[jacksteamdev]

Lahiri now reports the IAE 2027 mean ayanamsa plus the full 77-term IAU 2000B nutation through lahiri_ayanamsa and chart metadata. Chart projection and speeds retain the existing five-term model, preserving the source branch's sidereal calculations. The documented difference between the reported and projection offsets is up to 0.27 arcseconds.

Includes the initial Lahiri support from export source bc31f432e35d6a2da900e038fd8ca27ac8beeac8, which is not yet on main, plus the returned project's implementation and 1,593-case regression fixture. Base: refreshed main 032a3b653b50f1ac509b0488705914eb4b7a12ed.

Import reconciliation and review

• Mapped original AstrologyEngine-library/ and returned files/AstrologyEngine-library/ to repository root. Verified the matching returned ZIP against every supplied folder payload. Used the explicit partial-bundle diff; no deletions inferred from missing files.
• Imported 7 new files (nutation module, fixture and tests, fixture generator, plan and research record) and modified 10 existing files, including initial Lahiri prerequisite changes. No deletions or relocations.
• Reconciled exported content against the recorded source and current main. Adapted README edits to the real README; updated chart conventions and marked external experiment results as historical, unverified import claims.
• Excluded the export-only scripts/review.sh, loose PDF deletion, omitted benchmark/records.json, and unchanged export-only dataset/benchmark attachments. Dependencies, lockfiles, hooks and CI remain unchanged.
• Reviewed imported code and execution paths before running it. Hardened the fixture generator to require a full commit hash and reject untracked runtime files. Kept fixture bytes, values and provenance unchanged.

Validation

• Rust 1.96.0: workspace tests (40 passed, 2 existing ignored); default-feature tests (22 passed, 2 existing ignored).
• Formatting and Clippy with warnings denied: workspace, default-feature library, Worker verification crate and browser Wasm crate.
• Workspace release build, library wasm32 release build, and offline Cargo package verification.
• Browser production Wasm/Vite build and TypeScript check; Worker verification build.
• Local workerd: 6 structural cases and 11 adapter checks passed, including cache recovery, chart payload equivalence and Sun searches.
• CI-helper tests (10), CI TypeScript check, runtime dependency/I/O boundary check and usage-policy hook tests passed.
• Fixture-generator rejection checks: untracked source, mismatched source, option-like revision and failing benchmark report.
• Initial/final staged Gitleaks scans passed using external default-rule configuration, --ignore-gitleaks-allow --redact=100; staged policy scan and git diff --check passed.

The local harness needed an outside-sandbox rerun to bind localhost. Cargo package's registry refresh was blocked; offline package verification then passed with locked cached dependencies. Both Bun installs used frozen lockfiles and disabled lifecycle scripts.

Limits

External NOVAS reports, returned-environment Git history and the claimed large chart sweep were not supplied or rerun. Fixture regeneration therefore remains unavailable without recovering those artifacts or explicitly revalidating against a new benchmark. No frozen full-dataset parity, dataset regeneration, visual browser walkthrough, hosted checks or deployment was performed. Passing local regression checks does not establish independent astronomical accuracy.

Archive SHA-256:

• Original: 1ae213207684ba1764a3593926931a2d9205a335151d4061798edb0d737ce97b
• Returned: 12dee25cb72825e380cba579f2de615e28ec06f02c0e7da8c43275522bc89abd