Repository map¶

One page, the whole tree: what lives where, and which document explains it. Every folder carries its own README with deeper detail.

The pipeline¶

flowchart LR
    subgraph DATA["Frozen data layer"]
        Z["verification/data/<br/>2 001 058 ζ zeros · SHA-256"]
    end
    subgraph CORE["Canonical numerics"]
        J["code/ab_cloud_full.jl<br/>39 tests · two-pass"]
    end
    subgraph CROSS["Cross-language verification"]
        V["10 languages<br/>verification/"]
        S64["spinor64<br/>64/64 GUE-consistent"]
    end
    subgraph FORMAL["Formal layer"]
        L["Lean 4<br/>theorems + re-derivation"]
        C["Coq · Agda · Isabelle"]
    end
    subgraph OUT["Publications & apps"]
        M["projects/monographs/<br/>v22 RU/EN/ZH + v21 + v2.2.1"]
        R["results/<br/>flagship artifacts"]
        APP["apps/<br/>dashboard + 3D lab"]
    end
    Z --> J
    Z --> V
    Z --> L
    J --> R
    V --> S64
    J --> M
    R --> APP
    C -.-> L

The tree¶

Path	What lives there	Read this
`code/`	the canonical Julia suite (v23, 39 tests), RH sweep benches, the H-P audit battery	`code/README.md`
`code/julia/`	the v23 provenance copy + version lineage notes	`code/julia/README.md`
`verification/`	the 10-language cross-verification suite, spinor64, ζ datasets, per-section micro-verifications	`verification/README.md`
`formal/`	the formal layer — Lean 4 (theorems + executable), Coq, Agda, Isabelle/HOL, all dependency-free	`formal/README.md`
`formal/lean4/`	the flagship Lean project: exact layer, numeric port, `abcloud-verify` executable	`formal/lean4/ABCloud/README.md`
`projects/hilbert-polya/`	the Hilbert–Pólya bridge subproject: campaigns C1–C8, theorems T1–T10, monograph volumes	`projects/hilbert-polya/README.md`
`projects/monographs/`	five monograph editions (v22 RU/EN/ZH + original v21 RU/EN) + v2.2.1 corrections, 600-dpi figure apparatus	`projects/monographs/README.md`
`projects/lab-3d/`	the 3D lattice laboratory: code, committed outputs, preprint bundle	`projects/lab-3d/README.md`
`projects/dirac-lab/`	the Dirac laboratory: D1–D14 ledger, 96 checks, all PASS	`projects/dirac-lab/README.md`
`projects/vortex-optimization/`	the Test 34-ext vortex-density optimization study (best D = 0.0318)	`projects/vortex-optimization/README.md`
`projects/wave-attractors/`	the hyperbolic wave-attractors reproduction study (W1)	`projects/wave-attractors/README.md`
`projects/meridian-observatory/`	24 standalone Julia spectral benches	`projects/meridian-observatory/README.md`
`apps/`	the React dashboard and the WebGL 3D laboratory (with prebuilt bundles)	`apps/README.md`
`results/`	the flagship run artifacts (453 files, 39 directories) + reference logs	`results/README.md`
`docs/`	this MkDocs Material site	`docs/README.md`
`assets/`	the repo banner and shared art	`assets/README.md`
`.github/`	the nine CI workflows, templates, dependabot, labeler	`.github/WORKFLOWS.md`

The one rule¶

Every arrow in the pipeline reads the same frozen dataset. The Julia canon, the ten ports and the formal layer never re-download or regenerate the ζ zeros — so a disagreement between any two boxes is immediately visible rather than hidden by different inputs.