feat(feff10): add parity coverage and SFCONV support

[Ameyanagi]

Summary

• add FEFF10 parity and A*S02(k) stability regression coverage against regenerated ZnSe fixtures
• thread an optional use_sfconv flag through the Tauri and React fit flow for FEFF10 runs
• upgrade the FEFF10 runner to the 0.2 API and improve strict-card parsing error hints

Testing

• not run in this PR flow

[greptile-apps]

Greptile Summary

This PR wires an optional use_sfconv flag end-to-end through the FEFF10 pipeline — from the Tauri DTO and React checkbox down to FeffInput.other_cards injection in the runner — and upgrades the FEFF10 runner to the 0.2 API with stricter card-parsing error hints. New regression tests assert FEFF10/FEFF85L parity and A×S02(k) amplitude stability across varied k-ranges.

Confidence Score: 5/5

Safe to merge; only P2 style findings remain

No P0 or P1 issues found. The SFCONV injection logic is correctly guarded at the Tauri command layer, type safety is maintained across the Rust/TS boundary, the 0.2 API upgrade is clean, and the parity tests provide meaningful regression coverage. Two P2 suggestions (restore frontend CI and add a library-level guard for use_sfconv) do not block merge.

.github/workflows/rust.yml (CI coverage gap) and xraytsubaki-app/src-tauri/src/commands/fitting.rs (use_sfconv library-level validation)

Important Files Changed

Filename	Overview
.github/workflows/rust.yml	Removes the entire gui CI job (TypeScript type-check, lint, Tauri build check) with no replacement
crates/xraytsubaki/src/xafs/fitting/runner.rs	Upgrades FEFF10 runner to 0.2 API; adds SFCONV injection via ensure_other_card_present; improves strict-card parse error hints
crates/xraytsubaki/src/xafs/fitting/types.rs	Adds use_sfconv field to FeffRunRequest with serde default = false
crates/xraytsubaki/tests/feff10_s02k_parity.rs	Adds FEFF10 vs FEFF85L parity test and A×S02(k) amplitude stability regression test across k-ranges
xraytsubaki-app/src-tauri/src/commands/fitting.rs	Threads use_sfconv through run_feff_paths Tauri command with validation guard; adds unit tests for resolve_feff_run_mode
xraytsubaki-app/src-tauri/src/dto.rs	Adds use_sfconv bool field with serde default to FeffRunConfig DTO
xraytsubaki-app/src/backend/types.ts	Adds optional use_sfconv field to FeffRunConfig TypeScript interface
xraytsubaki-app/src/panels/FitPanel.tsx	Adds SFCONV checkbox to FEFF run section; wires use_sfconv into runConfig state
Cargo.lock	Routine dependency version bumps (anyhow, nalgebra, syn, bitflags, etc.)

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[FitPanel.tsx\nuse_sfconv checkbox] -->|FeffRunConfig| B[run_feff_paths\nTauri command]
    B --> C{resolve_feff_run_mode}
    C -->|explicit_executable AND use_sfconv| D[Error: SFCONV only in FEFF10 mode]
    C -->|no explicit_executable| E[Feff10Pipeline mode]
    C -->|explicit_executable\nuse_sfconv=false| F[Feff85LModules mode]
    E --> G[run_feff10_pipeline]
    G --> H{use_sfconv?}
    H -->|true| I[ensure_other_card_present\nappend SFCONV card]
    H -->|false| J[FeffInput unchanged]
    I --> K[FeffPipeline.run]
    J --> K
    F --> L[run_feff85l_modules]
    K --> M[FeffRunResult\npath_files]
    L --> M
    M --> N[FitPanel paths state\nfeff_dat_path list]

Loading

Prompt To Fix All With AI

This is a comment left during a code review.
Path: xraytsubaki-app/src-tauri/src/commands/fitting.rs
Line: 34-46

Comment:
**`use_sfconv` silently ignored at the library level**

The Tauri command layer correctly rejects `use_sfconv: true` combined with an explicit executable, but this guard exists only there. A library consumer who directly builds a `FeffRunRequest` with `mode: Feff85LModules` and `use_sfconv: true` and then calls `run_feff()` will get a silent no-op — the flag is never read by `run_feff85l_modules`. Consider adding a `debug_assert!(!(request.use_sfconv && request.mode == FeffExecutionMode::Feff85LModules))` inside `run_feff85l_modules`, or documenting clearly that `use_sfconv` is only honoured in `Feff10Pipeline` mode.

How can I resolve this? If you propose a fix, please make it concise.

---

This is a comment left during a code review.
Path: .github/workflows/rust.yml
Line: 76

Comment:
**Frontend CI coverage removed**

The `gui` job — which ran TypeScript type-checking (`npm run type-check`), ESLint (`npm run lint`), and `cargo check` on the Tauri backend crate — was dropped entirely with no replacement. This PR introduces new TypeScript types and a React checkbox component; without the job, regressions in the frontend can slip through undetected. Consider restoring the job or adding a leaner type-check/lint step.

How can I resolve this? If you propose a fix, please make it concise.

_{Reviews (3): Last reviewed commit: "fix: address FEFF10 PR review findings" | Re-trigger Greptile}