pythagorean-snap 0.1.0

O(log n) nearest-neighbor search on the full-circle Pythagorean manifold — zero drift, 1000x+ speedup

pythagorean-snap

O(log n) nearest-neighbor search on the full-circle Pythagorean manifold — zero drift, 1000x+ faster than brute force.

Install

pip install pythagorean-snap

Quick Start

from pythagorean_snap import PythagoreanManifold

m = PythagoreanManifold(50_000)
t = m.snap_angle(0.785)  # snap to nearest triple at ~45°
print(f"({t.x}, {t.y}, {t.c}) — angle: {t.angle():.4f} rad")
assert t.verify()  # x² + y² = c² — always exact

Performance

At max_c = 50,000 (41K+ triples):

• Binary search: O(log n) per query
• 100% agreement with brute-force
• Zero floating-point drift

API

Method	Description
PythagoreanManifold(max_c)	Build manifold
.snap_angle(theta)	Nearest triple to angle (rad)
.snap_brute(theta)	Brute-force (verification)
.constraint_distance(theta)	Angular distance
.angle_range()	(min, max) angle
.iter_triples()	Iterate all triples

License

MIT