All SDGFT implementations, analysis pipelines, and datasets are open-source under the MIT license. Hosted on github.com/cosmologicmind.
Install the core library and compute all 89+ observables in three lines:
pip install sdgft
# Compute all observables from two axioms
from sdgft import Registry
reg = Registry()
reg.compute_all()
print(reg.scorecard()) # → 89+ observables, 0 free parametersRequires Python 3.10+. Full documentation in the repository README.
The primary SDGFT packages — from axioms to observables.
Core library — 89+ observables derived from $\Delta = 5/24$ and $\delta = 1/24$. Full derivation chain, observable registry, and 1305 unit tests. Zero free parameters.
Parameter-free cosmological predictions: CMB spectra, Horndeski gravity, density parameters, $\Delta\chi^2 = -35.9$ vs $\Lambda$CDM. 16+ observables, 3 falsification tests (2028–2035).
Modified hi_class Boltzmann solver with SDGFT modifications for CMB power spectra and large-scale structure predictions.
LaTeX sources for the foundational paper (v1 & v2): all derivations, proofs, numerical appendix, and the complete scorecard. Published on Zenodo.
Independent computational validation of SDGFT predictions via neural surrogates and Bayesian inference.
GATv2 graph-neural-network ensemble, 100M-point Oracle database, and CVAE inverse solver. Maps 37 observables with $R^2 = 0.9995$ round-trip accuracy.
ML surrogate, inverter & anomaly detector. PDG 2024 scorecard: 20/22 within 2σ, $\chi^2/\text{ndof} = 1.48$. Full training pipelines and reproducible notebooks.
Full Bayesian MCMC comparison: SDGFT vs $\Lambda$CDM with emcee, dynesty, and cobaya samplers. Fits BAO, CMB, SNIa, RSD, and gravitational-wave data.
Interactive Flask web application for SDGFT parameter-space visualization, scenario comparison, and publication-quality plots.
GW strain analysis, external datasets, and download utilities.
Gravitational-wave strain analysis vs SDGFT predictions: modified dispersion relations, bright sirens, and waveform corrections across 14 GWOSC events.
Download scripts and documentation for external datasets: GWOSC strain data, Pantheon+ supernova catalog, SPARC galaxy rotation curves, and SDSS-III BAO measurements. Raw data not included — use the provided scripts to fetch from original sources.
All active public repositories under github.com/cosmologicmind.
| Repository | Description | Stack | License |
|---|---|---|---|
| sdgft | Core library — 89+ observables, zero free parameters | Python | MIT |
| sdgft-cosmology-engine | Parameter-free cosmological predictions (Horndeski / hi_class) | Python | MIT |
| sdgft-ml-toolkit | GATv2 GNN ensemble + 100M-point Oracle DB + CVAE | PyTorch | MIT |
| sdgft-machine-learning | ML surrogate, inverter & anomaly detector | PyTorch | MIT |
| sdgft-mcmc | Bayesian MCMC: emcee, dynesty, cobaya | Python | MIT |
| sdgft-gw-analysis | GW strain analysis (14 GWOSC events) | Python | MIT |
| sdgft-ml-explorer | Interactive Flask app for parameter-space exploration | Flask | MIT |
| class_blotzman | Modified hi_class Boltzmann solver | C / Python | MIT |
| sdgft-foundational-paper | LaTeX sources for the foundational paper (v1 & v2) | LaTeX | MIT |
| sdgft-data | External dataset download scripts (GWOSC, Pantheon+, SPARC, BAO) | Shell | — |
Every prediction is reproducible. Clone the repos, run the tests, verify every observable yourself.