antikythera-spectral roadmap¶

Status of advertised CLI / Bridge API methods. Mirror of the chess-spectral ROADMAP discipline: every entry the help text advertises is either ✅ wired (real implementation behind it) or ⏳ pending (planned for a specific version).

Released versions¶

Version	Date	Headline
v0.3.0	2026-05-02	Mars planetary models + bit-packed HDC ALU + default-backend flip. New `mars_models` and `bit_alu` facades; bronze projection + Freeth-reconstruction param sets; `compare.compare_models_at_jd` and `bridge.compare_models` accept `bronze` + `params`; ADR 0012 (algebra/eigenbasis vs CAD scope); Mars 38° gap audit reproduced as `mars_models.fj_38deg_finding()`; bit-ALU benchmark + cycle-alignment investigation. Default backend flipped to `bit` across `default_encode`, `bridge.get_dial_state` / `decode_dial` / `decode_to_jd`, and the CLI; `backend="complex128"` available as v0.2.x opt-out. 186 tests passing (50 immolation including 23 v0.3.0 release-gate rows; 16 default-backend tests; 10 bit-ALU primitive tests; 9 mars_models tests).
v0.2.1	2026-04-30	PyPI metadata hotfix (project URL `Demo` → kiosk site).
v0.2.0	2026-04-30	Self-contained mode: visibility / heliacal / elongation / eclipses all default to algebraic propagation; skyfield optional via `[ephemeris]` extras. ADR 0011.
v0.1.0	2026-04-30	First public release on PyPI. 28-method bridge, 9 facade modules, codegen with provenance manifest, 31-row H-battery, CLI console script.

Next planned¶

Version	Theme
v0.4.x	Numpy-fast `bit_alu.permute` + `np.bitwise_count`-backed similarity (close the few remaining bit-ALU operations that still trail the float reference). H-battery rows for the bit-ALU encoder (does it pass the same B-H1 / E-H1 round-trip checks as the complex128 encoder?).
later	Apply the #5–#10 Mars-gap framework to Venus / Mercury / Jupiter / Saturn (each has its own theoretical-error figure worth auditing under the same discipline).

The phase table below is preserved for historical context — most entries shipped during v0.1.0 / v0.2.0 / v0.3.0 even though the inline status says "pending."

Current release: v0.3.0 (released 2026-05-02)¶

Phase status¶

Phase	Status	Description
0 — Plan	✅ shipped	Planning document
1 — Skeleton + codegen	✅ in progress	`pyproject.toml`, version stamp, `codegen/` + `_data/`
2 — Facade re-exports	⏳ pending	encoder / decoder / dials / render / hypotheses / ephemeris / eclipses / periods / gears
3–11 — Bridge API (28 methods)	⏳ pending	state↔date, calendar, astronomy, operator, comparators, what-if, archaeology, Goal-Year, animation, H-battery
12 — H-battery facade	⏳ pending	wraps consolidated_tests
13 — CLI	⏳ pending	`antikythera-spectral` console script
14 — CodeQL paths	⏳ pending	update `.github/codeql/codeql-config.yml`
15 — Publish + autotag workflows	⏳ pending	`antikythera-spectral-publish.yml`, `antikythera-spectral-autotag.yml`
16 — Docs + READMEs	⏳ pending	bridge_api.md, DELTA_T_MODEL.md, EPHEMERIS_KERNELS.md, CALENDAR_SYSTEMS.md, OPERATOR_WORKFLOW.md, 10 ADRs
17 — TestPyPI dry-run	⏳ pending	workflow_dispatch with target=testpypi
18 — Pre-merge gate	⏳ pending	§15.1 of plan, all 9 boxes ticked against TestPyPI
19 — Version bump → merge	⏳ pending	bump 0.1.0rc1 → 0.1.0, merge PR
20 — Autotag → real PyPI	⏳ pending	autotag fires, publish workflow lands 0.1.0 on pypi.org

Bridge methods (target: 28)¶

Read methods (state ← date) — phase 4¶

Method	Status
`get_dial_state(jd_tdb)`	⏳ pending
`get_dial_angle(jd_tdb, dial)`	⏳ pending
`get_pointer_xy(jd_tdb, *, layout)`	⏳ pending
`get_all_dial_metadata()`	⏳ pending
`get_version()`	⏳ pending

Date ← state — phase 4¶

Method	Status
`decode_dial(state_vec, dial)`	⏳ pending
`decode_to_jd(state_vec)`	⏳ pending

Calendar conversion — phase 5¶

Method	Status
`jd_to_gregorian(jd_tdb)`	⏳ pending
`gregorian_to_jd(...)`	⏳ pending
`jd_to_julian_calendar(jd_tdb)`	⏳ pending
`jd_to_athenian(jd_tdb, archon_table)`	⏳ pending
`jd_to_olympiad(jd_tdb)`	⏳ pending

Astronomy — phase 6¶

Method	Status
`get_visibility_windows(jd_lo, jd_hi, planet)`	⏳ pending
`get_next_heliacal_rising(jd, planet)`	⏳ pending
`get_solar_elongation(jd, planet)`	⏳ pending
`get_eclipse_anchors(era)`	⏳ pending
`get_period_relations(source)`	⏳ pending
`find_eclipses(jd_lo, jd_hi, kind)`	⏳ pending

Operator workflow (§11.6.16) — phase 7¶

Method	Status
`start_operator_session(initial_jd, dials)`	⏳ pending
`operator_advance(state, delta_days)`	⏳ pending
`operator_observe(state, planet, observed_residue)`	⏳ pending
`operator_diagnostics(state)`	⏳ pending
`set_anchor(dial, jd, observed_residue)`	⏳ pending
`apply_anchor(state, calibration_delta)`	⏳ pending

Cross-comparators — phase 8¶

Method	Status
`compare_ephemerides(jd, body, kernel_a, kernel_b)`	⏳ pending
`compare_models(jd, body, model_a, model_b)`	⏳ pending
`compare_reconstructions(jd, dials)`	⏳ pending

What-if & archaeology — phase 9¶

Method	Status
`encode_with_custom_train(jd, dial, p, q)`	⏳ pending
`compare_to_ground_truth(jd, dial, p, q)`	⏳ pending
`get_fragment_inventory(fragment)`	⏳ pending

Babylonian Goal-Year — phase 10¶

Method	Status
`goalyear_predict(planet, jd, source)`	⏳ pending
`goalyear_compare(planet, jd)`	⏳ pending

Animation export — phase 11¶

Method	Status
`encode_range(jd_lo, jd_hi, step_days)`	⏳ pending
`export_animation(jd_lo, jd_hi, step_days, format)`	⏳ pending

H-battery — phase 12¶

Method	Status
`run_hypothesis_battery(ephemeris)`	⏳ pending
`get_hypothesis(id)`	⏳ pending

Beyond v0.1.0 (deferred)¶

Native C accelerator (only if profiling demands; would trigger dual-wheel pattern from chess-spectral)
DE-kernel cross-validation as a hypothesis row (H-H3)
Parapegma heliacal-event readout (needs digitisation work)
Drag-and-drop web UI (separate frontend project)
3D rendering of the bronze stack
Multi-language bindings (TS / Rust / Go) — Pyodide IS the JS surface

See ../ANTIKYTHERA_SPECTRAL_PYPI_PLAN_v0.1.0.md §11 for the full deferred-features list.