Research-Stack/0-Core-Formalism/lean
Brandon Schneider 81548958bb feat(lean): inline spectral refinement kernel into PistSimulation.lean
Port the Gauss-Jordan LSQ spectral pipeline from the cascade stash
into a self-contained, policy-clean module:

- BraggPeak, MagneticDomain, pseudoVoigtQ16, chiSqWindow inlined
  as pure Q16_16 types (no Float, no string fields in compute paths)
- MagneticRegime inductive type replaces string-based regime classification
- QuadraticPacket with vertex-form (a, h, k, Δ, width, χ²)
- MatrixPacket with Gram matrix G=AᵀA, det3, rank3, trace3, spectral placeholder
- ChaosState + chaosConverge affine IFS contraction loop (converges in 5 steps)
- 13 #eval! witnesses passing on synthetic 8-bin fixture
- Removes dependency on deleted Astrophysics modules

Build: lake build Semantics = 3541 jobs green.

Generated with [Devin](https://cli.devin.ai/docs)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-05-20 23:48:32 -05:00
..
external/OTOM Refactor provenance sources for open witness backends 2026-05-12 05:57:04 -05:00
LeanGPT Remove legacy Python prototypes superseded by Lean formalization 2026-05-19 15:34:17 +00:00
Semantics feat(lean): inline spectral refinement kernel into PistSimulation.lean 2026-05-20 23:48:32 -05:00
CHAIN_ALL_REVIEW_REPORT.md initial: sovereign research stack (consolidated, weightless, and lfs-optimized) 2026-05-04 18:11:36 -05:00