RUN LOG -- verification scripts for J. Thrussell, 'Riemannian three-manifolds as hypersurfaces of Euclidean four-space: the realizability operator, its symbol, and its linearisation at the round sphere' (2026). DOI 10.5281/zenodo.23765514. Final run of every script in its deposited form, 29 October 2026. Two docstring-only edits followed some of these runs and touch no code: the author's archive labels were removed from the docstrings of 326, 329, 330, 332 and 333, and 338's docstring writes R^4 for E^4. Each script's docstring records its edits. test_354 was run in its deposited form on 1 November 2026. Tests 625, 626, 627 and 629 were run in their deposited form on 12 November 2026 (Python 3.13.18, the same library versions). Test 640 was run in its deposited form on 17 November 2026 (Python 3.13.18, sympy 1.14.0). Python 3.13.16; sympy 1.14.0; numpy 2.5.3; scipy 1.18.1; mpmath 1.3.0 SUMMARY test_325_realizability_symbol.py 6 pass 0 fail exit 0 [81s] test_326_symbol_module_is_immersions.py 4 pass 0 fail exit 0 [18s] test_327_round_sphere_conformal_path.py 5 pass 0 fail exit 0 [485s] test_329_degenerate_locus.py 4 pass 0 fail exit 0 [2s] test_330_ma_small_identities.py 6 pass 0 fail exit 0 [5s] test_331_contracted_polynomial_form.py 6 pass 0 fail exit 0 [2s] test_332_theorem_C_exact.py 3 pass 0 fail exit 0 [9s] test_333_zeta_count_scope.py 4 pass 0 fail exit 0 [19s] test_334_spinors_and_the_chiral_form.py 6 pass 0 fail exit 0 [7s] test_335_reduction_and_symbol_against_F.py 5 pass 0 fail exit 0 [26s] test_336_round_sphere_linearisation.py 5 pass 0 fail exit 0 [63s] test_337_shell_metrics_first_order.py 9 pass 0 fail exit 0 [44s] test_338_swept_spheres.py 8 pass 0 fail exit 0 [87s] test_339_second_order_round_base.py 6 pass 0 fail exit 0 [32s] test_340_second_order_general_bases.py 11 pass 0 fail exit 0 [350s] test_341_polynomial_form_and_the_degenerate_symbol.py 3 pass 0 fail exit 0 [5s] test_342_rank_one_branch.py 18 pass 0 fail exit 0 [3s] test_343_chiral_form.py 5 pass 0 fail exit 0 [8s] test_344_second_read_checks.py 4 pass 0 fail exit 0 [48s] test_354_symbol_module_every_K_and_general_bases.py 6 pass 0 fail exit 0 [9s] test_625_dF_equals_dC.py 2 pass 0 fail exit 0 [39s] test_626_rivertz_dualisation.py 2 pass 0 fail exit 0 [0s] test_627_zeta_tt_count.py 5 pass 0 fail exit 0 [0s] test_629_branson_intertwiner.py 2 pass 0 fail exit 0 [42s] test_640_rivertz_span.py 4 pass 0 fail exit 0 [2s] total checks passed: 139 ================================================================================================== test_325_realizability_symbol.py ================================================================================================== P1 the kernel at every nondegenerate point [PASS] sigma(xi) kills xi(.)v and Q_xi = P K P exactly (symbolic K, xi); rank 2 at random rational K and xi -- exact ranks [2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2] P2 where the symbol is elliptic: the first Newton tensor [PASS] TT-injective at xi iff xi^T T1 xi != 0; diag(2,3,-3) counterexample exact; characteristic covectors iff k1+k2 <= 0 <= k2+k3; conformal coincidence at fixed xi iff Q scalar -- TT singular value at null covectors 1.5e-14; min over definite-T1 cases 1.14e-01 P3 four identities among the nine components; five genuine conditions [PASS] rank tau_3 = 5; the cyclic identity and the three Bianchi functionals B_b(C) are independent and span the left null space -- rank tau_3 5, rank tau_3 5, rank tau_3 5 P4 Cartan's test at order 3 [PASS] (re-posed) rank tau_4 = 12, its left null space = 12 prolonged identities + 3 functionals xi ^ C; dim g_4 = 78 = 55 + 19 + 4: involutive at order 3 -- K=[1.3, 2.1, 0.8]: rank tau_4 12, dim g_4 78, Cartan sum 55 + 19 + 4 = 78; K=[2.0, 3.0, -3.0]: rank tau_4 12, dim g_4 78, Cartan sum 55 + 19 + 4 = 78 P5 four functions of three variables; the joint kernel is K [PASS] dim g_q has second differences 4 (four functions of three variables); joint kernel = span{K} -- K=[1.3, 2.1, 0.8]: dim g_q (q=3..9) [55, 78, 105, 136, 171, 210, 253], second differences [4, 4]; K=[2.0, 3.0, -3.0]: dim g_q (q=3..9) [55, 78, 105, 136, 171, 210, 253], second differences [4, 4]; joint-kernel alignment with K 1.000000000000 P6 the relations are exact identities (cyclic, contracted Bianchi, d_nabla Cod K = R.K = 0 under Gauss) [PASS] cyclic sum of Cod K = 0; div adj K = B(Cod K); cyclic nabla Cod K = -R.K (Ricci + Bianchi), exactly at a curved point; R.K = 0 under Gauss, symbolically -- residuals 0, 0, 0; Gauss cancellation True; div adj K at 0 = [-39037/25200, -91057/12600, -29783/3150] test_325: ALL PASS exit 0 [81s] ================================================================================================== test_326_symbol_module_is_immersions.py ================================================================================================== P1 the two lemmas behind the symbol theorem, exactly [PASS] sigma_2(DM)(xi) h = -1/2 adj_2(PhP), zero on gauge; D adj_K polynomial for every K, D adj_K(xi xi^T) = adj_2(PKP) P2 the immersions' symbol module M_K: dimension (q+2)(2q+5), inside g_q [PASS] rank M_K = (q+2)(2q+5) and tau_q M_K = 0 for q = 3, 4, 5 at all fifteen K -- q=3: rank M_K 55 = 55, tau_3 M_K = 0: True P3 Cartan's test at fifteen special shape operators; g_q = M_K [PASS] dim g_3 = 55, g_{3,1} = 19, g_{3,2} = 4, dim g_4 = 78 = Cartan sum, dim g_5 = 105 at every K: g_q = M_K, involutive -- (1, 1, 1):ok, (1, 1, 2):ok, (1, 1, -1):ok, (2, 3, -3):ok, (-1, 1, 2):ok, (1, 2, 3):ok, (1, 1, -2):ok, (-1, 2, 2):ok, (2, 2, -1):ok, (1, -2, -2):ok, (3, -1, -1):ok, (1, 2, -1):ok, (5, -2, 1):ok, (1, 1, 1/10):ok, (1, -1, 1):ok P4 the Kulkarni-Nomizu map with K is injective iff det K != 0 [PASS] det(A -> A (o) K on Sym^2) = 2 (det K)^2 -- 2*(w0*w1*w2 - w0*w5**2 - w1*w4**2 - w2*w3**2 + 2*w3*w4*w5)**2 test_326: ALL PASS exit 0 [18s] ================================================================================================== test_327_round_sphere_conformal_path.py ================================================================================================== P1 the implementation: F(gamma) = 0 and F(radial graph) = 0 through t^2 [PASS] F(gamma) = 0, R = 6; F((1+u)^2 gamma + du du) = 0 exactly through t^2 for a random cubic u -- radial-graph residual coefficients: {0} P2 the kernel at gamma: conformal and Lie directions [PASS] t-linear part of F(gamma + t h) is 0 for h = 2u gamma and h = L_X gamma (random cubic u, X) -- second-order part nonzero for the conformal one: True P3 the straight conformal path leaves at second order; the radial lift does not [PASS] F(e^{2tf} gamma) = -t^2 DF_gamma(df df) + O(t^3) with DF_gamma(df df) != 0; F(radial lift) = 0 through t^2 -- DF_gamma(df df) components at the point: [384/845, -55296/105625, 7962624/17850625, 884736/3570125] P4 the scalar-curvature detector on conformal directions [PASS] delta R(2u gamma) = -4(Delta + 3)u; 0 on a first harmonic; 4 lambda (2 cot^2 r - 1) cos theta on the shell dipole -- dipole value at the point: 5/26 = 5/26 P5 the count of ten [PASS] zeta(0) = 2[zeta_H(-2,3) - 4 zeta_H(0,3)] = 10 for lambda_n with an even-power expansion (test_333); 4 over n >= 1; scalars n >= 3: -5 -- exact 10, continuation 10.0 test_327: ALL PASS exit 0 [485s] ================================================================================================== test_329_degenerate_locus.py ================================================================================================== P1 the adjugate's ranks; a rank-2 M is never adj K [PASS] rank adj K = 3, 1, 0 for rank K = 3, 2, <= 1; adj of rank-2 diag(m1, m2, 0) is m1 m2 e3 e3^T P2 not vacuous where rank M = 2 (the audit's example) [PASS] M = diag(-1,-2,0), det M = 0, d_r det M = -12, P = diag(0,0,2), E_ryy = 24 and nothing else -- nonzero E components {(0, 2, 2): 24} P3 identically zero on a realizable rank-one point: the cone over a Clifford torus [PASS] M = diag(-1/t^2, 0, 0), det M = 0, P = 0 identically, E = 0 at the point P4 the Bianchi split on the cone [PASS] v = d_t geodesic (Gamma^i_tt = 0); div(mu v) = 0 -- div = 0 test_329: ALL PASS exit 0 [2s] ================================================================================================== test_330_ma_small_identities.py ================================================================================================== P1 det D adj_K = -2 (det K)^2 [PASS] det D adj_K on Sym^2 = -2 (det K)^2, general symmetric K P2 Theorem A(2) on the indefinite branch [PASS] adj(sqrt(det M) M^{-1}) = M for M of signature (+,-,-), det M > 0 P3 the Flamm hypersurface [PASS] metric, principal curvatures, M's eigenvalues, det M (1/256m^6 at the throat), finite distance, T_1 -- distance from 2m to 3m = m*(2*acosh(sqrt(6)/2) + sqrt(3)) P4 the count [PASS] n^2(n^2-1)/12 - n(n+1)/2 = n(n-3)(n+1)(n+2)/12, zero only at n = 3 P5 Lemma 7.3 [PASS] E'/E = -W.n, |n| = 1 P6 Proposition 7.6 [PASS] Q(V) = 1, -Q(V') = |W|^2, E = B(V,N)/2 with N null, Q((Phi V)') = Phi'^2 - Phi^2 |W|^2 test_330: ALL PASS exit 0 [5s] ================================================================================================== test_331_contracted_polynomial_form.py ================================================================================================== P1 phi (F M) = R on a generic metric with det M > 0 det M at the point = 171/200 [PASS] det M > 0; phi (F M)_ab^d = R_ab^d (nine components); d det M = tr(P nabla M); F != 0 there -- max |F| ~ 1.081e-1 P2 R = 0 on the round S^3 [PASS] M = I and R = 0 at the point P3 the audit's rank-2 point [PASS] M = diag(-1,-2,0); R_rx^x = -6, R_ry^y = 6 (and antisymmetric partners) -- nonzero {(0, 1, 1): -6, (0, 2, 2): 6, (1, 0, 1): 6, (2, 0, 2): -6} P4 the realizable rank-one cone [PASS] M = diag(-1/t^2, 0, 0) and R = 0 at the point P6 the identities of R [PASS] R_ab^b = 0 (three); cyclic sum of R_ab^d P_dc = 0 -- cyclic = 0 P5 a non-umbilic realizable graph [PASS] M = diag(6,3,2), nabla M != 0, F = 0, R = 0 test_331: ALL PASS exit 0 [2s] ================================================================================================== test_332_theorem_C_exact.py ================================================================================================== P1 five conditions at every nondegenerate K (symbolic, principal frame) [PASS] a 5x5 minor of the coordinate-covector images is a monomial; a 4x4 minor of the identities is +-k1k2k3; the identities kill tau_3 -- 5x5 minor (scaled by (det K)^10): k1**8*k2**8*k3**9/256 P2 Cartan's test in exact arithmetic at sixteen shape operators [PASS] dim g_3 = 55, g_{3,1} = 19, g_{3,2} = 4, dim g_4 = 78, dim g_5 = 105, exactly over Q, at all sixteen -- ('1', '1', '1'):ok, ('1', '1', '2'):ok, ('1', '1', '-1'):ok, ('2', '3', '-3'):ok, ('-1', '1', '2'):ok, ('1', '2', '3'):ok, ('1', '1', '-2'):ok, ('-1', '2', '2'):ok, ('2', '2', '-1'):ok, ('1', '-2', '-2'):ok, ('3', '-1', '-1'):ok, ('1', '2', '-1'):ok, ('5', '-2', '1'):ok, ('1', '1', '1/10'):ok, ('1', '-1', '1'):ok, ('13/10', '21/10', '4/5'):ok P3 the kernel is gauge + span{K}; sigma(xi) is injective on S_xi [PASS] sigma(xi) kills xi(.)v and K (direct, dim 4), rank 2, injective on S_xi; TT rank shown for comparison -- TT rank 1, S_xi rank 2; TT rank 2, S_xi rank 2; TT rank 2, S_xi rank 2; TT rank 2, S_xi rank 2; TT rank 2, S_xi rank 2; TT rank 2, S_xi rank 2 test_332: ALL PASS exit 0 [9s] ================================================================================================== test_333_zeta_count_scope.py ================================================================================================== P1 the continuation against direct sums [PASS] continuation = direct sum at s = 2, 3 for all five lambda -- n^3 + 7n, 2n^5 + n^3/2, n^3 + n^2, n^3 - n^2/2, n^3 + 2 P2 even powers: zeta(0) = 10 [PASS] n^3 + 7n and 2n^5 + n^3/2 give 10 -- 10.0, 10.0 P3 a first odd power changes the count [PASS] n^3 + n^2 gives 112/9; n^3 - n^2/2 gives 313/36 -- 12.4444444444444, 8.69444444444444 P4 a third odd power changes it too [PASS] n^3 + 2 gives 26/3 -- 8.66666666666667 test_333: ALL PASS exit 0 [19s] ================================================================================================== test_334_spinors_and_the_chiral_form.py ================================================================================================== P1 the two half-spin representations [PASS] rho_+ and rho_- are homomorphisms for one overall sign c; kernels 3 + 3 complementary; image elements square to -|v|^2/4 -- c = -1 P2 the Dirac operator on a generalized Killing spinor [PASS] sum_i e_i K(e_i) = -(tr K) 1 for symmetric K P3 end to end on a generic diagonal metric and generic K [PASS] spinor curvature = rho_s(curvature of Omega); tangential block = Riemann - k^k; mixed block = Codazzi of K -- (i) True, (ii) True, (iii) True P4 consequences [PASS] (a) paraboloid in R^4, both chiralities flat; (b) 1.1 K, both curved -- (a) True, (b) True [PASS] (c) round S^3: K = Id flat for both s (Killing spinors), K = 2 Id not flat [PASS] (d) K = Id on the generic metric: Codazzi holds, spinor curvature = rho_s(Gauss residual) != 0 for both s -- Codazzi True, Gauss-only True, curved True test_334: ALL PASS exit 0 [7s] ================================================================================================== test_335_reduction_and_symbol_against_F.py ================================================================================================== P1 det adj K = (det K)^2 [PASS] det adj K = (det K)^2, general symmetric K P2 adj K~ = M on det M > 0 [PASS] adj(sqrt(det M) M^{-1}) = M, M = diag(a, b, c/(ab)), c > 0, a, b any sign P3 K~(adj K) = sign(det K) K [PASS] K~(adj K) = sign(det K) K, all sign patterns (symbolic) and four rotated rational K P4 adj K > 0 iff K definite [PASS] adj K positive definite iff all k_i of one sign (eight sign patterns) P5 the symbol formula against the full operator [PASS] O(eps) part of F(g + eps (xi.x)^3/6 h0) at 0 = sigma(xi) h0, both branches, nonzero -- M(0) = diag(1, 2, 3): True; M0 = diag(1, 2, 3), xi = [3/5, 4/5, 0]: equal; M0 = diag(1, 2, 3), xi = [3/5, 4/5, 0]: equal; M0 = diag(1, 2, 3), xi = [2/3, 1/3, 2/3]: equal; M0 = diag(1, 2, 3), xi = [2/3, 1/3, 2/3]: equal; M(0) = diag(2, -1, -3): True; M0 = diag(2, -1, -3), xi = [3/5, 4/5, 0]: equal; M0 = diag(2, -1, -3), xi = [3/5, 4/5, 0]: equal; M0 = diag(2, -1, -3), xi = [2/3, 1/3, 2/3]: equal; M0 = diag(2, -1, -3), xi = [2/3, 1/3, 2/3]: equal test_335: ALL PASS exit 0 [26s] ================================================================================================== test_336_round_sphere_linearisation.py ================================================================================================== P1 DF_gamma h = Cod (1/2)(-Lap + 2) h on transverse-traceless jets 120 jet coefficients, 50 constraint equations, 70 free [PASS] O(eps) of F(gamma + eps h) at 0 = Cod((1/2)(-Lap h + 2h))(0) on three random TT jets; differs off TT -- trial 0: equal; trial 1: equal; trial 2: equal; control differs: True P2 level 3 on SU(2): constant trace-free tensors in a left-invariant frame [PASS] sectional curvature 1; div h = 0; -Lap h = 6h; |nabla h|^2 = 6|h|^2; |Cod h|^2 = 18|h|^2 = 2|nabla h|^2 + 6|h|^2 P3 the augmented symbol [PASS] (xi ^ H, H xi, (tr H) xi) injective in H; |xi ^ H|^2 = 2|xi|^2|H|^2 on {H xi = 0, tr H = 0} P4 the count of ten and the determinant [PASS] zeta(0) = 10 for both multipliers; ln det_zeta(n(n^2-1)) on TT: numerical = closed form = -14.0371968... -- zeta(0) = 10.0, 10.0; ln det = -14.0371968383834 vs -14.0371968383834 P5 the shell dipole's decomposition [PASS] 2 lambda cos(theta) dr^2 = 2 w gamma + L_eta gamma, w = (lambda + c sin r) cos(theta), all c, q0 test_336: ALL PASS exit 0 [63s] ================================================================================================== test_337_shell_metrics_first_order.py ================================================================================================== [PASS] P1 background M0 = diag(Q, P, P), K~0 = diag(P/sqrt Q, sqrt Q, sqrt Q) on every base [PASS] P2 first family C_{r th r} = -((Lam-2)/2) A^3/(Phi Z') dY (l = 2, 3) [PASS] P3 second family C_{th ph ph} = -((Lam-2)/2) Z'/(A Phi P) sin^2 th dY (l = 2, 3) [PASS] P4 third family = D T[Y]; D measured = closed form = curvature form; l- and angle-independent -- round: D = 0; B2: D = -17/47; B2': D = -73/122; B3: D = 314163*sqrt(161)/14168; B4: D = -25911*sqrt(455)/902720; BC: D = 0 [PASS] P5 the five components with an odd number of phi indices vanish [PASS] P6 kernel: l = 0 and l = 1 give C = 0 on every base; l = 2, 3 do not P7 two further first-order kernel directions [PASS] P7 (a) l = 1 with r-dependent amplitude, (b) off-diagonal beta(r) sin(theta): all O(eps) components vanish -- B2: True, True; B3: True, True; B4: True, True P8 symbolic, general base [PASS] P8 closed form = curvature form (general base); D = 0 on the round base and on the constant-ratio family P9 D on a constant-ratio base and on the others [PASS] P9 D = 0 on the round base and BC (k_mer/k_sph = 1/2); D != 0 on B2, B3, B4 -- round: 0, B2: -17/47, B2': -73/122, B3: 314163*sqrt(161)/14168, B4: -25911*sqrt(455)/902720, BC: 0 [43s] test_337: ALL PASS exit 0 [44s] ================================================================================================== test_338_swept_spheres.py ================================================================================================== P1 the swept-sphere construction reproduces A^2 = H^2 + (alpha + Phi b.n)^2, round shells, no cross terms [PASS] explicit embedding: g = diag(H^2 + (alpha + Phi b.n)^2, Phi^2/E^2, Phi^2/E^2), cross terms 0 -- max relative deviation 1.3e-10; chart flow is a CKF to 8.2e-11; g_rr 0.1253966056 vs 0.1253966056 P2 tilt with no chart drift, convex branch, F = 0 [PASS] tilt-only family: exact F = 0, M > 0, chart drift W = 0, radial coefficient with an l = 1 part -- point (1/2, 1/3, -1/5): residual 0 True, M eigenvalues [0.322338, 0.074767, 0.116838]; point (2/3, -1/2, 1/4): residual 0 True, M eigenvalues [0.283054, 0.059569, 0.083753] P3 the general swept family, and its Szekeres subfamily [PASS] general swept family F = 0 (det M > 0); Szekeres subfamily A^2 = H^2/(1-k) identically, and F = 0 -- general: M eigenvalues [0.133625, 0.341712, 0.285426]; Szekeres: M eigenvalues [0.4, 0.455917, 0.455917] P4 route 1 of the second-order removal is the tilt [PASS] round base: A = 1 + eps cos(theta) + eps^2 [cot^2 r/6 + (cot^2 r/3) P_2] + O(eps^3) P5 first order, general alpha [PASS] delta A / A0 = Phi (Phi' W + alpha b).n / A0^2; with W = 0, delta(A^2) = 2 alpha Phi b.n -- Phi*(Phi_p*Wn + alpha*bn)/(Phi_p**2 + alpha**2) P6 the axisymmetric case, off the round base [PASS] axisymmetric tilt, W = 0, non-round base: F = 0 exactly, M > 0 -- M eigenvalues [0.377669, 0.414567, 0.411746] P7 each shell is intrinsically round [PASS] Gauss curvature of Phi^2 E^{-2}(dp^2 + dq^2) = 1/Phi^2 -- K = Phi**(-2) P8 the drift in shell coordinates [PASS] g_rr = A^2 + 2c'Phi' cos th + c'^2, g_{r th} = -c' Phi sin th, angular block round; round base beta = -tan r test_338: ALL PASS exit 0 [87s] ================================================================================================== test_339_second_order_round_base.py ================================================================================================== [PASS] P1 the O(eps^2) obstruction: the four displayed components, the rest zero, O(eps) zero -- r0 = atan(4/3), th0 = acos(3/5): C_rthr = -27/32; r0 = atan(4/3), th0 = acos(8/17): C_rthr = -3375/4624 [PASS] P2 route 1 (the tilt's second-order correction) removes every component through O(eps^2) [PASS] P3 route 2: O(eps) zero for every s [PASS] P3 route 2: C_{r th r} = -(1 - s)^2 2 sin th cos th / sin^4 r, double root at s = 1 [PASS] P3 route 2: s = 1, mu = sec^2 r: every component vanishes [PASS] P3 route 2: the O(eps^2) residual does not depend on mu (sec^2 r against 0, s = 0, 2) [31s] test_339: ALL PASS exit 0 [32s] ================================================================================================== test_340_second_order_general_bases.py ================================================================================================== base-radius pairs on the positive branch: 8; constant-ratio configurations: 6 mode F done [173s] mode C done [340s] [PASS] P1 (F) O(eps) zero for every s; f_X(s) quadratic [PASS] P2 (F) f_X(1) = 0 for all four components, 8 base-radius pairs and three angular points [PASS] P3' (F) all four families have a double root at s = 1, every configuration [PASS] P5' (F) constant-ratio bases: all four families double root at s = 1 [PASS] P1 (C) O(eps) zero for every s; f_X(s) quadratic [PASS] P2 (C) f_X(1) = 0 for all four components, 8 base-radius pairs and three angular points [PASS] P3 (C) (Lambda-2) families double root, D family simple root [PASS] P4' (C) the D family's second root differs between bases -- B2@1/2: 0.376941, B2@1/3: 0.442371, B2@3/4: 0.252259, B3@1/2: 0.617772, B3@3/4: 0.350832, B4@1/2: 0.285611, B4@1/3: 0.394797, B4@3/4: 0.105186 [PASS] P5 (C) constant-ratio bases: D slots simple root, (Lambda-2) slots double root, root at s = 1 P7 the two parametrisations: beta' enters only the D family [PASS] P7' (Lambda-2) components agree between (F) and (C); D components differ for s = 2, 3 (every configuration) P6 (C) the eps^2 radial companion does not enter [PASS] P6 (C) mu scaled by 2, 1/2, 0: every f_X(s) unchanged [348s] test_340: ALL PASS exit 0 [350s] ================================================================================================== test_341_polynomial_form_and_the_degenerate_symbol.py ================================================================================================== P1 the branch-free variable P [PASS] adj adj K = (det K) K; even in K; P = sqrt(det M) K~; P > 0 iff K~ definite P2 T = 2 (det M)^{3/2} F at a point, both branches [PASS] T = 2 (det M)^{3/2} F exactly, convex and indefinite branches -- M(0) = diag(1, 2, 3): equal True, F != 0 True; M(0) = diag(2, -1, -3): equal True, F != 0 True P3 the symbol as k3 -> 0 [PASS] P3' rank 2; sv_max ~ 1/k3 always; sv_min bounded iff xi is orthogonal to e_3, else ~ 1/k3 -- perp [1, 0, 0]: sv_max k3 -> 0.35355, sv_min -> 0.50000; perp [0, 1, 0]: sv_max k3 -> 0.08839, sv_min -> 0.25000; perp [3/5, 4/5, 0]: sv_max k3 -> 0.17337, sv_min -> 0.35043; oblique [0, 0, 1]: sv_max k3 -> 0.59293, sv_min k3 -> 0.50000; oblique [2/3, 1/3, 2/3]: sv_max k3 -> 0.41962, sv_min k3 -> 0.28112; oblique [12/13, 0, 5/13]: sv_max k3 -> 0.38666, sv_min k3 -> 0.14569 test_341: ALL PASS exit 0 [5s] ================================================================================================== test_342_rank_one_branch.py ================================================================================================== [PASS] Q1 adj diag(k1, k2, 0) = diag(0, 0, k1 k2): rank one [PASS] Q1 adj M = 0 on the branch (adj M vanishes) [PASS] Q1 M's nonzero eigenvector spans ker K (the ruling) [PASS] Q2 along v: the Bianchi component is exactly div(mu v) [PASS] Q2 across v: the components are mu times the acceleration of the ruling [PASS] P1 Codazzi with v: b21 k11 + k12 (b22 - b11) - b12 k22 = 0 (-b11*k12 - b12*k22 + b21*k11 + b22*k12) [PASS] P1 equivalently B K = K B^T (the diagonal is identically zero) [PASS] P1 principal frame: k2 b12 = k1 b21 (-b12*k2 + b21*k1) [PASS] P2 k1 = k2 = k forces b12 = b21 (no twist) [PASS] P2 a pure rotation of the rulings forces tr K = 0 on the block (so mu = det < 0) (-c*(k11 + k22)) [PASS] P3 the pointwise system has a curve of solutions with Jacobian rank 2 (one free parameter) (4 solutions on distinct k12, ranks [2, 2, 2, 2]) [PASS] P4 transport + Jacobi: v(ln det K) = -tr B = -div v, i.e. nabla(mu v) = 0 (-b11 - b22) [PASS] P5 cone over a Clifford torus: N is the unit normal and K kills the ruling d_t [PASS] P5 the block is diag(tan al, -cot al)/t and B = I/t (B [1/t, 0, 0, 1/t]) [PASS] P5 frame parallel along the ruling, and d/dt block = -B K exactly [PASS] P5 cone class B = f I: P1's equation is empty there [PASS] P6 where det M > 0, sqrt(det M) M^-1 = sign(det K) K (20 random symmetric K) [PASS] P6 across a seam K~ flips sign; sign(det K) K~ = K is continuous, so K is fixed there (K~_22: -2 | 2; K_22: [2, 2]) test_342: ALL PASS exit 0 [3s] ================================================================================================== test_343_chiral_form.py ================================================================================================== P1 so(4) = su(2)+ (+) su(2)- [PASS] each half closes and has dimension 3; the halves commute and sum to so(4) P2 Theorem 9.1 on a generic metric and K [PASS] curvature of each half = projection of the curvature of Omega; blocks = Gauss and Codazzi residuals P3 a realizable hypersurface, and a scaled K [PASS] paraboloid in R^4: both halves flat; 1.1 K: Gauss block nonzero, Codazzi block zero, both halves curved P4 Corollary 9.2: K = K~(g) on a non-realizable metric [PASS] Gauss block 0; Codazzi block = Cod K~ != 0; the halves are equal and opposite (su(2) labels opposite) -- Gauss 0 True, Codazzi != 0 True, = Cod K~ True, opposite True P5 Proposition 9.3: reversing the normal [PASS] K -> -K: Gauss block unchanged, Codazzi block reversed, Omega -> R Omega R, the halves exchanged test_343: ALL PASS exit 0 [8s] ================================================================================================== test_344_second_read_checks.py ================================================================================================== P1 the chart's rotation [PASS] d_r n = -W + (W.n) n + rho_c x n; the trajectory condition holds with rho = (u2', -u1', 0)/u3, not with +rho_c P2 the shell variation, coordinate components [PASS] the five printed components are the coordinate components; the rest vanish; lowered M_1 is symmetric -- (M_1)_r_theta = -Derivative(Phi(r), r)*Derivative(Y(theta), theta)/Phi(r) P3 T and R vanish at a point where rank M <= 1 [PASS] P(0) = 0, d det M(0) = 0, T(0) = R(0) = 0 at rank 1 and rank 0 (rank >= 2 nearby); nonzero at rank 2 -- rank 1: silent; rank 0: silent; rank 2 control: nonzero P4 the flag bounds from the linearised immersions [PASS] free of the first coordinate: rank 19 = 3*5 + 4; of the first two: rank 4 = 3 + 1 (injective), at four K -- diag(1, 2, 3): 19/19, 4/4; diag(2, 3, -3): 19/19, 4/4; diag(1, 1, 1): 19/19, 4/4; diag(1, 1, -2): 19/19, 4/4 test_344: ALL PASS exit 0 [48s] ================================================================================================== test_354_symbol_module_every_K_and_general_bases.py ================================================================================================== P1 the identities behind Theorem C(3) [PASS] Einstein symbol on K = -D adj_K(xi xi)/2; D adj and its inverse; the omega and Hessian potentials P2 g_3 = M_K^3 element by element [PASS] every element of an exact basis of g_3 is sym-grad(omega) + phi K, at diag(2,3,-3) and a non-diagonal K P3 Cartan's test at shape operators not among the sixteen [PASS] dim g_3, g_{3,1}, g_{3,2}, g_4 = 55, 19, 4, 78 -- [55, 19, 4, 78]; [55, 19, 4, 78]; [55, 19, 4, 78] P4 the fifteen functionals of the left null space of tau_4 [PASS] twelve prolonged algebraic identities and three cyclic symbols: rank 15 -- 15, 15, 15, 15 P5 Theorem 7.7: the combined drift and tilt [PASS] radial coefficient, cross term and difference from the target, exactly P6 scope examples [PASS] (a) local Helmholtz tensor; (b) R = 146/25 for the tilt; (c) TT kernel at a characteristic covector for all t; (d) scalar B leaves two parameters -- (a) True, (b) R = 146/25, (c) True, (d) True test_354: ALL PASS exit 0 [9s] ================================================================================================== test_625_dF_equals_dC.py ================================================================================================== [PASS] P1 F(gamma) = C(gamma) = 0, and DF_gamma(H) = DC_gamma(H) for general H at two rational points (e.g. (0,1,2): DF = 982939805/99574272, DC = 982939805/99574272) [PASS] P2 controls: both sides vanish on 2u gamma and on L_X gamma ALL PASS exit 0 [39s] ================================================================================================== test_626_rivertz_dualisation.py ================================================================================================== [PASS] P1 R_ab^d = eps_cab W^cd, and eps_ecd W^cd = R_de^d: the trace identity is exactly W's skew part [PASS] P2 eps^abc (cyclic sum) = 6 tr(W P), and the cyclic sum is totally antisymmetric (one component) ALL PASS exit 0 [0s] ================================================================================================== test_627_zeta_tt_count.py ================================================================================================== [PASS] P1 TT: sum_{n>=3} 2(n^2 - 4) regularised = +10 (10.0) [PASS] P2 im K exact part: sum_{n>=3} n^2 regularised = -5 (-5.0) [PASS] P3 im K coexact part: sum_{m>=3} 2(m^2 - 1) regularised = -5; im K = -10; Sym^2_0 = TT + im K = 0 (-5.0) [PASS] P4 scalar shape modes n >= 3: -5 = -dim ker(Obata's trace-free Hessian); all scalars and all 1-forms count 0 (scalars 0.0, 1-forms 0.0) [PASS] P5 the TT count is +10 for lambda = n^2 - 1, n(n^2 - 1), n^2 (n^2 - 1)^2/2 (continuation in s); control: an odd power, lambda = n(n + 1), shifts it (10.0, 10.0, 10.0; control 13.6666666667) ALL PASS exit 0 [0s] ================================================================================================== test_629_branson_intertwiner.py ================================================================================================== [PASS] P1 on TT: curl preserves TT, curl^2 = nabla*nabla + 3, eigenvalues +-n (n >= 3 integer), multiplicities n^2 - 4 each sign (n <= 7) (max |eig - integer| = 2.0e-14; weighted counts 3: 5/5, 4: 12/12, 5: 21/21, 6: 32/32, 7: 45/45) [PASS] P2 DC_gamma(Omega^-2 Phi^* H) = Phi^* DC_gamma(H) for the Moebius dilation y -> 2y (exact, two points) ALL PASS exit 0 [42s] ================================================================================================== test_640_rivertz_span.py ================================================================================================== [PASS] P1 controls: W = r_1 = r_2 = 0 on hypersurface jets; W, r_1 != 0 when beta is not totally symmetric [PASS] P2 r_1 and r_2 lie in the span of the six W^cd (rank 6) (rank W = 6; r_1 in span: True, r_2 in span: True) [PASS] P3 the frame relabellings of r_1, r_2 span 6 dimensions, inside span(W): span(W) = span(r_1..r_6) (rank of images 6, with W 6) [PASS] P4 (tr M)^2 x Cotton-York: symmetric, same bidegree, not in span(W) (rank of W with it: 11) ALL PASS exit 0 [2s]