improved depletion plan

[shimwell]

just making this issue to keep track of a possible break down of the PRs needed to close out #2757 on openmc

PR 1: Energy-dependent isomeric branching in chain files (@shimwell PR)

• Chain.from_endf() reads MF=9/MF=10
• Nuclide.isomeric_branching stores/reads/writes energy-dependent yields
• LFS ordering for state mapping

PR 2: ELIS matching for isomeric state identification (perhaps @jckica) could make this PR

• requires PR 1 to be merged (or can stack on top)
• demonstrated by @jckica on https://github.com/jckica/isomer_branching_jendl
• Read ELFS from MF=8 for each product state
• Read ELIS from the decay library for each candidate metastable nuclide
• Match by excitation energy within tolerance instead of assuming LFS=n → _m{n}
• Falls back to LFS ordering when no match found

PR 3: (n,n') reaction support (perhaps @yrrepy could make this PR)

• demonstrated by @yrrepy on branch fork
• Add "(n,n')": ReactionInfo({4}, ()) to REACTIONS dict
• Add "(n,n')": (0, 0) to MASS_NUMBER_CHANGE
• Add REACTION_MT["(n,n')"] = 4
• Add C++ alias in reaction_type()
• This unlocks inelastic scattering activation (isomeric state changes via (n,n'))

PR 4: Runtime spectrum-dependent branching collapse (perhaps @jbae11)

• requires PR 1 to be merged (or can stack on top)
• form_matrix() accepts isomeric_branching parameter (energy-collapsed ratios per material)
• New helper that takes local multigroup flux + Nuclide.isomeric_branching yield data and collapses to effective scalar ratios
• Thread through _matrix_funcs.py and pool.py

PR 5: Operator integration (perhaps @yrrepy)

• IndependentOperator reads isomeric_branching from chain, collapses with flux at each depletion step
• CoupledOperator gets direct_with_flux reaction rate mode to tally flux spectrum alongside reaction rates
• Per-material branching ratios updated each step

PR 6: GENDF cross-section support (optional) (perhaps @yrrepy could make this PR)

• demonstrated by @yrrepy on branch fork
• GENDF library parser
• get_gendfxs_and_flux() as alternative to CE cross sections
• Broader nuclide coverage than HDF5 transport libraries

Related

Review jbae11/openmc_activator#41
rerun openmc activator for V&V if all PRs get merged in. See if OpenMC and Fispact are closer and if OpenMC perhaps @jbae11 is best placed to do that
rerun fusion activation handbook if all PRs get merged. See if anything changes

@eepeterson can perhaps keep an overview and make sure we head in the right direction without overlapping with MIT work