MMA-DMF20 — Tools and Files Used in the Tests Generated (UTC): 2025-11-09 05:07:08Z Scope ----- This document enumerates the tools, inputs, and outputs employed across the validation pipeline (BAO rs-free, z3+Window, Lyα 10-point, nulls/stability sweeps) to demonstrate consistency without a dark-matter sector under MMA-DMF20. I) Model, Equations, and Configuration -------------------------------------- • Model set: MMA-DMF20 (Poisson/Weyl-only mu(a,k), Euler=GR, no slip; early-X bump in background only). • Early-X handling: injected via the microphysics mapping layer as a background-only, pre-drag deformation of H(a); it shifts r_d while preserving theta* and Delta N_eff ~ 0; no changes to Euler or baryon sound speed. • Primary equation pack: - MMA_DMF20.pdf (master specification) - Equations069d.zip (archived prior set; referenced for step-by-step BAO/CDM contrasts) • Config / activation presets: - activation_nocdm_vA_focus_A3.ini - activation_nocdm_vB_balanced.ini - activation_nocdm_vC_conservative.ini II) Core Pipelines (Guards and Criteria) ---------------------------------------- • BAO rs-free mapping (geometry only): alpha_parallel = (D_H/r_d) / (D_H/r_d)_fid alpha_perp = (D_M/r_d) / (D_M/r_d)_fid • z3+Window guard (survey-window robustness): - Input W(200x2000) -> square W_sq(200x200) via W @ pinv(W) @ W - Whitening (SVD): W_sq = U S V^T -> X = U V^T - Leakage metric: leak = || X X^T - I ||_F / sqrt(n); guard=OK if leak <= 0.25 - Verdicts: Excellent <= 0.15, Good <= 0.25, else Poor • Lyα 10-point (AP geometry): Delta chi^2 = chi^2(alpha ~ 1) - min_alpha chi^2 Thresholds: OK <= 2.30; Tense <= 6.17; Reject > 6.17 • Reinforcement suite: - Null tests (permutations) for Lyα series and W-column permutations - Stability sweep: rankQ x gamma; whitening (SVD) vs Cholesky(C_pre) - Variant check: C_post in [1000, 997, ...] bookkeeping; Lyα grade micro-sweep III) External Data and Matrices (BOSS/eBOSS/DR12/DR16) ------------------------------------------------------ A) BAO DR12 — spectra, windows, covariances • Spectra (post-recon): - Beutleretal_pk_monopole_DR12_NGC_z3_postrecon_200.dat - Beutleretal_pk_monopole_DR12_SGC_z3_postrecon_200.dat - (Quadrupole files if available; not required for rs-free) • Window matrices (z3 V6C): - W_BOSS_DR12_NGC_z3_V6C_1_1_1_1_1_10_200_2000_averaged_v1.matrix.gz - W_BOSS_DR12_SGC_z3_V6C_1_1_1_1_1_10_200_2000_averaged_v1.matrix.gz - M_BOSS_DR12_[NGC|SGC]_z3_V6C_1_1_1_1_1_1200_2000.matrix.gz (fallback when W missing) • Covariances: - C_1000_BOSS_DR12_SGC_z3_V6C_1_1_1_1_1_10_200_200_postrecon.matrix.gz - C_997_BOSS_DR12_NGC_z3_V6C_1_1_1_1_1_10_200_200_postrecon.matrix.gz - C_2048_BOSS_DR12_[NGC|SGC]_z3_V6C_*_prerecon.matrix.gz (for Cholesky fallback) B) eBOSS / DR16 — Lyα and QSO windows • W_eBOSS_DR16_QSO_NGC_1_1_1_1_1_10_200_2000_averaged_v1.matrix.gz • W_eBOSS_DR16_QSO_SGC_1_1_1_1_1_10_200_2000_averaged_v1.matrix.gz C) Lyα chi^2 grids / tracks (DR16) • sdss_DR16_LyAUTO_bao_chi2.dat • sdss_DR16_LYxQSO_bao_chi2.dat (aliases accepted during ingestion: *.dat.txt; case-insensitive normalization) D) Optional/ancillary • likelihood_data_PLANCK2018.zip (not used directly in geometric runs) • mcconnachie2012.pdf (bibliographic; not part of pipelines) IV) Produced Artifacts / Bundles (by stage) ------------------------------------------- • BAO (rs-free) bundle: - MMA_DMF20_BAO_DR12_rsfree_bundle_FINAL.zip Contents: PNGs (fits, residuals), CSV/JSON summaries, CALC.txt, CONCLUSIONS.txt • BAO (z3+Window) bundle: - MMA_DMF20_BAO_DR12_z3_window_bundle_FINAL.zip Contents: z3_window_summary.csv, heatmaps (if sweep), CALC.txt, CONCLUSIONS.txt, PNGs • Lyα 10-point bundle: - MMA_DMF20_LYA_proof_bundle_FINAL.zip Contents: LYA_10points.csv, plots (Delta chi^2 bars/grade sweep), CALC.txt, CONCLUSIONS.txt • Reinforcements: - MMA_DMF20_REFORCOS_FINAL.zip Contents: 01_selo1/ (clean re-run), 02_selo2/ (nulls), 03_selo3/ (sweep), 04_variants/ (C_post variants, grade micro-sweep), README, CALC.txt • Combined deliveries (where provided): - MMA_DMF20_RESULTS_FINAL.zip / MMA_DMF20_RESULTS_FINAL_BUNDLE.zip - MMA_DMF20_DELIVERY_*.zip (aggregator with multiple FINAL bundles) V) Intermediate and Summary Files --------------------------------- • z3_window_summary.csv • rsfree_summary.csv • lya_10p_summary.csv • nulltest_summary.csv • sweep_3x3.csv (or stability_sweep.csv) • resumo_global.csv (global index) • CALC.txt (per test; equations, thresholds, inputs, numeric summary) • CONCLUSIONS.txt (per test; verdicts: "Poor/Good/Excellent") VI) Tools / Environment ----------------------- • Python 3.x: numpy, pandas, matplotlib, gzip, json, zipfile, pathlib • Linear algebra: SVD (whitening), pseudo-inverse for W->W_sq • Integrity: deterministic seeds for permutation tests; structured CSV/JSON outputs • Notebooks/Cells: Colab cells A–E (setup, ingestion, guards, fits, packaging) VII) HTML / Documentation Utilities ----------------------------------- • Model comparison (AIC/BIC/LOOIC/WAIC): MMA_DMF20_model_comparison.html • Prefilled comparison: MMA_DMF20_model_comparison_prefilled.html • Long-form rationale: MMA_DMF20_NoDM_LongSummary.txt • Overleaf patch (addendum, macros, CALC template): mma_dmf20_overleaf_patch_v1.zip VIII) Expected Bundle Layout (per test) --------------------------------------- __bundle_FINAL.zip ├─ figures/ (all PNGs) ├─ tables/ (*.csv, *.json) ├─ CALC.txt (equations + thresholds + file list + numerics) ├─ CONCLUSIONS.txt (verdicts incl. "Poor/Good/Excellent") └─ README.txt (short how-to, versions, hashes if provided) Notes ----- 1) Early-X is always injected at the equation layer as a background-only, pre-drag deformation. It does not alter Euler's equation nor introduce slip; it only shifts r_d consistently. 2) For window robustness, prefer W; if only M is available, compress to 200x2000 (bin-averaging) and document. 3) Guards must be reported alongside results (leak, gamma, Delta chi^2 thresholds, etc.) in CALC.txt. End of document.