vibespatial.spatial.prepared_polygon_mask¶

Prepared device carrier for bounded polygon-against-one-mask classification.

Attributes¶

cp

Classes¶

`PreparedPolygonMaskPhysicalShape`	Indexed candidate-refine shape for one-mask classification.
`PreparedPolygonMaskClassification`	Device row classification relative to one polygonal mask boundary.
`PreparedPolygonMask`	One polygonal mask prepared as an indexed physical boundary table.

Functions¶

prepared_polygon_mask_fp64_plan(...)

Return the exact predicate plan implemented by the prepared mask kernels.

Module Contents¶

vibespatial.spatial.prepared_polygon_mask.cp = None¶

vibespatial.spatial.prepared_polygon_mask.prepared_polygon_mask_fp64_plan() → vibespatial.runtime.precision.PrecisionPlan¶: Return the exact predicate plan implemented by the prepared mask kernels.

class vibespatial.spatial.prepared_polygon_mask.PreparedPolygonMaskPhysicalShape¶

Indexed candidate-refine shape for one-mask classification.

row_count: int¶

segment_count: int¶

indexed_query_count: int¶

dense_candidate_work: int¶

candidate_work: int¶

scheduled_index_lane_bound: int¶

exact_lane_bound: int¶

total_scheduled_lane_bound: int¶

free_device_bytes: int¶

output_bytes: int¶

candidate_tile_capacity: int¶

row_tile_size: int¶

segment_tile_size: int¶

tile_count: int¶

class vibespatial.spatial.prepared_polygon_mask.PreparedPolygonMaskClassification¶

Device row classification relative to one polygonal mask boundary.

valid: object¶

covered_by: object¶

exterior: object¶

boundary_unresolved: object¶

class vibespatial.spatial.prepared_polygon_mask.PreparedPolygonMask¶

One polygonal mask prepared as an indexed physical boundary table.

mask: vibespatial.geometry.owned.OwnedGeometryArray¶

boundary_lines: vibespatial.geometry.owned.OwnedGeometryArray¶

segment_index: object¶

precision_plan: vibespatial.runtime.precision.PrecisionPlan¶

classmethod from_owned(mask: vibespatial.geometry.owned.OwnedGeometryArray, *, precision_plan: vibespatial.runtime.precision.PrecisionPlan) → PreparedPolygonMask | None¶

classify_polygon_rows(rows: vibespatial.geometry.owned.OwnedGeometryArray) → PreparedPolygonMaskClassification | None¶

Classify rows through an indexed, capacity-backed candidate relation.

Boundary MBR queries and representative-point ray queries share one reusable segment index. Count/scan/scatter emits only refined candidates into structural tile capacity; exact-sign ray parity then classifies rows without a candidate-count host fence.