EolisaSpaceEngine v7.0.0 — Dual-Source Bayesian Framework for Shadow-Deviation Constraints on Exotic Compact Objects (Sgr A and M87)**

Abstract

This release provides the complete, reproducible pipeline used by the Eolisa Space Research Team to place quantitative constraints on exotic compact object (ECO) alternatives to the Kerr black hole, using the public Event Horizon Telescope visibilities of both Sagittarius A* (EHT release 2022-D02-01) and M87* (EHT release 2019-D01-01), together with GRAVITY astrometry of the Sgr A* infrared flares. The framework computes the Kerr shadow from first principles by integrating the Bardeen (1973) photon region, and expresses each competing spacetime Simpson–Visser wormholes, solitonic boson stars, gravastars, Johannsen spacetimes, and regular black holes as an explicit, citation-backed fractional shadow deviation δ from that prediction. Two dominant systematic uncertainties are marginalised directly inside the joint likelihood rather than appended afterward: the source's mass-to-distance ratio θ_g (the GRAVITY 2022 precision measurement for Sgr A*, and a GR-independent stellar-dynamics mass for M87*, chosen specifically to avoid circularity with the EHT ring-derived mass), and the documented coefficient uncertainty σ_δ of every ECO ansatz. Visibility amplitudes, band-separated closure phases built directly from the released data, and GRAVITY astrometry enter as three genuinely active likelihood terms. Bayesian evidence and posteriors are obtained through nested sampling (dynesty, with a dependency-free NumPy fallback for minimal environments), yielding model comparison via Bayes factors, a direction-aware Jeffreys-scale interpretation, AIC/BIC, and credible intervals. Beyond model comparison, the engine reports its headline, model-agnostic result: the marginal posterior on the fractional shadow-size deviation δ, cross-checked against an exact per-term goodness-of-fit diagnostic and validated through synthetic injection recovery testing. Version 7.0.0 is a full audit-drivenn rewrite resolving critical issues identified in v6 a misspecified error model, spurious Bayesian evidence, cross-band closure-phase contamination, and missing likelihood normalisation and extesnds the framework from one source to two. Every run is deterministic given a seed; every figure, posterior, and report shipped with this release is generated by the code itself, not transcribed by hand. Scope. This is a constraint framework, not a discovery claim. It answers one precise question for each source how large a shadow deviation from Kerr the 2017 data permit, once every known systematic is honestly included and it reports non-detections. Mapping a measured δ to a specific metric parameter requires full GRMHD ray-tracing and is flagged as future work. The repository includes the full analysis pipeline, two methodology papers (Paper I: Framework; Paper II: Validation and Constraints), complete documentation, and machine-readable citation metadata. Released under the MIT License. Onur H. Evgin, President of Eolisa Space — Eolisa Space Research Team eolisaspace.com · presidency@eolisaspace.com