Formalizing the Infinite-Dimensional Sato Grassmannian \(\mathrm{Gr}_p(H)\), Plabic \(\mathrm{Gr}_{\ge 0}(2,4)\) KP-II / KdV / AKNS Soliton PDE Solvers, Baker–Akhiezer Lax Pairs, and 6-Channel Dual-Schubert QFIM in Lean 4, Rust/PyO3, and SymPy


Abstract

Finite-dimensional Grassmannians \(\mathrm{Gr}_{\mathbb{C}}(2,4)\) and positive Plabic parametrizations \(\mathrm{Gr}_{\ge 0}(2,4)\) capture single-soliton and two-soliton interactions, but the full Kadomtsev–Petviashvili (KP) hierarchy lives on the infinite-dimensional Sato–Segal–Wilson Grassmannian \(\mathrm{Gr}_p(H)\) of a polarized Hilbert space \(H = H_+ \oplus H_-\). We present a complete, zero-sorry Lean 4 formalization of the infinite-dimensional Sato Grassmannian and its integrable PDE hierarchy inside the Signals library (Signals.SatoGrassmannian.*), paired with a compiled Rust/PyO3 PlabicGr24Solver and exact SymPy symbolic verification.

Our formalization establishes six main results:

  1. Infinite-Dimensional Polarized Hilbert-Space Grassmannian & Riesz–Schauder Fredholm Theory (Base.lean, RieszSchauder.lean, CircleL2.lean): We formalize PolarizedHilbertSpace, Fredholm projections IsFredholm, the Riesz–Schauder theorem (isFredholm_id_add_compact, index zero for compact perturbations of the identity), the Hardy space polarization on \(H = L^2(S^1, \mathbb{C})\) with unilateral shift \(S = M_z\) (shiftCLE), and the restricted general linear group \(\mathrm{GL}_{\mathrm{res}}(H)\) preserving \(\mathrm{Gr}(H)\) with additive Fredholm virtual dimension.

  2. Exact \(\mathrm{Gr}_{\mathbb{C}}(2,4)\) Truncation, Klein Quadric, and Two Formalization Bug Fixes (Gr24Truncation.lean, LaxPair.lean): Embedding \(\mathbb{C}^4 \cong \operatorname{span}\{z^{-2}, z^{-1}, 1, z\} \subset L^2(S^1)\) yields an explicit isomorphism onto the big cell of \(\mathrm{Gr}_{\mathbb{C}}(2,4)\) cut out by the Klein quadric \(p_{01}p_{23} - p_{02}p_{13} + p_{03}p_{12} = 0\). Formalizing the chat transcript caught and proved two genuine algebraic errors in the source derivation: (i) a sign inversion in the big-cell frame reconstruction kleinQuadric_reconstruct_bigCell (Block 005, lines 2475–2995), and (ii) a coefficient error in the third spatial derivative \(\partial_x^3 \psi\) of the Baker–Akhiezer wave function (baWave_lax_equations, Block 019, lines 12135–12380), where the true term is \(+6 z k^3 S_2 T - 4 k^4 S_2 T^2\) rather than \(+4 k^3 S_2 T (z - k T)\).

  3. Hirota Bilinear KP-II Identity, Nonlinear \(u = 2\partial_x^2\log\tau\), and Smooth Kodama–Williams \(\mathrm{Gr}_{\ge 0}(2,4)\) Soliton Webs (KPFlow.lean, HirotaKP.lean, KP2Nonlinear.lean, PDESolver.lean): We prove hirotaKP_tauU_eq_zero identically via the Klein quadric, hirotaKP_eq_zero_iff_kp2 (equivalence between Hirota’s bilinear form and the nonlinear \((2+1)\)-D KP-II PDE \(\partial_x(4 u_t - u_{xxx} - 6 u u_x) - 3 u_{yy} = 0\)), strict positivity and global smoothness of \(\tau(x,y,t) > 0\) on the totally nonnegative cell \(\mathrm{Gr}_{\ge 0}(2,4)\) (solitonTau_pos_of_totallyNonneg), \(y\)-invariance of KdV 2-solitons (kdv2Soliton_y_invariant_and_smooth), and the exact KdV 1-soliton PDE (kdv_oneSoliton_pde).

  4. Baker–Akhiezer Lax Pair & AKNS Focusing NLS / Defocusing GPE Solitons (LaxPair.lean, AKNS.lean): We prove both the \(B_2\) (\(\partial_y \psi = (\partial_x^2 + u)\psi\)) and \(B_3\) (\(\partial_t \psi = (\partial_x^3 + \tfrac{3}{2}u\partial_x + w)\psi\)) Lax equations for the KdV 1-soliton, the \(2\times 2\) AKNS zero-curvature equivalence akns_zero_curvature_iff, the focusing NLS bright soliton brightSoliton_nls_pde, and the defocusing GPE dark soliton darkSoliton_gpe_pde.

  5. 3-Channel Plücker–Madelung & 6-Channel Dual-Schubert QFIM (DualSchubertQFIM.lean): We prove \(F_Q = 4\,g_{\mathrm{FS}}\) (qfim_eq_four_fubiniStudy), the 3-channel Plücker–Madelung decomposition (three_channel_plucker_madelung_qfim), and the additive 6-channel \(\mathrm{Gr}_{\mathbb{C}}(2,4)\times\mathrm{Gr}_{\mathbb{C}}(2,4)^\vee\) decomposition (dual_schubert_six_channel_decomposition).

  6. Overflow-Proof Log-Sum-Exp Rust/PyO3 PDE Solver (sato_grassmannian.rs): Using normalized Boltzmann weights \(w_r = c_r e^{\theta_r - M}/Z\) and central cumulants up to 6th order, PlabicGr24Solver evaluates arbitrary \(\mathrm{Gr}_{\ge 0}(2,4)\) KP-II soliton webs even at \((x, y, t) = (100, -50, 20)\) without floating-point overflow, achieving kp2_residual = 3.11e-15, lax_b2 = 1.11e-16, lax_b3 = 4.97e-16, akns_bright = 2.48e-16, akns_dark = 2.29e-16, and qfim_res = 0.0.


1. Architecture of the Signals.SatoGrassmannian Core & PDE Stack

        flowchart TB
    subgraph Core["Hilbert-Space Sato Grassmannian Core"]
        B1["Base.lean (Blocks 001–002)\nPolarizedHilbertSpace, IsFredholm,\nSatoGr, GL_res(H)"]
        B2["RieszSchauder.lean (Block 003)\nCompact perturbations of Id,\nindex(I + K) = 0"]
        B3["CircleL2.lean (Block 004)\nL²(S¹) Fourier basis e_n,\nHardy H_+ ⊕ H_-, shiftCLE"]
        B4["Gr24Truncation.lean (Block 005)\nGr_C(2,4) ↪ Gr(L²(S¹)), Klein Q_4,\nfixed big-cell frame signs"]
    end

    subgraph PDE["Integrable PDE Hierarchy & Lax/AKNS/QFIM"]
        P1["KPFlow.lean & HirotaKP.lean\n(Blocks 006–007)\nexp(∑ t_n z^n), schurTau,\nhirotaKP_tauU_eq_zero"]
        P2["KP2Nonlinear.lean & PDESolver.lean\n(Blocks 012, 018)\nu = 2 ∂_x² log τ, Gr_≥0(2,4) > 0,\nKdV 1- & 2-solitons"]
        P3["LaxPair.lean & AKNS.lean\n(Blocks 019–020)\nFixed B_2/B_3 Lax pair,\nNLS bright & GPE dark solitons"]
        P4["DualSchubertQFIM.lean (Block 022)\nF_Q = 4 g_FS, 3-channel Madelung,\n6-channel Gr(2,4) × Gr(2,4)^∨"]
    end

    subgraph Runtime["Compiled Rust/PyO3 + SymPy Verification"]
        R1["sato_grassmannian.rs (Block 021+)\nLog-sum-exp PlabicGr24Solver,\nBA Lax, AKNS, Dual-Schubert QFIM"]
        R2["sato_grassmannian.py & pytest\n13 SymPy exact proofs +\nPyO3 residual audits (< 10^-14)"]
    end

    B1 --> B2 --> B3 --> B4 --> P1 --> P2 --> P3 --> R1
    B4 --> P4 --> R1 --> R2
    

Every theorem in the 11 Lean 4 modules compiles under Lean v4.35.0-rc3 and Mathlib with zero sorry, zero admit, and zero module-level warnings, integrated into the repository’s top-level Signals library (3,830 Lake jobs via make signals_build).


2. Polarized Hilbert Space, Riesz–Schauder Fredholm Theory, and \(L^2(S^1)\)

Following Segal–Wilson [Segal and Wilson, 1985] and Pressley–Segal [Pressley and Segal, 1986], a polarized complex Hilbert space (Base.lean, sourced from data/chats/Gemini_Formalizing-Sato-Grassmannian-in-Lean_20261120-2337.md, lines 73–1160) is an orthogonal decomposition \(H = H_+ \oplus H_-\) into closed subspaces with orthogonal projections \(\pi_+ : H \to H_+\) and \(\pi_- : H \to H_-\). A closed subspace \(W \subset H\) belongs to the Sato Grassmannian \(\mathrm{Gr}(H)\) if:

  1. The restricted projection \(\pi_+|_{W} : W \to H_+\) is Fredholm (finite-dimensional kernel, closed range, finite-dimensional cokernel).

  2. The cross projection \(\pi_-|_{W} : W \to H_-\) is compact (and in the Hilbert–Schmidt Grassmannian \(\mathrm{Gr}_2(H)\), Hilbert–Schmidt).

In RieszSchauder.lean (lines 1255–1989), we establish the complete Riesz–Schauder Fredholm package from scratch over Mathlib’s IsCompactOperator: for every compact \(K \in \mathcal{K}(H)\), the operator \(I + K\) has finite-dimensional kernel, closed range, finite-dimensional cokernel, and Fredholm index zero (isFredholm_id_add_compact).

In CircleL2.lean (lines 1997–2326), we instantiate \(H = L^2(\mathrm{AddCircle}(2\pi), \mathbb{C})\) using Mathlib’s orthonormal Fourier basis fourierBasis \(\{e_n\}_{n \in \mathbb{Z}}\) with \(e_n(\theta) = e^{in\theta}\), splitting into the Hardy subspace \(H_+ = \overline{\operatorname{span}}\{e_n : n \ge 0\}\) and its orthogonal complement \(H_- = \overline{\operatorname{span}}\{e_n : n < 0\}\). Multiplication by \(e^{i\theta}\) defines the unitary bilateral shift shiftCLE : HCircle ≃L[ℂ] HCircle, whose restriction to \(H_+\) has kernel \(\{0\}\) and 1-dimensional cokernel \(\mathbb{C} e_0\), giving Fredholm index \(-1\).


3. Klein Quadric Truncation \(\mathrm{Gr}_{\mathbb{C}}(2,4) \hookrightarrow \mathrm{Gr}(L^2(S^1))\) and Two Machine-Caught Bug Fixes

Mapping \((v_0, v_1, v_2, v_3) \in \mathbb{C}^4\) to \(v_0 e_{-2} + v_1 e_{-1} + v_2 e_0 + v_3 e_1 \in L^2(S^1)\) embeds any 2-plane \(U = \operatorname{span}\{v, w\} \subset \mathbb{C}^4\) into the virtual-dimension-\(0\) Sato Grassmannian via \(W_U = U \oplus \overline{\operatorname{span}}\{e_n : n \ge 2\}\) (Gr24Truncation.lean, lines 2475–2995). Its six Plücker minors \(p_{ij} = v_i w_j - v_j w_i\) (\(0 \le i < j \le 3\)) satisfy the Klein quadric equation:

\[ p_{01} p_{23} - p_{02} p_{13} + p_{03} p_{12} = 0. \]

3.1 Machine-Caught Fix 1: Big-Cell Frame Reconstruction Signs (Gr24Truncation.lean)

On the big cell \(p_{23} \ne 0\), the unique reduced row-echelon frame \((v_{\mathrm{norm}}, w_{\mathrm{norm}})\) whose \((2,3)\)-columns form \(I_2\) and whose Plücker coordinates recover \(p_{ij} / p_{23}\) is:

\[ v_{\mathrm{norm}} = \left(\frac{p_{03}}{p_{23}},\; \frac{p_{13}}{p_{23}},\; 1,\; 0\right), \qquad w_{\mathrm{norm}} = \left(-\frac{p_{02}}{p_{23}},\; -\frac{p_{12}}{p_{23}},\; 0,\; 1\right). \]

In the source chat (Block 005, lines 2475–2995), w_norm was originally written with positive signs \((+p_{02}/p_{23}, +p_{12}/p_{23}, 0, 1)\), which gives \(v_{\mathrm{norm},0} w_{\mathrm{norm},2} - v_{\mathrm{norm},2} w_{\mathrm{norm},0} = -p_{02}/p_{23}\) — off by a sign from \(p_{02}/p_{23}\) (chat_block005_bug_confirmed_nonzero = True in SymPy). Lean 4’s kernel rejected the erroneous sign during kleinQuadric_reconstruct_bigCell, and correcting the signs of \(w_{\mathrm{norm}}\) makes both kleinQuadric_reconstruct_bigCell and the infinite-dimensional Segal–Wilson theorem segalWilson_tau_gr24_eq_tauU (Paper 2) close identically.

3.2 Machine-Caught Fix 2: Baker–Akhiezer \(\partial_x^3 \psi\) Coefficient (LaxPair.lean)

For the KdV 1-soliton potential \(u(x, t) = 2 k^2 \operatorname{sech}^2(\theta)\) with \(\theta = k x + k^3 t\), the Baker–Akhiezer wave function (LaxPair.lean, lines 12135–12380) is:

\[ \psi(x, y, t; z) = \bigl(z - k\tanh\theta\bigr)\exp\bigl(z x + z^2 y + z^3 t\bigr). \]

Setting \(T = \tanh\theta\), \(S_2 = \operatorname{sech}^2\theta = 1 - T^2\), and \(E = e^{zx + z^2 y + z^3 t}\), the exact third spatial derivative is:

\[ \partial_x^3 \psi = \Bigl[z^3(z - k T) - 3 z^2 k^2 S_2 + 6 z k^3 S_2 T + 2 k^4 S_2(1 - 3 T^2)\Bigr] E. \]

The source chat draft (Block 019) wrote \(+4 k^3 S_2 T (z - k T) + 2 k^4 S_2 (1 - 2 T^2)\) instead of \(+6 z k^3 S_2 T + 2 k^4 S_2 (1 - 3 T^2)\), leaving an exact discrepancy of \(2 k^3 S_2 T (z + k T) E\) (chat_block019_psi_xxx_bug_confirmed_nonzero = True in SymPy). With the corrected derivative, baWave_lax_equations proves both the \(B_2\) and \(B_3\) Lax equations unconditionally:

\[ \partial_y \psi = (\partial_x^2 + u)\psi, \qquad \partial_t \psi = \left(\partial_x^3 + \frac{3}{2} u\,\partial_x + w\right)\psi, \qquad w = -3 k^3 \tanh\theta \operatorname{sech}^2\theta. \]

4. Hirota Bilinear KP-II, Kodama–Williams \(\mathrm{Gr}_{\ge 0}(2,4)\) Soliton Webs, and AKNS Solitons

Under the 3-time KP flow \(\exp(t_1 z + t_2 z^2 + t_3 z^3)\) (KPFlow.lean, lines 3001–3617), the Sato \(\tau\)-function of a 2-plane \(U = \operatorname{span}\{v, w\} \in \mathrm{Gr}_{\mathbb{C}}(2,4)\) expands in Schur polynomials indexed by partitions inside the \(2\times 2\) Young diagram (HirotaKP.lean, lines 3736–4227):

\[ \tau_U(t_1, t_2, t_3) = p_{23} + p_{13} s_{(1)} + p_{03} s_{(2)} + p_{12} s_{(1,1)} + p_{02} s_{(2,1)} + p_{01} s_{(2,2)}, \]

where \(s_{(1)} = t_1\), \(s_{(2)} = t_2 + \tfrac{1}{2} t_1^2\), \(s_{(1,1)} = \tfrac{1}{2} t_1^2 - t_2\), \(s_{(2,1)} = \tfrac{1}{3} t_1^3 - t_3\), and \(s_{(2,2)} = \tfrac{1}{12} t_1^4 + t_2^2 - t_1 t_3\). Both HirotaKP.hirotaKP_tauU_eq_zero and SymPy (hirota_kp_eq_24_times_klein) prove the exact algebraic identity:

\[ \bigl(D_{t_1}^4 - 4 D_{t_1} D_{t_3} + 3 D_{t_2}^2\bigr)\tau_U \cdot \tau_U = 24\,(p_{01} p_{23} - p_{02} p_{13} + p_{03} p_{12}) = 0. \]

In KP2Nonlinear.lean (lines 7410–7748) and PDESolver.lean (lines 11446–12015), we prove:

  1. Bilinear-to-Nonlinear Equivalence (hirotaKP_eq_zero_iff_kp2): Wherever \(\tau(x, y, t) \ne 0\), setting \(u = 2\,\partial_x^2 \log\tau\) satisfies \(\partial_x(4 u_t - u_{xxx} - 6 u u_x) - 3 u_{yy} = 0\) if and only if Hirota’s bilinear expression vanishes.

  2. Wronskian Positroid Soliton Webs (solitonTau_eq_wronskian, hirotaKP_solitonTau_eq_zero, solitonTau_pos_of_totallyNonneg): For real ordered wavenumbers \(\kappa_0 < \kappa_1 < \kappa_2 < \kappa_3\) and a totally nonnegative frame \((v, w) \in \mathrm{Gr}_{\ge 0}(2,4)\) (all \(p_{ij} \ge 0\), not all zero),

\[ \tau_{\kappa, v, w}(x, y, t) = \sum_{0 \le i < j \le 3} p_{ij}(\kappa_j - \kappa_i)\exp\Bigl((\kappa_i + \kappa_j)x + (\kappa_i^2 + \kappa_j^2)y + (\kappa_i^3 + \kappa_j^3)t\Bigr) > 0 \]

is strictly positive on all of \(\mathbb{R}^3\), guaranteeing global smooth, blowup-free \((2+1)\)-D resonant soliton webs [Kodama and Williams, 2014].

  1. KdV Reduction (kdv2Soliton_y_invariant_and_smooth, kdv_oneSoliton_pde): Choosing symmetric wavenumbers \((-\kappa_2, -\kappa_1, \kappa_1, \kappa_2)\) and the KdV positroid cell annihilates the \(y\)-dependence (\(\partial_y u = \partial_{yy} u = 0\)), yielding exact smooth KdV 1- and 2-solitons solving \(4 u_t - u_{xxx} - 6 u u_x = 0\).

  2. AKNS Hierarchy (AKNS.lean, lines 12388–12739): Passing to \(2\times 2\) matrix Lax pairs [Ablowitz et al., 1974], we prove akns_zero_curvature_iff together with the focusing NLS bright soliton brightSoliton_nls_pde (\(q(x,t) = A\operatorname{sech}(Ax)e^{i A^2 t}\) solving \(i q_t + q_{xx} + 2|q|^2 q = 0\)) and the defocusing GPE dark soliton darkSoliton_gpe_pde (\(q(x,t) = A\tanh(Ax)e^{-2 i A^2 t}\) solving \(i q_t + q_{xx} - 2|q|^2 q = 0\)).


5. Three-Channel Plücker–Madelung, Six-Channel Dual-Schubert QFIM, & Evanescent Speed Law

In DualSchubertQFIM.lean (lines 13049–13254), we lift the projective geometry of \(\mathrm{Gr}_{\mathbb{C}}(2,4) \hookrightarrow \mathbb{P}^5(\mathbb{C})\) into a machine-checked quantum-metrology identity connecting directly to our companion paper Why the Quantum Fisher Information Metric Is Four Times the Fubini–Study Kähler Metric:

\[ \|\partial_x \Psi\|^2 = \rho\,\bigl(\operatorname{Im}\langle\hat{p}, \partial_x\hat{p}\rangle\bigr)^2 + \frac{(\partial_x\rho)^2}{4\rho} + \rho\,\frac{F_Q(\hat{p}, \partial_x\hat{p})}{4}, \qquad F_Q(p, v) = 4\,g_{\mathrm{FS}}(p, v) = 4\left(\frac{\|v\|^2}{\|p\|^2} - \frac{|\langle p, v\rangle|^2}{\|p\|^4}\right). \]

On the dual-Schubert product \(\mathrm{Gr}_{\mathbb{C}}(2,4) \times \mathrm{Gr}_{\mathbb{C}}(2,4)^\vee\), dual_schubert_six_channel_decomposition proves that the total QFIM splits additively into six nonnegative channels (three forward Plücker channels plus three dual-Schubert channels). Furthermore, Theorem 4.3 (sharoglazova_klaers_evanescent_speed_law in DualSchubertQFIM.lean, linked to er_epr_wormhole_bohm_blowup_and_sharoglazova_klaers_qfim_cushion_identities in Signals.QuantumTrajectories Section 9) formalizes the stationary coupled-waveguide evanescent state \(\Psi(x) = e^{-\bar{\kappa} x}(\cosh(\delta\kappa\,x), \sinh(\delta\kappa\,x))\) where \(\kappa_{\pm}^2 = 2m(|E_{\mathrm{kin}}| \mp \hbar J_0)/\hbar^2\):

\[ v_{\mathrm{Bohm}}^2 = 0, \qquad g_{\mathrm{FS}}(\hat{p}, \partial_x\hat{p}) = \frac{\delta\kappa^2}{\cosh^2(2\delta\kappa\,x)}, \qquad \left(\frac{\hbar}{m}\right)^2\left[0 + \left(\frac{\partial_x\rho}{2\rho}\right)^2 + \frac{F_Q}{4}\right] = \frac{2}{m}\bigl(|E_{\mathrm{kin}}| - \hbar J_0\tanh(2\delta\kappa\,x)\bigr). \]

Thus Channel 1 (Bohmian phase velocity) vanishes identically while Channel 2 (scalar osmotic gradient) and Channel 3 (projective Fubini–Study / QFIM) jointly recover the exact evanescent speed law without singularity. As recorded in Signals.Pending.SatoRiemannSpectralPendingBoundary, these exact geometric identities do not by themselves establish experimental tunneling speeds in coupled-waveguide cavities [Sharoglazova et al., 2025] or physical wormhole-throat stabilization.


6. Overflow-Proof Log-Sum-Exp Rust/PyO3 PDE Solver & Tri-Language Verification Summary

Directly evaluating \(\tau(x,y,t) = \sum_r c_r e^{\theta_r}\) in IEEE-754 f64 overflows above \(\theta_r \approx 709\) (e.g., at \((x, y, t) = (100, -50, 20)\)). In src/signals/numerical_adapter/src/sato_grassmannian.rs (extending chat Block 021, lines 12747–13021), PlabicGr24Solver shifts all exponents by \(M = \max_r (\ln c_r + \theta_r)\), forms normalized Boltzmann weights \(w_r = e^{\phi_r - M} / Z\), and computes the exact analytical KP-II jet up to 4th derivatives via central cumulants of \((\alpha_r, \beta_r, \gamma_r)\):

\[ u = 2\,\mu_{2,\alpha}, \qquad u_x = 2\,\mu_{3,\alpha}, \qquad u_{xx} = 2(\mu_{4,\alpha} - 3\mu_{2,\alpha}^2), \]
\[ u_{xxx} = 2(\mu_{5,\alpha} - 10\mu_{3,\alpha}\mu_{2,\alpha}), \qquad u_{xxxx} = 2(\mu_{6,\alpha} - 15\mu_{4,\alpha}\mu_{2,\alpha} - 10\mu_{3,\alpha}^2 + 30\mu_{2,\alpha}^3). \]

Verification Artifact

Domain / Benchmark Point

Lean 4 Theorem (Zero sorry)

SymPy Exact Proof

Rust/PyO3 Measured Residual

Big-cell Klein quadric (Gr(2,4))

Postnikov weights \((1.2, 0.8, 1.5, 2.1)\)

kleinQuadric_reconstruct_bigCell

big_cell_reconstruction_corrected_signs

\(|p_{01}p_{23}-p_{02}p_{13}+p_{03}p_{12}| < 10^{-15}\)

KP-II top-cell resonant web

\(\kappa=(-2,-0.5,0.7,1.8)\), \((1.3,-2.4,0.9)\)

hirotaKP_solitonTau_eq_zero

hirota_kp_eq_24_times_klein

\(u = 1.162107\), residual \(3.11 \times 10^{-15}\)

KP-II extreme-coordinate stability

\(\kappa=(-1,0,1,2)\), \((100,-50,20)\)

solitonTau_pos_of_totallyNonneg

Exact cumulant invariance

No overflow, residual \(< 10^{-11}\)

KdV 2-soliton \(y\)-invariance

\(k_1=0.6, k_2=1.3\), \(y=0\) vs. \(y=17.5\)

kdv2Soliton_y_invariant_and_smooth

Exact \(\beta_r\) symmetry

\(|u(y=0)-u(y=17.5)| < 10^{-13}\)

Baker–Akhiezer \(B_2\) & \(B_3\) Lax pair

\(k=0.8, z=0.3+1.1i\), \((0.4,-0.2,0.5)\)

baWave_lax_equations

ba_lax_b2_exact, ba_lax_b3_exact

\(B_2\): \(1.11 \times 10^{-16}\); \(B_3\): \(4.97 \times 10^{-16}\)

AKNS bright NLS & dark GPE solitons

\(A=1.3\), \((x,t)=(0.7,-0.4)\)

brightSoliton_nls_pde, darkSoliton_gpe_pde

akns_bright_nls_exact, akns_dark_gpe_exact

NLS: \(2.48 \times 10^{-16}\); GPE: \(2.29 \times 10^{-16}\)

6-channel Dual-Schubert QFIM

Normalized 6+6 Plücker densities

dual_schubert_six_channel_decomposition

dual_schubert_three_channel_plucker_madelung_qfim

\(F_Q^{\mathrm{tot}} = 0.988767\), residual \(< 10^{-15}\)

Sharoglazova–Klaers evanescent speed law

\(\hbar=m=1\), \(E_{\mathrm{abs}}=2.5\), \(J_0=1.5\), \(x=0.37\)

sharoglazova_klaers_evanescent_speed_law

sharoglazova_klaers_evanescent_speed_law

\(v_{\mathrm{Bohm}}^2 = 0\), residual \(< 10^{-14}\)


7. Reproducibility Commands

# 1. Verify all 21 Signals.SatoGrassmannian modules, SignalsTests, and SignalsPendingTests
make signals_build

# 2. Run Rust unit tests and build/install the PyO3 extension
cargo test --manifest-path src/signals/numerical_adapter/Cargo.toml sato_grassmannian
maturin develop --manifest-path src/signals/numerical_adapter/Cargo.toml --release --features extension-module

# 3. Run the SymPy + mpmath + PyO3 cross-verification test suite (15 symbolic proofs + PyO3/mpmath)
pytest -q tests/test_sato_grassmannian_sympy_pyo3.py

References