ComputeRemainderWithin only returned the larger of two possible free
rectangles, permanently losing usable area on the other axis after each
remainder item was placed. Replace the single shrinking box with a list
of free rectangles using guillotine cuts so both sub-areas remain
available for subsequent items.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
StripNestEngine only overrode Fill(NestItem), so ActionClone.Fill
and Pack operations fell through to the empty base class defaults.
Now all virtual methods delegate to DefaultNestEngine.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Wrap IProgress with AccumulatingProgress so remnant fills prepend
previously placed strip parts to each report. The UI now shows the
full picture (red + purple) instead of replacing strip parts.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Nest() now deducts placed counts from input NestItem.Quantity so the
UI loop doesn't create extra plates. All inner DefaultNestEngine.Fill
calls forward the IProgress parameter for live progress updates.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The auto-nest code paths (MainForm, MCP, Console) now call
engine.Nest(items, progress, token) instead of manually orchestrating
sequential fill+pack. The default implementation in NestEngineBase
does sequential FillExact+PackArea. StripNestEngine overrides with
its strip strategy. This makes the engine dropdown actually work.
Also consolidates ComputeRemainderWithin into NestEngineBase,
removing duplicates from MainForm and StripNestEngine.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
New NestEngineBase subclass that dedicates a tight strip to the
largest-area drawing and fills the remnant with remaining drawings.
Tries both bottom and left orientations, uses a shrink loop to find
the tightest strip, and picks the denser result.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The forward bounding-box gap check (gap < 0) incorrectly skipped obstacles
for irregular shapes like SULLYS-003 whose narrow handle extends past an
adjacent part's BB edge while the wide body still needs contact detection.
Replaced with a reverse-direction gap check that only skips obstacles the
moving part has entirely cleared. Also fixed edge distance check to prevent
overshooting the work area boundary when already at the limit.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
PushSelected now calls Compactor.Push instead of duplicating the push
logic. Compactor.Push moves parts as a group (single min distance) to
preserve grid layouts. Compact tries both left-first and down-first
orderings, iterating up to 20 times until movement drops below
threshold, and keeps whichever ordering traveled further.
Also includes a cancellation check in FillWithProgress to avoid
accepting parts after the user stops a nest.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Strip mode was adding thousands of candidates (7600+) when the work area
was narrow. Now caps at 100 total, sorted by utilization descending so
the best candidates are tried first.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The binary search was passing null for progress, so the NestProgressForm
showed all dashes during the entire search (potentially minutes). Now
each iteration reports progress — the user sees phases, part counts, and
density updating as the search runs.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Track productive angles across Fill calls; subsequent fills skip
angles that never produced results (knownGoodAngles)
- Binary search uses utilization-based range estimates (70%-25%)
instead of starting from the full work area dimension
- Quick bounding-box capacity check skips binary search entirely
when the plate can't fit more than the requested quantity
- Use full Fill (not rect-only) for binary search iterations so
the search benefits from pairs/linear strategies
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Previously each moving part only checked against the original stationary
set. Parts pushed earlier in the loop were invisible to later parts,
causing overlaps (utilization > 100%). Now each pushed part is added to
the obstacle set so subsequent parts collide correctly.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Show width x length and total part area on the "Nested:" row
in the nesting progress dialog, using the existing GetBoundingBox
extension to compute the extents.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The BestFitFilter's aspect ratio cap of 5.0 was rejecting valid pair
candidates needed for narrow plates (e.g. 60x6.5, aspect 9.2) and
remainder strips on normal plates. Three fixes:
- BestFitFinder: derive MaxAspectRatio from the plate's own aspect
ratio so narrow plates don't reject all elongated pairs
- SelectPairCandidates: search the full unfiltered candidate list
(not just Keep=true) in strip mode, so pairs rejected by aspect
ratio for the main plate can still be used for narrow remainder
strips
- BestFitCache.Populate: skip caching empty result lists so stale
pre-computed data from nest files doesn't prevent recomputation
Also fixes console --size parsing to use LxW format matching
Size.Parse convention, and includes prior engine refactoring
(sequential fill loops, parallel FillPattern, pre-sorted edge
arrays in RotationSlideStrategy).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Move NFP-based AutoNest logic (polygon extraction, rotation computation,
simulated annealing) into dedicated AutoNester class. Consolidate duplicate
FillWithPairs overloads, extract BuildCandidateAngles and BuildProgressSummary,
reorganize NestEngine into logical sections. Update callers in Console,
MCP tools, and MainForm to use AutoNester.Nest.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace recursive FillRecursive with flat FillGrid that tiles along primary
axis, then perpendicular. Extract FindPlacedEdge, BuildRemainingStrip,
BuildRotationSet, FindBestFill helpers. Use array-based DirectionalDistance
to eliminate allocations in FindCopyDistance and FindPatternCopyDistance.
Simplify FindSinglePartPatternCopyDistance to delegate to FindCopyDistance.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Vector implements IEquatable<Vector> with proper GetHashCode for HashSet usage.
Polygon.FindCrossing uses bounding-box pruning to skip non-overlapping edge pairs.
Helper.DirectionalDistance deduplicates vertices via HashSet, sorts edges for
early-exit pruning, and adds a new array-based overload that avoids allocations.
PartBoundary sorts directional edges and exposes GetEdges for zero-alloc access.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Linear phase shows "Linear: 12/36 angles, 45° = 48 parts" with a
running count. Pairs phase shows "Pairs: 8/50 candidates, best = 252
parts" tracking the best result seen so far. Reports on every
completion so the UI always reflects current state.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Parallel loops were flooding the UI with per-angle/per-candidate reports
faster than WinForms could render them. Use Interlocked timestamp checks
to report at most every 150ms, keeping descriptions readable.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Don't overwrite the Detail label with phase-level reports — let the
per-angle and per-candidate descriptions from the parallel loops remain
visible. Only clear the label on completion.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
ReportProgress was not setting Description, so the Detail row always
showed the default em-dash. Now each phase report includes a meaningful
description, and UpdateProgress always updates the label (resetting to
em-dash when null).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Mark _loadAttempted as volatile for correct double-checked locking
- Add Content item to copy Models/ directory to output for ONNX inference
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
