- FillContext gains a Policy property (init-only) carrying the IFillComparer
- DefaultNestEngine.Fill sets Policy = BuildPolicy() on every context
- RunPipeline now uses context.Policy.Comparer.IsBetter instead of IsBetterFill
- RunPipeline promoted to protected virtual so subclasses can override
- BuildAngles/RecordProductiveAngles overrides delegate to angleBuilder
- RunPipeline calls virtual BuildAngles/RecordProductiveAngles instead of angleBuilder directly
- TODO comment added in group-fill overload for Task 6 Comparer pass-through
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add virtual comparer, direction, and angle-building hooks to NestEngineBase
so subclasses can override scoring and direction policy. Rewire IsBetterFill
to delegate to the comparer instead of calling FillScore directly.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Implements two IFillComparer strategies that preserve axis-aligned remnants:
VerticalRemnantComparer minimizes X-extent, HorizontalRemnantComparer minimizes
Y-extent, both using a count > extent > density tiebreak chain. Includes 12
unit tests covering all tiebreak levels and null-guard cases.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Extracts the fill result scoring contract into IFillComparer with a DefaultFillComparer implementation that preserves the existing count-then-density lexicographic ranking via FillScore.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Convergence loop now uses FillLinear internally to measure actual
waste with geometry-aware spacing instead of bounding-box arithmetic
- Each candidate pair is tried in both Row and Column orientations to
find the shortest perpendicular dimension (more complete stripes)
- CompleteStripesOnly flag drops partial stripes; remnant strip is
filled by a full engine run (injected via CreateRemnantEngine)
- ConvergeStripeAngleShrink tries N+1 narrower pairs as alternative
- FillResultCache avoids redundant engine runs on same-sized remnants
- CLAUDE.md: note to not commit specs/plans
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The convergence loop now ensures the pair starts with its short side
parallel to the primary axis, maximizing the number of pairs that fit.
Also adds ConvergeStripeAngleShrink to try N+1 narrower pairs, and
evaluates both expand and shrink results to pick the better grid.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Wire IDistanceComputer into RotationSlideStrategy, replacing inline
CPU/GPU branching. BestFitFinder constructs the appropriate implementation.
Replace PushDirection enum with direction vectors in BuildOffsets.
Rename IBestFitStrategy.Type and PairCandidate.StrategyType to StrategyIndex
for clarity (JSON field name unchanged for backward compatibility).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Extract distance computation from RotationSlideStrategy into a pluggable
IDistanceComputer interface. CpuDistanceComputer adds leading-face vertex
culling (~50% fewer rays per direction) with early exit on overlap.
GpuDistanceComputer wraps ISlideComputer with Line-to-flat-array conversion.
SlideOffset struct uses direction vectors (DirX/DirY) instead of PushDirection.
SpatialQuery.RayEdgeDistance(dirX,dirY) made public for CPU path.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Three bugs fixed in NfpSlideStrategy pipeline:
1. NoFitPolygon.Reflect() incorrectly reversed vertex order. Point
reflection (negating both axes) is a 180° rotation that preserves
winding — the Reverse() call was converting CCW to CW, producing
self-intersecting bowtie NFPs.
2. PolygonHelper inflation used OffsetSide.Left which is inward for
CCW perimeters. Changed to OffsetSide.Right for outward inflation
so NFP boundary positions give properly-spaced part placements.
3. Removed incorrect correction vector — same-drawing pairs have
identical polygon-to-part offsets that cancel out in the NFP
displacement.
Also refactored NfpSlideStrategy to be immutable (removed mutable
cache fields, single constructor with required data, added Create
factory method). BestFitFinder remains on RotationSlideStrategy
as default.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
NFP strategy has coordinate correction issues causing overlaps.
The slide-based approach is fast and accurate — keeping it as default.
NfpSlideStrategy and PolygonHelper remain in the codebase for future use.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Calling ConvexMinkowskiSum directly with manual reflection produced
wrong winding/reference-point handling, causing all pairs to overlap.
Route through Compute which handles reflection correctly. Hull inputs
keep it fast — few triangles means trivial Clipper union.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
ConvexMinkowskiSum is O(n+m) with no boolean geometry ops.
The concave Minkowski path was doing triangulation + pairwise
sums + Clipper2 Union, which hung at 100% CPU for complex parts.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace RotationSlideStrategy with NfpSlideStrategy in BuildStrategies,
and add integration tests covering the end-to-end FindBestFits pipeline.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Creates PolygonHelper.cs in OpenNest.Engine.BestFit with ExtractPerimeterPolygon
(returning PolygonExtractionResult with polygon + correction vector) and RotatePolygon.
AutoNester.ExtractPerimeterPolygon and RotatePolygon become thin delegates.
Adds MakeSquareDrawing/MakeLShapeDrawing to TestHelpers and 6 PolygonHelperTests.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
In strip mode, build candidate list entirely from pairs whose
ShortestSide fits the narrow work area dimension, sorted by
estimated tile count. Previously, the top-50 utilization cut
ran first, excluding good strip candidates like #183.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Remove NeedsSweep that triggered a 5-degree sweep (36 angles) when
the work area was narrower than the part. Position matters more than
angle for narrow areas, and the base angles (bestRotation + 90deg)
cover the useful cases. ForceFullSweep still works for training.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Normalize pair bounding box to landscape (width >= height) in
PairEvaluator for consistent display and filtering. Fix
BestFitViewerForm where BoundingWidth/BoundingHeight were passed
in the wrong order to the plate Size constructor.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add Compactor.Settle and AutoNester.Optimize post-passes to
NestEngineBase.Nest, StripNestEngine, and PlateView.FillWithProgress
so all fill paths benefit from geometry-aware compaction.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add Optimize method that re-places parts using NFP-based BLF, keeping
the result only if it improves density without losing parts. Fix
perimeter inflation to use correct offset side. Add NfpNestEngine
that wraps AutoNester for the registry. Register NFP engine.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add Settle method that repeatedly pushes parts left then down until
total movement falls below a threshold. Replaces manual single-pass
push calls for more consistent gap closure.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Refactor BLF to compute NFP paths as Clipper PathsD with offsets
instead of translating full polygons. Add spatial pruning to skip
NFPs that don't intersect the IFP bounds. Clamp placement points
to IFP bounds to correct Clipper2 floating-point drift. Add
progress reporting to simulated annealing. Add debug logging.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Move PlacedPart to its own file. Replace tuple-based sequences with
SequenceEntry struct for clarity. Add IProgress parameter to
INestOptimizer. Add IFP caching to NfpCache to avoid recomputing
inner fit polygons for the same drawing/rotation/workArea.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
PairFiller now returns PairFillResult (Parts + BestFits) instead of
using a mutable BestFits property. Extracted EvaluateCandidates,
TryReduceWorkArea, and BuildTilingAngles for clarity. Simplified the
candidate loop by leveraging FillScore comparison semantics.
Removed FillRemainingStrip and all its helpers (FindPlacedEdge,
BuildRemainingStrip, BuildRotationSet, FindBestFill, TryFewerRows,
RemainderPatterns) from FillLinear — these were a major bottleneck in
strip nesting, running expensive fills on undersized remnant strips.
ShrinkFiller + RemnantFiller already handle space optimization, making
the remainder strip fill redundant.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
StripNestEngine was passing progress directly to DefaultNestEngine.Fill
inside the ShrinkFiller loop, causing every per-angle/per-strategy report
to update the UI with overlapping layouts in the same work area.
Now inner fills are silent (null progress) and ShrinkFiller reports its
own progress when the best layout improves. IterativeShrinkFiller tracks
placed parts across items and includes them in reports. The trial box is
reported before the fill starts so the work area border updates immediately.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When the first pair candidate places more parts than needed (e.g., 17
when target is 10), sort by BoundingBox.Top, trim from the top until
exactly targetCount remain, and use that Top as the new work area
height. All subsequent candidates fill this smaller area, dramatically
reducing fill time.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
When a target count is known, ShrinkFiller now uses FillBestFit (fast
rectangle packing) to estimate how many parts fit on the full area,
then scales the shrink axis proportionally to avoid an expensive
full-area fill. Falls back to full box if estimate is too aggressive.
Also shrinks to targetCount (not full count) to produce tighter boxes
when fewer parts are needed than the area can hold.
IterativeShrinkFiller passes NestItem.Quantity as the target count.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
PairEvaluator already computes OptimalRotation via RotatingCalipers on
the pair's convex hull, but PairFiller.EvaluateCandidate only passed
hull edge angles to FillPattern. Now includes the optimal rotation
angle (and +90°) so tiling can use the mathematically tightest fit.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Both types were only used internally by the old StripNestEngine.Nest
strip-orientation logic, which has been replaced by IterativeShrinkFiller.
No references remain outside of these files.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replaces the old orientation-based strip nesting (TryOrientation,
SelectStripItemIndex, EstimateStripDimension, ShrinkFill helpers) with
a call to IterativeShrinkFiller.Fill for multi-quantity items, plus a
RemnantFinder-based PackArea pass for singles and leftovers.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Introduces IterativeShrinkFiller.Fill, which composes RemnantFiller and
ShrinkFiller by wrapping the caller's fill function in a closure that tries
both ShrinkAxis.Height and ShrinkAxis.Width and picks the better FillScore.
Adds IterativeShrinkResult (Parts + Leftovers). Covers null/empty inputs and
single-item placement with three passing xUnit tests.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Kept using OpenNest.Api in Timing.cs and EditNestForm.cs alongside
remote's reorganized usings and namespace changes.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- NestRunner now assigns Material to plates from request.Material
- NestResponse.LoadAsync uses descriptive exceptions instead of null-forgiving operators
- Fix pre-existing FillExtents.Fill signature mismatch (add bestFits parameter)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Extract magic numbers into named constants (MaxTopCandidates,
EarlyExitMinTried, etc.), extract candidate evaluation into
EvaluateCandidate method, and expose BestFits property so
PairsFillStrategy can reuse without redundant BestFitCache call.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The pattern bounding box already computes max(upper) - min(lower), so the
manual loop was redundant. Extract the N×N pair distance loop into a static
FindMaxPairDistance helper. Drop pre-cached edge arrays since GetEdges()
returns stored references with zero allocation.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
