Search all 707 registered Python modules by subject, filename, claim or source section. This catalogue records source presence and links. It does not report a new execution or a passing result for every module.
707 matching modules · showing 341–360
Grav backreaction
v361_grav_backreaction.py · core
Read the registered purpose, outcome and evidence links
Direction 5: the matter-gravity backreaction with the explicit J3 flux -- two forced facts + an honest decline. [E] the carrier's gravitational anomaly is FORCED: c_-=8 (16 Majoranas=dim S+=2^(g_car-1); chiral central charge 16/2=8=g_car+N_fam, the same c_- as v358). [E] the backreaction is FINITE -> Lambda from alpha: the matter vacuum energy through G_ab=(1/c3)T_ab is gap-suppressed (Decoupling Thm v337) to rho_Lambda/Mbar^4=(3/(4 pi^2))e^{-2 alpha^-1} (v60), 123 orders = 119.03+3.92 -- NOT the M_Pl^4 catastrophe; the loop (v359+v60+carrier vacuum) closes. [E] J3 explicit (v358): delta<K_B>=(8 pi^2 R^4/15) delta<T_00>, carrier T_ab=the SM's (v159). [C] SM trace anomaly reproduced (content=SM). [O] DECLINE: no new atom-forced read-out of individual SM quantum numbers beyond c_-=8 and Lambda -> per v354/v355 declined. Finite & consistent backreaction, no new dial.
Read the original Python source
Grav gap corrections
v360_grav_gap_corrections.py · core
Read the registered purpose, outcome and evidence links
Direction 2: the gap-induced higher-curvature correction to GR -- the disciplined result. Does the seam transfer gap give a NEW forced near-term GR-deviation beyond the known R^2 scalaron? With the v354/v355 discriminator: NO. [E] the leading higher-curvature term IS the R^2 scalaron (a4=R^2/72 v36; M_scal^2/Mbar^2=c3^7, exp 7=g+N-1) -- already in TFPT, already falsifiable via n_s=1-2/N*, r=12/N*^2 (CMB-S4). [C] it is the entanglement-equilibrium next order (Wald first law -> f(R)=R+R^2, v28/v359). [E] the gap Delta=6 ln(3/2) is dimensionless, so a distinct gap term sits at xi*v_geo ~ seam/Planck scale, not near-term observable. [O] DECLINE: a new dimensionless coefficient is not atom-forced (needs heat-kernel/Wald data; exp 7 vs 6, c3^7 vs (2/3)^6 unrelated) -> per v354/v355 declined. No new near-term forced prediction; the answer is the scalaron.
Read the original Python source
Grav nonlinear einstein
v359_grav_nonlinear_einstein.py · core
Read the registered purpose, outcome and evidence links
The FULL covariant Einstein equation, parameter-free (Direction 1, extends v358). The Jacobson-2015 fixed-VOLUME entanglement equilibrium brings in the EINSTEIN TENSOR G_ab (not just R_ab) -- the unique divergence-free metric tensor (Lovelock) -- so equilibrium yields the full covariant G_ab + Lambda g_ab = (1/c3) T_ab with conservation built in. [C/E] fixed-volume -> Einstein tensor (trace=-R in d=4). [E] Lovelock => nabla^a T_ab=0 is an output. [E] coeff 1: 8pi=1/c3 fixed (v358). [E] coeff 2: Lambda from alpha -- rho_Lambda/Mbar^4=(3/(4pi^2))e^{-2 alpha^-1}, prefactor (8pi)^2*48c3^4=3/(4pi^2) (v60); BOTH coefficients TFPT-fixed. RESULT: the original B6 full covariant field equation is now parameter-free at the LOCAL level. [O] residual: the equation-of-state status + the global ambient measure (C7).
Read the original Python source
Grav entropy equilibrium
v358_grav_entropy_equilibrium.py · core
Read the registered purpose, outcome and evidence links
The entanglement-equilibrium derivation (Jacobson 2015 / Faulkner et al., delta S = delta<K>) with TFPT's atoms -> the LINEARISED Einstein equation falls out PARAMETER-FREE, and the two formerly-open pieces are closed. [E] parameter-free Einstein coeff G_ab=(1/c3)T_ab, 1/c3=8pi fixed (Unruh 1/(2pi)=4c3, EH 1/(16pi)=c3/2, Einstein 8pi=1/c3, Bekenstein 1/4=1/|mu4|). [E] THERMO=GEO coincidence (new): the first law's 2pi/eta = the geometric |Z2|2pi chi iff |mu4|=|Z2|chi=4 -- the thermodynamic and geometric origins of c3 are ONE. [C] J3 solved: the CHM ball modular Hamiltonian (boost via BW v323) gives the matter flux delta<K_B>=8pi^2 R^4/15 delta<T_00>. [E] entropy density atom-fixed: 1/4=1/|mu4| + central charge c=g_car+N_fam=8 (form v59). [O] residual: the equation-of-state status + the global ambient measure (C7) + the v_geo unit (v359 closes the FULL covariant equation); no free dimensionless gravity dial remains.
Read the original Python source
Continuum mmst applicability
v356_continuum_mmst_applicability.py · core
Read the registered purpose, outcome and evidence links
Direction A: the continuum chain of SEAM.EQUIV.01 assembled theorem-by-theorem, residual reduced to ONE lattice-realization input tied to the seam's one-sidedness. Chain: gapped (v302) quasi-free 16-Majorana collar (v155/v160) -> invertible (Kitaev, v301) -> the NONTRIVIAL invertible phase (c_-=8) -> a non-gappable anomalous chiral edge (anomaly inflow: edge EXISTENCE follows, reducing the v336/v351 'does the edge exist?' residual) -> chiral-CFT scaling limit (MMST v336, 8<=8<=16) -> order-4 mu4 clock + AQFT stack pin (E8)_1 (v351/v297). The only non-theorem link is the seam one-sidedness = the c3 origin (c_-!=0 = one-sided = the |Z2| oint K = 8 of c3=1/(8pi), v58). [C]/[E]/[O]; ADVANCES SEAM.EQUIV.01 to one realization input, does not close. (Direction B was found already closed by v40.)
Read the original Python source
E8 unmapped bandwidth
v355_e8_unmapped_bandwidth.py · registry
Read the registered purpose, outcome and evidence links
A disciplined bandwidth search of the unmapped E8 region (next step after v354), run with a STRICT discriminator: a candidate counts only if it is a FORCED identity (a theorem about E8), never a numerical coincidence -- because mining a rich object for matches is the numerology temptation. [E] FORCED SET-LEVEL (the real collective content): sum(degrees)=128=dim S+ the spinor half of 248=120+128; sum(exponents)=120=#pos roots; product(degrees)=|W(E8)|; exponents = the phi(30)=8 totatives of 30 (v223); max(degree)=30=h(E8). [C] SUGGESTIVE not forced: 12=h(E6),18=h(E7),30=h(E8) (the McKay ladder v219), but 'deg(E8) contains subalgebra Coxeter numbers' is not a theorem -- flagged. [O] DECLINED: the per-degree atom matches (14=dim G2, 20=det L, 24=|W(A3)|, 12=|R(A3)|, 18=p4(a); v66 flagged 12,18 fittable) each admit multiple readings, and the |W| prime 7='scalaron exponent' is one of many -- declined. [E] RECONCILES v66 (degrees=atoms) and v354 (only 2,8,30 primary): NO new forced physical hit; the disciplined decline IS the anti-numerology result.
Read the original Python source
E8 reverse audit
v354_e8_reverse_audit.py · registry
Read the registered purpose, outcome and evidence links
The REVERSE numerology audit + what the golden ratio means. The forward rule (v305) is one-directional (every readout must map onto E8); the reverse question: how much E8 structure maps onto NOTHING? [E] E8 has 8 Casimir invariant degrees {2,8,12,14,18,20,24,30}; THREE feed a readout (2 metric, 8 rank->c3, 30 Coxeter->g_car) and FIVE are UNMAPPED {12,14,18,20,24} -- 3/8 load-bearing, 5/8 not. [E] ECONOMY, NOT PROMISCUITY: a small fixed set (rank 8, h=30, det 1, D5(+)A3, 16-spinor) generates ALL readouts; the higher Casimirs are never reached into to fit -- the anti-numerology signature (economy + a large untouched remainder). [E] the golden ratio is UNMAPPED (the affine-E8 attractor eigenvalue v312, the icosian lattice v348; no FROZEN readout is phi -- the sole [C] exception, cos(theta_i=3pi/5)=-1/(2phi) on the axion spine branch, refined in v429) -> NOT numerology (which needs a number fitted to an observation); phi is the algebraic signature of the 5-fold h=30=2*3*5, the carrier announcing it is icosahedral. [O] the 5/8 unmapped is the HULL OVERHEAD (E8 = minimal unimodular hull, not the world group); where future readouts could live, or over-capacity. The reverse complement of v305.
Read the original Python source
Selfloop capstone
v353_selfloop_capstone.py · core
Read the registered purpose, outcome and evidence links
The bird's-eye rethink (refining v352): TFPT is not a LINEAR theory (axioms -> theorems) but a CLOSED SELF-CONSISTENT LOOP whose outputs fix its own inputs, with ZERO free adjustable dimensionless parameters. [E] IT IS A LOOP: outputs fix inputs (g_car=5 = max prime of h(E8)=30; the 8 in c3 = rank E8) and inputs fix outputs; the arrows close into a circle (v6) with a unique fixed point (v56). [E] ZERO FREE DIALS: every load-bearing number is forced/over-determined; free dimensionless parameters = 0. [E] pi is a FIXED geometric constant (the boundary 2-sphere, oint K = 4pi -> c3=1/(8pi)), not a dial. [A] refining v352: pi is not a free input and the framework is a SELECTION PRINCIPLE (the unique self-consistent loop), not a free meta-axiom. [O] the one residual is the CONTINUUM closure (the chiral-edge existence, v336/v351); structural-realism (loop complete) vs constructivism (continuum open) is a philosophical, not numerical, fork.
Read the original Python source
Framework irreducible
v352_framework_irreducible.py · core
Read the registered purpose, outcome and evidence links
The framework reduction -- the deepest answer to 'the inputs are not axioms, they are back-determined'. [E] P2 (g_car=5) FORCED three ways (v6): rank-fill, Coxeter-match (g_car=max prime of h(E8)=30), Pascal. [E] P1 (c3=1/(8pi)) FORCED: c3 is the boundary Gauss-Bonnet coefficient (v216/v342), the 8 over-determined four ways (8=rank E8=h(D5)=phi(30)=det R, v54/v6); only pi is non-forced (the boundary transcendental). [E] the E8 hull is forced (unique even unimodular rank-8, v83/v154). [E] so the ONLY irreducibles are the FRAMEWORK (a discrete reflection-positive boundary with a finite Clifford carrier) + pi -- the precise content of 'the inputs are not axioms'. [O] the framework is the META-AXIOM (not math-reducible, the foundational stance, tested by its consequences not proved). TFPT's true input is ONE framework + pi.
Read the original Python source
Bootstrap inputs correction
v350_bootstrap_inputs_correction.py · core
Read the registered purpose, outcome and evidence links
The honest CORRECTION of v349's framing (prompted by the right objection: the inputs are not axioms, they are back-determined by the theory). [E] THE INPUTS ARE BOOTSTRAP-FORCED, NOT FREE AXIOMS (v6): g_car=5 over-determined three ways (rank-fill, Coxeter-match g_car=max prime of h(E8)=30, Pascal). [E] g_car=5 IS THE ICOSAHEDRAL 5-FOLD, not the D5 rank: g_car = max prime of h(E8)=30=2*3*5; the atoms (2,3,5) ARE the primes of the icosahedral Coxeter number 30. [E] THE GOLDEN RATIO IS EMERGENT FROM THE GLUE, not external: D5 (h=8)/A3 (h=4) non-golden alone, but the mu4 extension D5+A3 -> E8 (v154) gives h(E8)=30 (golden, 5|30) -- the seam's own mu4 produces the 5-fold; v349 checked the un-glued pieces, the wrong object. [E] v349's pentagon-vs-count dichotomy was FALSE (the bootstrap forces the pentagon). [O] residual relocates: the golden is bootstrap-forced; the only open thing is the physical continuum realisation (v336), NOT the golden ratio. Caveat: self-consistency within the framework, not from-nothing. Supersedes v349's interpretation.
Read the original Python source
Raw seam golden test
v349_raw_seam_golden_test.py · core
Read the registered purpose, outcome and evidence links
The honest test of the v348 question 'does the RAW seam carry the golden ratio (before assuming E8)?'. Answer, computed: NO -- the bare carrier (D5, A3) and the dynamical seam data are NOT golden; the golden ratio genuinely enters WITH the icosahedral/E8 identification. A NEGATIVE result that prevents a false closure. [E] RAW CARRIER NOT GOLDEN: golden needs 5|h; D5 has h=8 (2cos(2pi/8)=sqrt2), A3 has h=4 -- no 5-fold, bare carrier gives sqrt2. [E] DYNAMICAL DATA NOT GOLDEN: cusp {0,1/3,2/3} rational, recovery (2/3)^6 rational, seed phi0 transcendental. [E] GOLDEN IS THE ICOSAHEDRAL/E8 SIDE: 5|h only for A4=SU(5) (h=5), H3 (h=10), E8 (h=30). [E] phi alone insufficient: 2I != C5 x mu4 (20 != 120). [O] RESOLUTION: L2 reduces to 'is g_car=5 a PENTAGON (5-fold => golden => icosahedral) or a COUNT (rank D5 => sqrt2)?' -- the golden ratio is the pentagon-reading of P2, NOT derivable from P2-as-a-count; no hidden phi. Foundational: P2-as-pentagon closes it (mode C) else rigidity must produce the 5-fold (mode B); NOT closed.
Read the original Python source
Seam rigidity route
v348_seam_rigidity_route.py · core
Read the registered purpose, outcome and evidence links
Route B (the rigidity route) for the one open arrow L2, carried out -- it lands on a strikingly SIMPLE statement: the whole keystone reduces to ONE question, does the raw seam carry the golden ratio? [E] rigidity closes the crystallographic part: the order-4 clock has a UNIQUE normal form z->iz (Nielsen/Kerckhoff, v180), Troyanov gives the unique flat pillowcase metric (v284). [E] downstream is a LATTICE identity: E8 = the ring of icosians (Conway-Sloane), rank 8, so 2I->E8 is a lattice identity (stronger than McKay v219). [E] the single bridge is phi: crystallographic orders {1,2,3,4,6}, the 5-fold non-crystallographic; mu4 crystallographic, 2I (5-fold) not, phi=2cos(pi/5) extends one to the other (quasicrystal). [O] the keystone reduces to ONE question: 'does the raw seam carry phi?' -- if yes, mu4+phi -> 2I -> E8 closes L2. [O] honest subtlety: TFPT's phi is currently E8-side (v312/v313); the raw-seam data (cusp {0,1/3,2/3}, recovery (2/3)^6, seed phi0) are not manifestly golden, so showing the RAW seam carries phi non-circularly is the residual (= L2). NOT closed.
Read the original Python source
Seam closure modes
v347_seam_closure_modes.py · core
Read the registered purpose, outcome and evidence links
The closure-mode classification of the one open arrow L2 (the seam-orbifold realisation, v346) -- a direct honest answer to 'is there ANY other way to FULLY solve it?'. [E] THE PRECISE LOCUS: the seam gives the pillowcase base S^2(2,2,2,2), chi_orb=0 (EUCLIDEAN, v216), but the 2I/E8 structure is the Seifert base S^2(2,3,5), chi_orb=1/30 (SPHERICAL); L2 must bridge a FLAT 4-marked base to a SPHERICAL 3-marked base -- why L2 needs the carrier (2,3,5) input. [E] FOUR CLOSURE MODES: A construct (the analytic wall, v336); B rigidity (unique realisation => C^2/2I forced, the only theorem-route w/o new input, v284/v300/v331); C axiom P3 (=> TFPT complete rel. 3 axioms); D empirical (test the consequences). [E] no-new-state (tfpt_3/v246) blocks deriving L2 from extra physics => only rigidity (B) closes it as a theorem. [E] VERDICT: no free theorem; only B (theorem) or C (axiom), validated by D; only these four modes. [O] NOT closed.
Read the original Python source
Seam geometric bridge
v346_seam_geometric_bridge.py · core
Read the registered purpose, outcome and evidence links
The geometric bridge as an explicit six-link chain raw-seam => det K=1, with the ONE open arrow pinned -- the honest capstone of the SEAM.EQUIV.01 investigation. L1 [E/C] raw seam -> mu4 + four marks (v216); L2 [O] mu4 + (2,3,5) atoms -> 2I orbifold C^2/Gamma (THE OPEN ARROW = SEAM.EQUIV.01); L3 [E] 2I -> affine E8 McKay (v219); L4 [E] 2I -> E8 du Val (v232); L5 [E] C^2/2I -> Poincare link H_1=0 (v345); L6 [E] H_1=0 -> det K=1. [E] 5 of 6 links are theorems, only L2 open. [E] the group is FORCED by the seam atoms: axis signatures (2,3,3)/(2,3,4)/(2,3,5) -> 2T/2O/2I, (2,3,5) unique to 2I, and (|Z2|,N_fam,g_car)=(2,3,5) ARE those axes. [E] downstream of L2 forced. [O] L2 (the seam IS an ADE orbifold point) = SEAM.EQUIV.01, NOT closed -- the keystone reduced to one geometric-realisation arrow, group pinned by (2,3,5).
Read the original Python source
Hypergraph homotopy
v345_hypergraph_homotopy.py · core
Read the registered purpose, outcome and evidence links
EXECUTE the combinatorial core of route R3 (the hypergraph-homotopy route to det K=1, v344): computes explicitly that the (2,3,5)-rewrite attractor graph's plumbing link is the POINCARE homology sphere (H_1=0), only E8 among ADE; R3's combinatorial half is now [E]-complete, the geometric bridge (= SEAM.EQUIV.01) stays [O]. Plumbing topology (Brieskorn): link H_1 = coker(Cartan). [E] SNF(E8 Cartan) = identity => coker(E8)=0 => the E8 plumbing boundary is an INTEGRAL HOMOLOGY SPHERE (the genus-1 meaning of det E8=1). [E] only E8: coker(A_n)=Z/(n+1), D_n->Z/4, E6->Z/3, E7->Z/2 -- all nontrivial, only E8=0. [E] pi_1=2I=SL(2,5) PERFECT (commutator subgroup order 120=|G|, computed) => H_1=0 => THE Poincare sphere; |2I|=120=|mu4|*h(E8). [O] does NOT close det K=1: the graph->geometry half is meanwhile Kronheimer-cited (v479); what stays open is the raw seam supplying the ALE/orbifold datum = SEAM.EQUIV.01.
Read the original Python source
Detk synthesis
v344_detk_synthesis.py · core
Read the registered purpose, outcome and evidence links
The bird's-eye synthesis of the ONE open keystone bit, det K=1 -- a full-spectrum map of SEAM.EQUIV.01's only open residual. [E] det K=1 is ONE statement with SIX equivalent faces (holomorphy / homology-sphere link H_1=0 / one 1-dim irrep / det Cartan=1 / full mu4 condensation / tau=i CM), unified by |det Cartan| = |H_1(link)| = #1-dim irreps across ADE (A_n->n+1, D_n->4, E6->3, E7->2, E8->1) -- ONLY E8 gives 1. [E] THE GENUS-1 OBSTRUCTION: the easy arguments are genus-0, necessary but not sufficient (v87); det K is the torus degeneracy. [E] THREE genus-1 access routes: R1 continuum-modular (v336), R2 anyon-condensation (v281), R3 hypergraph-homotopy. [C] the HYPERGRAPH route is the most promising fresh angle: the (2,3,5) attractor is the affine-E8 graph (v312), the E8 du Val link is the Poincare sphere (v232), so det K=1 becomes COMBINATORIAL. [O] none closes it; R3's combinatorial-to-geometric bridge is meanwhile discharged to Kronheimer 1989 (v479), so the residual transforms to 'the raw seam supplies the ALE/orbifold datum' = SEAM.EQUIV.01.
Read the original Python source
Four routes analysis
v343_four_routes_analysis.py · frontier
Read the registered purpose, outcome and evidence links
The honest investigation of the four black-hole-birth solution routes (A finite causal diamond, B Carlip-Cardy near-horizon CFT, C modular/thermal flow, D self-reproduction attractor) -- direct answer to 'which closes ALL problems and is 100% provable': NONE individually, but they CONVERGE on the same two irreducible residuals as v335 (SEAM.EQUIV.01 + the decoupled QG contour). [E] ROUTE A reframes QG.AMB to a FINITE thermal trace (de Sitter finite, gapped chi=729/665) but c_conf(4)=-3/2<0 still needs the GHP contour (v334 [O]); does NOT touch SEAM.EQUIV.01. [C] ROUTE B: Carlip+Cardy consistent with c=8 but L0 free (ambiguity) and gives c only not holomorphy det K=1 -- inherits SEAM.EQUIV.01. [E] ROUTE C: modular flow KMS beta=1 (v239), geometric identification = SEAM.EQUIV.01 (v335) -- a reduction. [E]/[P] ROUTE D: unique Perron-Frobenius attractor (v56) but cyclic dynamics [P] -- not a closure. [E] VERDICT: independent closures = 0; irreducible residuals = 2 (= v335); BH premise makes both more natural but eliminates neither; focus = SEAM.EQUIV.01.
Read the original Python source
Em ward heatkernel
v342_em_ward_heatkernel.py · em
Read the registered purpose, outcome and evidence links
The heat-kernel sharpening of EM.WARD.01: DERIVES the ORIGIN and STRUCTURE of the F_{U(1)} determinant-line terms from textbook Seeley-DeWitt/Gilkey coefficients, advancing the EM-Ward origin from [P] to [E] for the STRUCTURE; NO new gate. [E] c3=1/(8pi) the boundary GAUSS-BONNET coefficient (seam = flat pillowcase chi=2, oint K = 4pi, v216). [E] the alpha^2 term is the GILKEY gauge-curvature coeff: a4 has 30 Omega^2/360 = (1/12)Omega^2, Omega=i alpha F => -(alpha^2/12)F^2, so the Calderon -2 c3^3 a^2 is the gauge-curvature piece, not an ansatz. [E] c3-ladder {0,3,6} arithmetic (3 channels; C6 hexagon at 6). [E] multiplicities 2=|Z2|, 8 b1=(rank E8)b1. [E] pi-power test: 2 c3^3 ~ pi^3 => THREE boundary insertions. [C] exact coefficient needs the seam F-normalisation. [O] the cubic alpha^3 + full identification = EM.WARD.01 residual (tied to SEAM.EQUIV.01).
Read the original Python source
Alpha quillen
v341_alpha_quillen.py · em
Read the registered purpose, outcome and evidence links
The alpha fixed point reformulated as the STATIONARITY of a U(1) determinant line, every ingredient identified as an index / heat-kernel coefficient / discriminant form -- NOT a free knob (the 'derive the FORM' step). F_{U(1)} = a^3 - 2 c3^3 a^2 - (4/5) c3^6 M ln(1/phi_seam) = 0 in Quillen form: [a^3 - 2 c3^3 a^2] (det-line variation) = [8 b1 c3^6] (anomaly counterterm) * [ln(1/phi_seam)] (seam response). [E] M=41 a U(1) INDEX (EM Ward, v48): (4/5)M = 8 b1, b1=41/10 the SM 1-loop U(1)_Y beta. [E] c3=1/(8pi) GAUSS-BONNET boundary coeff (8=rank E8). [E] -5/4 = -q(D5) DISCRIMINANT; phi_base=(4/3)c3, c3/phi_base=3/4=q(A3). [E] delta_top=Omega_adm c3^4. [E] Quillen split holds at the alpha^-1=137.036 root. [E] NO NEW GATE: alpha stays [E]; v341 SHARPENS the pre-existing [P] EM.WARD.01 (no new alpha problem). [O] the one still-open part IS EM.WARD.01's residual: the AQFT proof that this is the EXACT Quillen functional (variational principle derived, not assembled). Sharpens v3/v48/EM.WARD.01 from fixed point to stationarity of a named determinant line.
Read the original Python source
Qg oneloop saddle
v365_qg_oneloop_saddle.py · core
Read the registered purpose, outcome and evidence links
Finite saddle diagnostics [E]/[C]/[O]. The scalar stiffness is 1/c3=8pi and the gap margin is 6 ln(3/2)-31/(4pi^2)>0. On the stated two positive OS modes, Tr' H=-log det' T=6 ln(9/2). If the fluctuation operator M is identified with this H, log det' M=log[36 ln(3/2) ln3]; stiffness, base normalization and original relative/ghost factors remain. The GHP Gaussian integral is conditional. chi=729/665 is a finite susceptibility, not a geometric tightness or projective limit construction. No whole common field measure or primitive marked value is evaluated. The separate conditional IR/ambient redundancy route is retained.
Read the original Python source