Task 133: 各解析ごとの再現コマンド（解析テスト）

Status: DONE

Human Directive

「再現コマンドは、レポートに対してひとつではなく、各解析に対してひとつ作っておく。それをテストのように扱うことで、TDD的に解析を行うことができる。また、自動テストのように、常に解析されている状態を維持する。それによって、前提条件が変わった際にDAGを確認するよりも確実に再度解析を行うことができる。」

Scope

Audit existing analyses (EP01-05 transfers, parameter explorations) for reproduction commands
Design per-analysis reproduction command format (CLI invocation that re-derives specific analysis results)
Implement individual reproduction scripts for each transfer analysis:

- EP01: Mars→Ganymede 72h brachistochrone, mass boundary, Hohmann reference, Jupiter arrival

- EP02: Jupiter escape, ballistic coast, Saturn capture, Enceladus descent

- EP03: Enceladus departure, Saturn→Uranus transfer, Titania approach, navigation accuracy

- EP04: Titania departure, thrust-limited transfer, plasmoid encounter, Earth approach

- EP05: Uranus departure, Jupiter flyby, composite route, LEO insertion, nozzle analysis

Register these as automated tests (npm run test:analyses or similar)
Integrate with CI so analyses are always verified
Update npm run recalculate to use the per-analysis commands

Design Notes

Each analysis test should:

- Run the calculation script/WASM function

- Compare output against expected values (with tolerance)

- Fail clearly when preconditions change (mass, Isp, orbital parameters)

Think of them as "golden file" tests for analysis results
When a test fails due to changed preconditions, the analysis must be re-examined (not just the test updated)
This is more reliable than DAG invalidation because it catches actual numerical discrepancies

Dependencies

solar-line-core (Rust orbital mechanics)
solar-line-wasm (WASM bridge)
Existing calculation scripts in ts/
reports/data/calculations/ (current calculation outputs)