show).
The geometry channel-swap mutant follows the encoder's new spelling.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
startRig puts one element per lit card, farthest first, at its rows with transform-origin on the
horizon centre, animated by +-scale_i x heading with scale(scale_i); the still draws the same cards at
heading 0. index.ts lights in prepare, so a sky-change rebuild lights again and any failure (no WebGL2
included) is planesFail and the sky alone. ridgeAssets.test gates the committed v2 bake.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
lightCards.ts: one program on its own canvas, per card sun and ambient in the column frame,
aerial perspective from the geometry distance and the atmosphere's extinction, contour ink,
the sky's own display; wall code is the face up. Context released after the last card.
Backdrop.ts fetches and decodes every card straight and lights once; any failure is null.
toneMap exports its ACES constants and earthAtmosphere its ground extinction, so the shader copies no number.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Heading each side is the baker's range capped so every card still covers the viewport;
card i pans scale_i x heading about the horizon centre. Tests on phone, tablet, desktop
and ultrawide viewports and for the cap.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Three cards (splits 6, 40 km), camera r +200 m, heading range 23 deg, per-card scale
1.156 / 1.018 / 1.004. The six-key sun-frame bake vienna-12e733dc9f09 is removed.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Cards i > 0 flag the extruded face under their first sample as wall (R = G = 255 in the
geometry file, albedo with the face up); the manifest gains camera {r_m, heading_max_deg} and
per-card distance_m (non-wall median) and scale = d / (d - r). The baker refuses |r| >= half the
nearest distance and any camera that uncovers the frame. Tests and mutants for each rule.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
frameChoice, frameLuma, the two-frame Backdrop, the v1 ridgeAssets gate, DRIFT_DEPTH,
the plane rig and paintPlanes are gone; parseManifest reads schema 2 (K cards, encoding,
camera.heading_max_deg, pxPerDegV). Until the lit-card rig lands the scene is the sky alone.
The plane/index tests of the deleted behaviour are removed here and return rewritten.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
The slide contract (viewport-aspect drift, simultaneous slides) lives in the runtime cut; a baker
bound of 5% of the card width was unsound and did not lower the decode total. cards() is again one
rectangle of rows with coverage; the schema is regenerated and committed.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Side-by-side nearer ridges, overlap at the edge of their union and a ridge across the window each
need a different term of the cut (single-card slide, slide width, frame edge). The frame-edge mutant
now treats what lies beyond the frame as held (the old one was equivalent).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The slide check shifted the albedo along rows, so erosion 0, single-card and frame-edge mutants
survived; it now shifts the last axis. A second test sets the seam prior to 0.1 on the 4 to 15 km
ridge, where only the swept-rows term prevents the hop (at 0.02 the mutant without it cannot tear).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Decode the written WebPs against the cards (channels, range, straight alpha); the six receding-ridge
fixtures with the 0.35 alpha-jump metric; no slide within the bound reveals a cut pixel; empty card;
SEAM_SEP, elev curvature, coverage rows, interpolation and renormalisation. mutate.py grows to 33
mutants. README describes the reveal bound.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
cards() cuts from each card the pixels the nearer cards hide at every offset within MAX_SLIDE_PX
(ceil(5% of the card width), the runtime's DRIFT_FRAC), also when one card slides alone; the composite
at rest is unchanged. Manifest v2 carries max_slide_px. pack() refuses an empty card naming its splits.
coverage() no longer counts the rows below the first sample twice. SEAM_PRIOR 0.02.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The prior is the value where the side-by-side test passes with the swept-rows term and fails without it
(0.05 and below let a seam without the term follow the ridge, 0.15 and above hop across it with the term).
mutate.py holds 16 mutants of the card rules, the cache keys and the export checks, every one killed.
README rewritten for cards, manifest v2, the seam rule and the low-pass.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
ridge.py: elev (shared by march and prefilter), prefilter, normals, set_view from the march skyline,
coverage, seams (one forward-energy DP per authored split), cards (first visible hit at or beyond the
near seam, crest coverage, cross-fade over seam +- XF, nested alphas so the composite equals the opaque
coverage, cropped to the row band). load_preset requires card_split_km (strictly increasing, at most
MAX_CARDS - 1). export.py: manifest v2, two WebPs per card, check_cover, check_budget on count, total
bytes and decoded bytes, sheet. places.json: Vienna card_split_km [6, 40], prefilter_m 100.
The seam prior is 0.05 rows/nat^2 (tie-break only): at 2.0 the seam hopped across a receding ridge
between two columns, which the spec's own side-by-side test forbids.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Cut 1 of ridge-gbuffer-baker (rulings ridge-bake-gbuffer-depth-bands, ridge-cards-runtime-lit,
ridge-cast-shadows-dropped-for-now). Deletes first: extract, chains, sweep, slots, tapers, the painter,
the relight family, shadows, sun.py, sun_set, luma_range, check_near_cover, the painter inputs in
places.json, and every test and mutant that exercised them. export.py is left unrunnable until the
card writer replaces write_bake; the ray caches stay valid (march, rays_key, stage0 unchanged).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>