The histogram search over a free/blocked cell grid never depended on the
obstacles being boxes. Move it to OpenNest.Core as MaximalRectangles.FromGrid
so other callers can build their own grids; RemnantFinder keeps building its
obstacle grid and calls it. No behavior change.
Parts drawn a few millionths over their sheet's work area (for example a
36.125006 panel on a 36.125 sheet with no edge spacing) pass NestLayoutCheck,
which allows Tolerance.Epsilon of overhang, but the Rectangles engine refused
them and left them unplaced. The engine now allows 90% of that slack: a box
that exceeds the sheet by no more than the allowance packs as exactly the
sheet's size, so it spans the axis and the overhang lands only past the edge.
The check's bounds slack is named NestTolerances.WorkAreaSlack (same value,
no behavior change) so the engine and the test layout assertion share it.
Rectangle-lane benchmark (91 strict + 77 boxable jobs): every layout valid;
4 jobs that left panels unplaced now place them (strict complete 86 -> 88,
boxable complete 74 -> 75, one more boxable job places 2 more parts); no other
job changed.
DefaultPlateFiller.PackArea (Default, Strip and both Remnant strategies)
now packs leftover parts with the shared maximal-rectangles packer instead
of bottom-left corner points. Every fit rule and pick mode is tried; the
layout placing the most parts, then box area, per priority tier wins, and
the strategy's own fill comparer breaks ties so remnant strategies keep
their clear side. PackBottomLeft and PackEngine are removed.
Rectangle-lane benchmark (91 strict + 77 boxable jobs) against the old
packer: no strategy lost a valid layout; Default, Vertical Remnant,
Horizontal Remnant and Strip complete 2-3 more strict jobs, and Strip
gains a valid one. Cost on jobs complete in both runs falls for every
strategy (Default -0.6% strict, -2.7% boxable). Default is about 7-10%
slower on the changed jobs.
Golden layouts for Default and both Remnant strategies are re-captured;
each is complete, passes NestLayoutCheck and repeats exactly.
Move the Rectangles engine's global and ordered pick loops into
RectanglePacking.MaxRectsPacker, which works on plain box sizes so other
fill paths can reuse it. SheetPacker maps part orientations onto it.
No behavior change: rectangle-lane benchmark results are identical job for
job (91 strict and 77 boxable jobs).
MaxRectsSheet only tracks free rectangles and has no job dependencies, so it
moves from the Rectangles engine into OpenNest.Engine.RectanglePacking where
the interactive fill packer can use it. No behavior change.
Desktop Auto Nest, Console --autonest, MCP autonest_plate and the API
NestRunner all run through NestPipeline, so nothing calls the old
orchestration any more.
Delete MultiPlateNester (with MultiPlateNestOptions, MultiPlateResult,
PlateResult, PartClass and PartSortOrder), PlateOptimizer and
PlateOptimizerResult, plus their tests. The explicit-strategy contract
those tests checked now lives at PlateFillService.ResolveStrategy, which
keeps its null-means-Default, canonical-name and unknown-name tests.
CreateFiller loses its internal visibility, which only the deleted
orchestrators used.
This breaks source and binary compatibility for external callers of the
removed types; whole-job callers use NestPipeline.Run or INestingEngine.
Skip FillGrid's Step 2 overlap check only when Step 1 checked this exact
row clean and the perpendicular tiling appended zero parts, so gridResult
still holds the same Part objects in the same order and poses. The local
rowIsVerified flag is cleared by every Step 1 skip or bbox fallback, and
any nonzero append (including partial-copy parts) keeps the eager check.
Debug work assertions pin the reduced overlap work (horizontal stripe
8->4 exact calls, 4->2 preparations, 16->8 triangulations; vertical
stripe 36->18, 4->2, 38->19) and keep single-seed, Horizontal full-grid,
partial-only and PerpOnly work at base; the invalid overlapping-seed
control pins both fallback stages plus the retained exact-call total.
Three mutants (forced eager check, dropped flag, complete-rows-only)
are caught by their designated controls. Serial whole-job layout is
byte-identical to the frozen 62b5a8d0 oracle before and after.
Moves the two production plug-in engines into OpenNest.Engine under names
that describe the jobs they suit:
- Rectangles: plain and near-rectangular plates, maximal-rectangles box
packing (was the RectanglesNestingEngine plug-in)
- Irregular: irregular profiles, no-fit-polygon frontier packing (was the
Opus55NestingEngine plug-in)
Their tests and the shared engine contract/layout test kit move into
OpenNest.Engine.Tests/NestingEngines.
The registry maps the old plug-in names to the new engines, so saved desktop
selections, scripts and API requests keep working, and a leftover plug-in DLL
under an old name cannot shadow its replacement. Desktop startup passes the
registry's lookup when restoring the saved Auto Nest engine.
build NestJob -> resolve engine by name -> Solve -> independent
NestLayoutCheck validation -> bind placements to caller drawings.
The pipeline never commits to plates and never mutates caller items;
validation failures are returned as messages naming real drawings
(unknown-requirement placements are reported, not dropped). Console,
MCP, API and desktop Auto Nest will adopt this path in later phases.
- NestPipeline.Run(request): registry-resolved engine, unknown names
list the registered engines; cancellation propagates untouched.
- NestPipeline.Run(engine, ...): stub/plug-in engines take the same path.
- NestResultBinder: pose semantics identical to NestResultMaterializer.
- NestLayoutCheck: Violations overload with per-requirement display names.
Tests: overlap stub -> violations w/ drawing names, no throw; ghost
placement -> violation + excluded from further checks; unknown engine
-> NotSupportedException listing engines; cancelled token -> no result;
happy path -> bound by reference, caller quantity untouched.
Grows part-to-part spacing of a selected group with the work area and
non-selected parts as hard boundaries. Doubling + bisection search over
the target spacing; Gauss-Seidel straight-line relaxation with anchor
mover policy (later-index selected part moves, first selection never).
Overlapping input is separated along penetration MTVs instead of being
rejected. Cancel/failure never mutates part positions. Clearance gains
BoundaryDistance for ring-pair gaps (part-in-cutout legality).
Both HasOverlappingParts loops rebuilt each part's polygon from its
Program on every pair. PartOverlapChecker prepares each distinct Program
(reference identity) once and each part's world polygon once per call,
then uses the overlap-only Collision.HasOverlap. Loop order, bounding-box
prefilter, early exit and returned indices are unchanged; Part.Intersects
shares the material/polygon recipe and still returns crossing points.
Verification:
- Frozen LegacyPartOverlap differential (original Intersects and both
loops): verdicts, indices and world polygons bit-identical across fill
grids, patterns, touching/epsilon gaps, scribe/rapid/empty programs.
- Debug OverlapPolygonPreparations: 246 -> 1 and 64 -> 2 per check.
- Corpus job (169 parts, --engines Default --parallel 1, with 1a):
median 18,885 -> 13,464 ms over 4+4 alternating runs, identical
outcomes; serialized layout byte-identical to the base.
FillLinear re-prepared offset perimeter geometry (ConvertProgram ->
ShapeProfile -> OffsetOutward) for every part it measured, although
tiled copies share one Program and differ only by Location. A CPU
profile of a 169-part Default job put 62% of wall time there.
Prepare each distinct Program (reference identity) once per public
Fill/FillRow call in local frame, then clone and translate for each
location. The cache is created per call and passed down privately
because FillHelpers.FillPattern calls Fill concurrently on one
instance. PartGeometry gains a local-frame Program overload that the
Part overload now delegates to.
Evaluation order, lazy preparation, fallbacks and tiling are
unchanged. Differential tests against a frozen copy of the previous
FillLinear check bitwise equality, including concurrent calls; Debug
work tests pin preparation counts. With the thread pool capped at one
worker, before/after whole-job layouts are byte-identical. The
Default corpus job median drops from 40,715 to 18,810 ms.
Thin-framed, hollow, or concave parts (e.g. SULLYS-035's frame) have
inherently low part-to-bbox utilization yet nest tightly, so the 30%
MinUtilization floor wrongly dropped every candidate for them. Pair
quality is judged by the rotated pair bounding-box area the results
are already sorted on; utilization now only ever serves as the
high-aspect exception (UtilizationOverride), never as a rejection.
Adds a hollow-frame helper plus regression tests that kept pairs
exist with low utilization and results stay sorted by pair area.
Layouts placed exactly at the part spacing can land ~1e-4 short once
rotated, rounded (e.g. PEP's 4-decimal exports) and snapped to the
Clipper grid, so both validators rejected layouts that were correct in
practice. NestTolerances.SpacingSlack (0.0005, far below anything a
cutting machine resolves) is now subtracted from the spacing by
NestLayoutCheck's inflation and NestJobPlacementValidator's edge-distance
check. The frozen LegacyNestValidator takes the same rule so the
equivalence tests keep comparing like with like.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The benchmark validator and every engine test re-triangulated both parts
for each nearby pair. NestLayoutCheck now uses TriangulatedRegion, with
Collision.HasOverlap as the fallback when it cannot decide. Verdicts are
unchanged (the frozen-validator equivalence tests still pass); validating
100 discs went from 1,254 ms to 94 ms.
Co-Authored-By: Codex <noreply@openai.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The job runner already checked sheet bounds on material contours only,
but the benchmark validator and salvage scoring used Part.BoundingBox,
which includes scribe/etch moves. A PEP bend tick that ends a hair past
the part's edge passed the runner yet failed the benchmark when placed
flush to the sheet edge, and it could shrink the credited offcut. Marks
only mark the surface, so bounds and salvage now use material only.
Benchmark before/after (all five built-in engines, local fixtures,
salvage 0.5): no job changed validity or cost. Regression tests pin the
new rule: a protruding tick flush to the sheet edge is valid in all four
quadrants, and a tick past the parts envelope no longer shrinks salvage
(targeted fixture cost 130 -> 120).
Co-Authored-By: Codex <noreply@openai.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Engines had to reverse-engineer the benchmark validator: Opus55 assumed a
0.01 arc tolerance (the validator uses 0.001), Gpt6Astra added hand-tuned
paddings and copied the validator's check order, Qwen picked its chord
tolerance to stay under a constant it could not reference.
NestTolerances publishes the validator's arc tolerance, the Clipper grid
and SafeClearanceMargin (with its derivation). NestLayoutCheck moves the
benchmark NestValidator's checks into OpenNest.Engine as a public API
(Clears for a part pair, Violations for a whole result); NestValidator is
now a thin wrapper. Verdicts are unchanged: tests compare ordered
violation lists against a frozen copy of the old validator, and a
tangent-disc stress test covers 432 pairs at the safe margin.
Co-Authored-By: Codex <noreply@openai.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
All three plugin engines turned a RotationPolicy into trial angles by hand
(fixed angle, stepped sweep, or right angles plus the minimum-bounding-
rectangle angle for Automatic), each with its own normalization, dedup and
sweep caps. EnumerateAngles gives one deterministic, Allows-checked list;
RotationCandidates.ForShape adds the MBR-aligning angles via the existing
Polygon.FindBestRotation, and DistinctOutlines drops angles where the part
looks identical. A cap of one returns the sweep start rather than throwing,
since engines request a single sample for small orientation budgets.
Co-Authored-By: Codex <noreply@openai.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Engines optimized guesses at the benchmark cost: Opus55 re-implemented
salvage credit, Qwen used plate area per part area, Gpt6Astra ignored
salvage. NestJobCost moves StockLadder's EstimateNetArea into a public
home (net sheet area, unplaced-part penalty, whole-result Evaluate) and
the benchmark and StockLadder now call it. Scores are unchanged: tests pin
it against a frozen copy of the old computation and real benchmark runs.
Bounds still include marks, as before, so scores do not move.
Co-Authored-By: Codex <noreply@openai.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Every plugin engine rebuilt part material from a snapshot by hand and
filtered only rapids, so all three kept counting scribe/etch marks as
material after 1b5e1b1 fixed it in the host. JobPartGeometry is the
validator's own reader made public: SpecialLayers.IsMaterial, validated
closed contours, material area, and TryRead returning null for unreadable
parts. The job validators now use it, so engines and validation agree.
Co-Authored-By: Codex <noreply@openai.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Engines assembled NestJobResult by hand - instance indices, fulfillment,
stock usage, status and PlateCommitted progress - and Qwen38FlashNext got
PlateIndex wrong (stock index instead of sheet order). The builder assigns
plate and instance indices itself and rejects overproduction and exhausted
stock, so engines only decide placements.
Co-Authored-By: Codex <noreply@openai.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Every plugin engine re-derived the quadrant/edge-spacing work area by hand
(Gpt6Astra, Opus55 and Qwen each had a copy, as did the placement
validator). One definition on the stock removes that duplication and the
chance of an engine disagreeing with the validator's bounds.
Co-Authored-By: Codex <noreply@openai.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Every nesting-geometry consumer filtered only rapids, so scribe/etch
moves counted as part material. An etch tick that ends a hair outside
the outline (PEP bend ticks start on the notch edge) made the part
"open geometry leaving the material region": the job validator threw
and every built-in engine plus Gpt6Astra crashed on real PEP jobs
(PT75, drawing 4980 A01 PT77). Marks are only on the surface, so they
should never affect placement, collision, area, or validation.
- SpecialLayers.IsMaterial excludes Rapid and Scribe; used by drawing
area, canonical angle, part collision, PartGeometry, plate perimeter,
best-fit/pair evaluation, rotation analysis, GPU evaluators, and both
validators. Timing, display, splitting and posts still see marks.
- ConvertGeometry also maps the saved SCRIBE layer name to Scribe, so
programs rebuilt from stored entities keep their marks.
- NestReader repairs older files (e.g. PepNestExport output) whose
programs saved etch as cut moves while source entities kept SCRIBE.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Every DefaultPlateFiller.Fill makes a fresh canonical copy of the drawing,
and BestFitCache/FillResultCache keyed by drawing reference, so fills never
shared results and the static caches grew without bound.
CanonicalFrame now records which drawing each canonical copy came from.
Both caches key weakly on that source drawing, so every canonical copy
shares one entry and released drawings can be collected. An entry is
dropped when the drawing's Program instance or canonical angle changes.
Best-fit candidates are computed once per (drawing, spacing) through
BestFitFinder.FindCandidates and filtered per plate size with the same
filter FindBestFits uses. FillResultCache keeps canonical and
non-canonical callers apart.
Adds Debug-only PerfCounters for best-fit runs, offset perimeter builds
and Part.Intersects calls.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The candidate validator flattened every arc into 1000 segments and rebuilt
both parts' polygons and edge lists for every pair it compared, so spacing
checks on filleted parts cost millions of edge pairs each. Validation, not
the fill pipeline, was nearly all of a solve's wall time.
Placed contours are now flattened once with ToPolygonWithTolerance(0.001),
the tolerance the benchmark NestValidator and Part.Intersects already use,
and each part's shape is built once per candidate. Arcs stay inscribed, so
a layout placed exactly at the spacing still passes.
12-nest PEP corpus, Default + StockLadder, --parallel 1: 2820 s -> 227 s.
Every run that finished before gives the same validity, count, plates and
cost. Three StockLadder runs that used to hit the 5-minute timeout now
finish.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Per-entity offsetting left spikes and inverted loops wherever a feature is
narrower than the spacing (1.nest), and RemoveSelfIntersections only
caught proper crossings. The callers that already flatten to polygons now
take a single Clipper region offset instead:
- PolygonHelper (BestFit) and PartBoundary use the conservative mode, which
keeps their never-under-estimate guarantee. PartBoundary also keeps holes
that appear when a perimeter curls back on itself.
- NestValidator offsets perimeter and cutouts in one region; Clipper drops
collapsed cutouts, so the collapsed-or-flipped heuristic goes away.
- CutOff.IntersectPerimeter offsets through the bridge. The old
OffsetEntity(Left) grew CW perimeters but shrank CCW ones, so with the
plate's perimeter cache a cut-off ran through the part; slots narrower
than twice the clearance now close up instead of leaving a gap.
- GetOffsetPartLines (3 overloads) and the AddOffset* helpers had no
callers and are removed, as is EntityView's never-defined DRAW_OFFSET
block.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
aj
