Tearing - Record the per-surface Zeff and the separatrix q-law in gpec.h5 #527

Two things the tearing analysis computes are not in the output file, so a result cannot be audited from gpec.h5 alone (follow-up to PR #298).

• Zeff at each rational surface. SLAYER takes Zeff(ψ) from the kinetic file's zeff dataset, else from n_i/n_e with zimp, else from the scalar fallback. The source is logged but not stored, and neither are the values. Tearing/PerSurface/ already holds the other derived layer inputs (rs, sval_r, lu, D_norm, P_perp); Zeff and its source belong beside them.
• The separatrix q-law. ForceFreeStates/LayerOverlap/ records which surfaces were placed on the law (extrapolated) but not the law itself: X-point flux and position, coefficient per X-point, anchor, and accepted shear ratio. The X-point flux in particular shows whether a g-file's boundary sits inside its separatrix, which decides whether the outermost pair can be scored.

The scan's caveats and notes are also only logged; consider storing them.

Done when:

• Names and locations follow docs/development/hdf5-conventions.md (derived output only).
• Writer, schema test and docs are updated together.
• diiid_slayer_n1 tracks Zeff and diiid_truncation_n1 tracks the X-point flux.

🤖 Generated with Claude Code