Add k-infinity estimator for IndependentOperator

[EdenRochmanSharabi]

Summary

• Add calculate_kinf=True parameter to IndependentOperator and from_nuclides (mutually exclusive with keff)
• Estimate k-infinity as the ratio of nu-fission production rate to neutron loss rate, following OpenMC's keff convention: (n,xn) reactions reduce the loss term by (x - 1) rather than contributing to production
• Add 'nu-fission' as a valid reaction in MicroXS, with a new _collapse_nu_fission() helper that integrates nu(E)*sigma_f(E) from pointwise ENDF data against the user-provided multigroup flux
• The estimate is reported as the operator eigenvalue and flows into Results.get_keff()

Changes

• openmc/deplete/independent_operator.py: _neutrons_emitted() parser, _estimate_k_inf() method (MPI-aware), calculate_kinf parameter with validation
• openmc/deplete/microxs.py: 'nu-fission' in _valid_rxns, _collapse_nu_fission() for pointwise collapse, from_multigroup_flux() routing for nu-fission
• docs/source/usersguide/depletion.rst: user guide section on calculate_kinf
• tests/unit_tests/test_deplete_independent_operator.py: analytic k_inf (with (n,2n)), multigroup flux, decay step, _neutrons_emitted table, error paths
• tests/unit_tests/test_deplete_microxs.py: nu-bar validation against ENDF data (U235 thermal ~2.43)

Test plan

• Analytic k_inf matches expected value including (n,2n) contribution
• Multigroup flux weighting produces correct result
• Decay step (zero source rate) still reports k_inf estimate
• _neutrons_emitted handles all reaction name patterns correctly
• ValueError raised for missing nu-fission data and keff/calculate_kinf conflict
• CI (17/17 passed, coverage +0.01%)

Closes #2132