caelum0x 2b734f3a20 feat(v2-B): scikit-learn adapter (LawSynthDynamics/Regressor/Transformer)
The #1 adoption lever from the competitive analysis: a sklearn-compatible package
(python/lawsynth-sklearn) so the scikit-learn ecosystem adopts LawSynth with
near-zero switching cost (why gplearn/PySINDy win mindshare).
- LawSynthDynamics: fit(X, t=)/predict/simulate/score/equations over the real
  Study discovery; full get_params/set_params/__sklearn_tags__/NotFittedError.
- LawSynthRegressor (RegressorMixin), LawSynthTransformer (feature library +
  deterministic decorrelation prune), parsimony='auto' (lambda = Cov/Var).
- Inherits sklearn mixins when present; faithful standalone fallback otherwise.
Demo recovers an oscillator R2=0.999996; 50 tests pass (7 sklearn-integration
skip cleanly when sklearn absent). Deterministic, offline.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-08-18 02:34:31 +03:00

29 lines
694 B
Python

"""Shared fixtures: deterministic synthetic dynamical systems."""
from __future__ import annotations
import math
import pytest
@pytest.fixture
def oscillator():
"""Harmonic oscillator: x = cos t, v = -sin t (x' = v, v' = -x)."""
n, dt = 160, 0.05
t = [i * dt for i in range(n)]
X = [[math.cos(ti), -math.sin(ti)] for ti in t] # columns [x, v]
return X, t
@pytest.fixture
def damped():
"""A lightly damped linear system with two coupled states."""
n, dt = 200, 0.03
t = [i * dt for i in range(n)]
X = []
for ti in t:
env = math.exp(-0.15 * ti)
X.append([env * math.cos(2.0 * ti), -2.0 * env * math.sin(2.0 * ti)])
return X, t