Equilibrium - BUGFIX - Build the two-pass ψ grid from kinetic profiles that have no omega_E #614

A run whose kinetic file has no omega_E now stops in the equilibrium stage on the default grid (mpsi = 0), even when only SLAYER uses the profiles. Before #581, the same deck ran: SLAYER warned and took Ω_E = 0. The HDF5 kinetic schema treats omega_E as optional.

Reproduced on develop 4f33fcf60:

• Deck: examples/DIIID-like_SLAYER_example.
• Kinetic file: its kinetic file copied without the omega_E dataset (n_e, n_i, T_e, T_i, chi_e, chi_i and chi_phi kept).
• Result: main fails with ErrorException: kinetic profiles (...) missing required 'omega_E'.

Stack, innermost first:

kinetic_profile_splines        Equilibrium/KineticProfiles.jl:405   (_need(data.omega_E, "omega_E"))
ntv_profiles                   KineticForces/Utils.jl:304
maybe_reform_equilibrium       GeneralizedPerturbedEquilibrium.jl:498
main_from_inputs               GeneralizedPerturbedEquilibrium.jl:262

Cause:

• maybe_reform_equilibrium builds the NTV view of the attached profiles (KineticForces.ntv_profiles) whenever equil.kinetic !== nothing. It feeds refined_psi_grid(kin=...) and implied_knot_count as well as the kinetic-resonance nodes, so it is built before any check on the kinetic source.
• The NTV view requires omega_E.
• The same path runs for a scripting-API solve on an equilibrium with such profiles attached.
• prepare_force_free_states! likewise builds the NTV view whenever kinetic_factor > 0, including kinetic_source = "fixed", which does not need profiles.

Suggested fix:

• Give the grid refinement a view that tolerates missing NTV fields, or pass kin only when the NTV fields are present.
• Build the NTV view in FFS preparation only for kinetic_source = "calculated".
• Add a test that runs maybe_reform_equilibrium with an omega_E-less KineticProfileData. The current omega_E-less test covers only Runner._slayer_profiles.

Found in review of #581.