[jacksteamdev]

Summary

Add current PAC/IAE Lahiri alongside existing chart references in Rust, Wasm, and the browser example.

Lahiri selection
  → PAC mean ayanamsa + engine nutation
  → shared longitude, speed and house projection

Benchmark investigation
  → separate time conversion, bodies, geometry and projection
  → identify baseline discrepancies requiring follow-up

Existing defaults and Fagan–Bradley formulas remain unchanged. No new regression fixtures have been frozen.

Evidence

• Before: the initial five-chart comparison passed only one complete chart; time, geometry and projection discrepancies were mixed together.
  After: explicit-time comparisons isolate the failures. Lahiri and Fagan–Bradley projection transforms pass all 125 sampled dates. Tropical bodies at matched TT pass 123/125.
• Remaining body failures: Saturn on 1887-07-19, approximately 1.325″ longitude and 2.092″/day speed; Ceres on 2097-04-16, approximately 1.150″ longitude.
• Geometry remains incomplete: matched-input Whole Sign/Equal geometry passes 110/125 cases per zodiac; Placidus passes 109/121 admitted cases. Four additional polar Placidus cases per zodiac are rejected by the candidate.
• Independent east-vector checks confirm western-horizon root selection in the four disputed polar ascendants. No geometry repair is included.
• Rust workspace tests, Clippy, formatting, Worker structural/adapter and dataset checks, and Python verification passed during the investigation. Two dataset-dependent Rust tests remain ignored.
• Earlier browser-example build, TypeScript check and five native/Wasm comparisons passed. Browser walkthrough, exact frozen-dataset parity and full coverage validation remain outstanding.
• Requests, reference outputs and diagnostic harnesses remain outside the repository. Projection agreement does not establish full-chart accuracy.

Follow-up order: polar orientation and cusp fallback; explicit TT/UT1 contract; remaining frame discrepancies; Saturn/Ceres fitting-versus-source investigation. Preserve existing fixtures and unchanged tolerances.

Merge Danger

Door: two-way

The additive Lahiri implementation can be reverted. Baseline numerical repairs are not included and require separate compatibility review.

Blast Radius: charts

Keep this PR draft until its benchmark acceptance requirements are resolved. Future corrections may change historical chart positions, house assignments and event times. No consumer rollout or release is included.